SH // ' A 2. f!S>H U.S. Department of Commerce Volume 112 Numbers 2-3 April-iuly 2014 Fishery Bulletin US. Department of Commerce Penny S. Pritzker Secretary National Oceanic and Atmospheric Administration Kathryn D. Sullivan NOAA Administrator National Marine PssSieries Service Eileen Sobeck Administrator for Fisheries Scientific Editor Bruce C. Mundy Associate Editoo- Kathryn Dennis National Marine Fisheries Service Pacific Islands Fisheries Science Center Fisheries Research and Monitoring Division 1845 Wasp Blvd., Bldg. 176 Honolulu, Hawaii 96818 Managing Editor Sharyn Matriotti National Marine Fisheries Service Scientific Publications Office 7600 Sand Point Way NE Seattle, Washington 98115-0070 The Fishery Bulletin (ISSN 0090-0656) is published quarterly by the Scientific Publications Office, National Marine Fia- heries Service, NOAA, 7600 Sand Point Way NE, Seattle, WA 98115-0070. Although the contents of this publica- tion have not been copyrighted and may be reprinted entirely, reference to source is appreciated. The Secretary of Commerce has de- termioned that the publication of this periodical is necessary according to law for the transaction of public business of this Department. Use of funds for print- ing of this periodical has been approved by the Director of the Office of Manage- ment and Budget. For Sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, DC 20402. Subscrip- tion price per year: $32.00 domestic and $44.80 foreign. Cost per single issue: $19.00 domestic and $26.60 foreign. See back for order form. Editorial Committee Richard Brodeur National Marine Fisheries Service, Newport, Oregon John Carlson National Marine Fisheries Service, Panama City, Florida Kevin Craig National Marine Fisheries Service, Beaufort, North Carolina Jeff Leis Institute for Marine and Antarctic Research Univ. Tasmania, Hobart, Australia Rich McBride National Marine Fisheries Sen/ice, Woods Hole, Massachusetts Rick Methot National Marine Fisheries Service, Seattle, Washington Adam Moles National Marine Fisheries Service, Auke Bay, Alaska Frank Parrish National Marine Fisheries Service, Honolulu, Hawaii Dave Somerton National Marine Fisheries Seivice, Seattle, Washington Ed Trippel Department of Fisheries and Oceans, St. Andrews, New Brunswick, Canada Mary Yoklavich National Marine Fisheries Service, Santa Cruz, California Fishery Bulletin web site: www.fisherybulletin.noaa.gov The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and economics. It began as the Bulletin of the United States Fish Commission in 1881; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Separates were issued as documents through volume 46; the last document was No. 1103. Beginning with volume 47 om 1931 and continu- ing through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bul- letin. Beginning with Volume 70, number 1, January 1972, the Fishery Bulletin became a peri- odical, issued quarterly. In this form, it is available by subscription from the Superindentent of Documents, U.S. Government Printing Office, Washington, DC 20402. It is also available free in limited numbers to libraries, research institutions, State and Federal agencies, and in exchange for other scientific publicarions. U S. Department of Commerce Seattle, Washington Volume 112 Numbers 2-3 April-July 2014 Fishery Bulletin Contents Articles 101-111 Long, W. Christopher, Peter A. Cummiskey, and J. Eric Munk Effects of ghost fishing on the population of red king crab ( Paralithodes camtschaticus) in Womens Bay, Kodiak Island, Alaska 112-130 Clemento, Anthony J., Eric D. Crandall, John Carlos Garza, and Eric C. Anderson Evaluation of a single nucleotide polymorphism baseline for genetic stock identification of Chinook Salmon ( Oncorhynchus tshawytscha ) in the California Current large marine ecosystem The National Marine Fisheries Service (NMFS) does not approve, recommend, or endorse any proprie- tary product or proprietary material mentioned in this publication. No reference shall be made to NMFS, or to this publication furnished by NMFS, in any advertising or sales promotion which would indicate or imply that NMFS approves, rec- ommends, or endorses any propri- etary product or proprietary mate- rial mentioned herein, or which has as its purpose an intent to cause directly or indirectly the advertised product to be used or purchased be- cause of this NMFS publication. The NMFS Scientific Publications Office is not responsible for the contents of the articles or for the standard of English used in them. 131-143 Callihan, Jody L., Charlton H. Godwin, and Jeffrey A. Bucket Effect of demography on spatial distribution: movement patterns of the Albemarle Sound-Roanoke River stock of Striped Bass iMorone saxatilis) in relation to their recovery 144-158 Manderson, John P., Linda L . Stehlik, Jeff Pessutti, John Rosendale, and Beth Phelan Residence time and habitat duration for predators in a small mid-Atlantic estuary 159-177 Henderson, E. Elizabeth, Karin A. Forney, Jay P. Barlow, John A. Hildebrand, Annie B. Douglas, John Calambokidis, and William J. Sydeman Effects of fluctuations in sea-surface temperature on the occurrence of small cetaceans off Southern California 178-197 Clardy, Samuel D., Nancy J. Brown-Peterson, Mark S. Peterson, and Robert T. Leaf Age, growth, and reproduction of Southern Kingfish (Menticirrhus americanus ): a multivariate comparison with life his- tory patterns in other sciaenids 198-220 Douglas, Annie B., John Calambokidis, Lisa M. Munger, Melissa S. Soldevilla, Megan C. Ferguson, Andrea M. Havron, Dominique L. Camacho, Greg S. Campbell, and John A. Hildebrand Seasonal distribution and abundance of cetaceans off Southern California estimated from CalCOFI cruise data from 2004 to 2008 221-236 Rulifson, Roger A., and Christopher F. Batsavage Population demographics of Hickory Shad (Alosa mediocris) during a period of population growth 237 Errata 238 Announcement of Best Paper Awards for 2013 239-241 Guidelines for authors 101 Effects of ghost fishing on the population of red king crab ( Parali th odes camtschaticus ) in Womens Bay, Kodiak Island, Alaska Email address for contact author: chris.long@noaa.gov Resource Assessment and Conseivation Engineering Division Kodiak Laboratory Alaska Fisheries Science Center National Marine Fisheries Sen/ice, NOAA 301 Research Court Kodiak, Alaska 99615 Abstract— Ghost fishing, the capture and killing of marine organisms by lost or abandoned fishing gear, is a serious ecological and economic problem confronting fisheries. In this study, we quantify the rate of ghost fishing on the population of red king crab (Paralithodes camtschaticus ) in Womens Bay, Kodiak Island, Alaska. From 1991 to 2008, crabs with cara- pace lengths (CLs) from 42 to 162 mm were tagged with acoustic tags and tracked both from a boat at the surface and by divers. Diver observa- tions were used to determine whether a crab molted or died and, in many cases, to determine the cause of death. Of 192 crabs tracked during this study in association with other projects, 13 were killed in ghostfishing gear (12 in ghost crab pots and 1 in a ghost gill net) and 20 were captured in ghost pots and released alive by divers. An additional 13 died of other causes, in- cluding predation by sea otters and an octopus and poaching by humans. We estimate that between 16% and 37% of the population of red king crab with CLs >60 mm in Womens Bay were killed by ghost fishing per year during the period of this study, making ghost fishing a substantial source of mortal- ity. These results indicate that steps to reduce ghost fishing in Womens Bay are warranted. Manuscript submitted 6 March 2013. Manuscript accepted 6 February 2014. Fish. Bull. 112:101-111 (2014). doi:10.7755/FB. 112.2-3.1 The views and opinions expressed or implied in this article are those of the author (or authors) and do not necessarily reflect the position of the National Marine Fisheries Service, NOAA. W. Christopher Long (contact author) Peter A. Cummiskey J. Eric Munk Ghost fishing is the capturing and killing of marine organisms by fish- ing gear that has been lost or aban- doned (Smolowitz, 1978) and is a seri- ous economic and ecological problem in fisheries around the world (Breen, 1990; Laist, 1996; Matsuoka et ah, 2005). Even after they are lost, nets can continue to entangle and kill or- ganisms (Kaiser et ah, 1996; Santos et al., 2003; Baeta et ah, 2009), and fish traps and crustacean pots can continue to attract, trap, and kill tar- get and nontarget species (Stevens et al., 2000; Hebert et al., 2001; Erzini et al., 2008; Ramirez-Rodriguez and Arreguin-Sanchez, 2008), such as reptiles, birds, and mammals (Ha- vens et al., 2008; Good et ah, 2009; 2010). Dead animals in crab pots and fish nets can then act as bait to at- tract even more organisms (Havens et al., 2008). Although in some cases these effects are negligible, in many cases, depending on the type of gear and the environment (Gerrodette et al., 1990; Santos et al., 2003), ghost fishing represents a substantial eco- nomic loss to the fishery (e.g., Breen, 1987). One of the major difficulties in the estimation of the effects of ghost fish- ing is the methods that are typically used. Many studies either recover lost gear and determine the number of organisms caught (e.g., Stevens et al., 2000) or deliberately “lose” and follow the gear over time (Bullimore et al., 2001). In both cases, however, extrapolation of the results to the population level with any certainty is difficult because of the many impor- tant factors that must be estimated on the basis of limited information, including the rate of gear loss, rate of gear decay, and population size of the study organism. In addition, ani- mals that escape may suffer delayed mortality in response to starvation during the captivity period (Paul et al., 1994), and this effect often is un- accounted for (e.g., Breen, 1987). In this study, we took the unique approach of tracking individuals in a population of red king crab (Para- lithodes camtschaticus ) over time and observed their fates by tagging them with acoustic tags and by mak- ing in situ observations during scuba diving. This work allowed us to cal- culate the mortality rate caused by ghost fishing at the population level independent of population size (e.g., Lambert et ah, 2006) and to compare it directly with other causes of mor- tality, such as predation. In addition, rather than focusing on one partic- ular type of gear, this approach in- cluded all types of ghostfishing gear that were catching red king crabs in the study area and allowed compari- sons among these gear. 102 Fishery Bulletin 112(2-3) This study took place in Womens Bay, which is lo- cated in the Gulf of Alaska near the city of Kodiak on Kodiak Island, Alaska, and is a popular site for com- mercial, sport, and subsistence fishing for a variety of finfish and shellfish species (Fig. 1). Red king crab was the major fishery species in the Gulf of Alaska during the 1960s, but populations of this crab crashed in the 1970s and 80s. The commercial fishery was closed in 1983 and has not been reopened (Orensanz et al., 1998). Since that time, red king crab has been harvested in this region only in a subsistence fishery for which the catch limit has been 3 crabs per household per year. In this fishery, only male crabs with a carapace width >178 mm (or -154 mm carapace length [CL]) may be taken legally (Orensanz et al., 1998). Crab pots or traps are used to capture red king crabs in this subsistence fishery. Harvest information has been available since 1995 for the Chiniak Bay area, which covers 321.0 km2 and includes Womens Bay. With an area of 8.5 km2, Womens Bay represents only about 2.5% of Chiniak Bay (Fig. 1). Harvest levels in Chiniak Bay from 1995 to 2012 ranged from 10 to 1178 crabs per year (me- dian=66 crabs per year) (ADFG)1 (Fig 2), which would indicate a low exploitation rate. The population of red king crab in Womens Bay is not surveyed discretely, and no direct estimate of this population is available. However, the Kodiak district as a whole is surveyed by the Alaska Department of Fish and Game during its westward region trawl survey, when data are collected for estimates of the population size of red king crab and southern Tanner crab ( Chion - oecetes bairdi ) in the northeast section of the Kodiak district, which includes Womens Bay (Fig. 1). For that survey -90 tows are conducted, each 1.85 km long, in the northeast section of the Kodiak district with a 400- mesh eastern otter trawl constructed of 8.9-cm mesh in the body and 3.2-cm mesh in the codend to estimate population size by using the area-swept, method (for complete survey methods and design, see Spalinger2). Over the period of 1991-2012, estimates of the size of the population of red king crab in the northeast sec- tion have been low, ranging from about 160,000 to 0 (median=9500 crabs) (Fig. 2); these estimates are not precise because red king crabs typically are caught only at a few stations (Spalinger2). Our study area in Womens Bay accounts for <1% of the northeast section, a total area of 1978.5 km2; therefore, the population in Womens Bay is likely a small proportion of the popula- tion estimated for this region. 1 ADFG (Alaska Department of Fish and Game). 2012. Ko- diak shellfish subsistence database. [Data available upon request from ADFG, 351 Research Ct, Kodiak, AK 99615- 7400.] 2 Spalinger, K. 2009. Bottom trawl survey of crab and groundfish: Kodiak, Chignik, South Peninsula, and Eastern Aleutians Management Districts, 2008. Fishery Manage- ment Report 9-25, 121 p. Alaska Department of Fish and Game, Anchorage, AK. Materials and methods Our study is based on a 17-year data set, from 1991 to 2008, of tracking red king crab in Womens Bay with acoustic tags. Crabs were tracked during other projects where behavior and habitat use by red king crab were examined (Dew et al., 1992; Dew and McConnaughey, 2005; Dew, 2010) and were found primarily in pods, which are aggregations of crabs (Dew, 1990). Red king crabs (identified per Donaldson and Byersdorfer, 2005) were captured by divers throughout Womens Bay, and acoustic tags (Sonotronics; Tucson, AZ3) were affixed to each crab’s carapace with marine grade epoxy. Only ac- tive, healthy-looking crabs were used, and crabs typi- cally were new-shell ( i.e. , they had recently molted). Crabs were released from the surface at the same location where they were captured (Fig. 1). Tags emit- ted unique acoustic sequences, allowing for the iden- tification and tracking of individual crabs, and track- ing was performed from the surface with a Sonotron- ics USR-4D surface acoustic receiver deployed from a boat. Each time a crab was located, the position of the vessel at the sea surface above the crab was recorded with a GPS unit. Accuracy of the boat’s location in re- lation to the crab’s location was generally about 40 m and was dependent on depth and weather. Crabs also were tracked with a Datasonics DPL-275A, underwater acoustic dive receiver (Teledyne Benthos, North Fal- mouth, MA), which allows divers to use acoustic sig- nals to locate tags in situ. Crabs were located on the seafloor by scuba divers, and data were collected on the behavior and habitat of all live, tagged crabs, and any closely aggregated crabs; data included whether crabs were trapped alive in ghostfishing gear. We classified the final condition of all crabs that had been tagged, including crabs no longer attached to their tag, and recorded the status of unrecovered tags in the following manner. When a tag was found on a complete, newly shed carapace, the final condition of the crab was classified as “molted.” When a tag was found attached to a dead or partially eaten crab, final condition of the crab was classified as “dead.” When the condition of the crab could not be determined with any confidence, its condition was classified as “unknown.” Tags that could not be located from the surface were recorded as “lost,” indicating either that the tag had run out of power or had malfunctioned or that the crab had moved into an area where it could not be tracked. Because we tracked all tagged crabs until they died, molted, or became lost, no tagged crabs had their final condition classified as live (see the last paragraph of this section). When a crab was classified as dead, the cause of death was ascertained when possible. Ghostfishing-induced mor- tality (hereafter termed “ghostfishing mortality”) was recorded when that crab was found dead in ghost- 3 Mention of trade names or commercial companies is for iden- tification purposes only and does not imply endorsement by the National Marine Fisheries Service, NOAA. Long et al.: Effects of ghost fishing on the population of Paralithodes ccimtschaticus in Womens Bay, Alaska 103 Figure 1 Map showing study area and sites (A) where tagged red king crabs (Paralithodes camtschaticus) were released in Womens Bay, Kodiak Island, Alaska, during this study in association with other projects over the period of 1991-2008: (A) Alaska and Kodiak Island, (Bi the northeast section of the Kodiak district, as defined for the westward region trawl survey of the Alaska Department of Fish and Game, and (C) Chiniak Bay. The southwestern portion of the bay is a shallow area where the red king crab does not occur, and no crabs were released there. In Inset A, lines are drawn to indicate Kodiak Island and the northeast section (N.E. section). In Inset B, lines show boundaries of the northeast section (adapted from the Fishery Management Report 09-25 by K. Spalinger published in 2009 by the Alaska Department of Fish and Game) and the rectangle around Chiniak Bay shows the extent of Inset C. In Inset C, the rect- angle around Womens Bay shows the extent of the main map fishing gear. On one occasion, 3 crabs released on the same day all died almost immediately for no appar- ent reason, and each of these crabs was recorded as a handling-induced mortality (hereafter “handling mor- tality”). There was no evidence of any latent handling mortality for other crabs in this study. All other causes of death were classified as “other mortality.” One tag became detached from its crab because the epoxy failed to adhere to the carapace and was classified as a “tag- ging failure.” Occasionally, surface tracking located a tag that had stopped moving for a long period of time and was not recovered because either it was in an un- suitable diving location or there were other logistical concerns (e.g., ice cover). Such tags were denominated as “derelict.” Generally, all tagged crabs were tracked and diver observations were made weekly on a sub- set of tags, but effort varied with season, weather, and availability of field personnel. When divers discovered ghost pots, both during this and other projects (effort on other projects var- ied throughout the study period) in Womens Bay, they made notes that included the type of gear, whether it was found intact or upside down, and the approximate number and species of crabs entangled or entrapped in the gear. Documentation of whether the pot was up- side down is important because Dungeness pots have legally mandated biodegradable release mechanisms on their tops (ADFG4), making release ineffective if the 4 ADFG (Alaska Department of Fish and Game). 2011. Shell- fish gear requirements. Accessed September 2011. http:// www. adfg.alaska.gov/ind ex. cfm?adfg= per sonaluseby a reasout heastGear.main. 104 Fishery Bulletin 112(2-3) ' 1200 - 1000 I - 800 < CD - 600 ~ c - 400 | - 200 - 0 1990 1995 2000 2005 2010 Year Figure 2 Population size and harvest levels for red king crab (Paralithodes camtschaticus ) in 1991—2012. Population size (in thousands of indi- viduals) is for the northeast section of the Kodiak district, as defined for the westward region trawl survey of the Alaska Department of Fish and Game, and harvest levels are for Chiniak Bay. The study area, Womens Bay, is part of Chiniak Bay, which is in turn part of the northeast section of Kodiak district (Fig. 1). Data used in this graph came from Alaska Department of Fish and Game reports on bottom trawl surveys of crabs and groundfishes conducted in 1991-2012: Technology Fishery Reports 93-16, 93-17 and Regional Information Report 4K95-1 by D. Urban, and Fishery Management Reports 05-48 and 13-27 by K. Spalinger. pot is flipped over. Field personnel made observations on both ghost and actively fishing pots to determine the likely causes of pot loss in Womens Bay. Divers rou- tinely disabled all ghost pots that they discovered, ren- dering them incapable of catching animals, and they released all trapped animals. We assessed the effects of ghost fishing on the popu- lation of red king crab in Womens Bay by modeling the rate of tag loss. We assumed, on the basis of hundreds of hours of in situ observations during which we did not see any differences in behavior between tagged and untagged crabs, that tagged animals were representa- tive of the population (except in cases of handling mor- tality, which we quantified and explicitly corrected for) and that tagging did not change behavior (Pollock et al., 1991). In particular, in the context of this study, we assumed that tagged crabs did not suffer higher natural or fishing mortality, that they did not molt at a higher rate, and that they were no more likely to en- ter a crab pot than were untagged crabs. Given these assumptions, the mortality rate of the population as a whole would be similar to the calculated mortality rates of tagged crabs (Pollock et al., 1991; Lambert et al., 2006). As part of our determination of mortality rates, we estimated the number of days between the release date and date of final condition (e.g., the day a crab molted or died). If a tag had been located from the surface at the same position several times, then a dive was per- formed to determine the final condition of a crab. In most of these cases, we used the date from the second time the crab was located at that same position as the estimated date of final condition. When the time be- tween the surface locations was greater than a month, the date half-way between them was used. If there was no indication that the crab had remained in the same location for a length of time (fewer than 2 surface ob- servations at the same location), we used the day of the final dive observation as the end date. We fitted the data using maximum likelihood to an exponential loss model assuming a binomial distribu- tion such that P = e-rt, where P = the probability that a tagged crab molted, died, or was lost; /- = the instantaneous loss rate; and t = the time in days. We calculated the loss rate due to each cause (i.e., molting, ghostfishing mortality, handling mortality, or tag detachment, other mortality, and tag malfunction) with the proportional number of crabs in each category. For example, the rate of mortality from ghost fishing was calculated as Long et al.: Effects of ghost fishing on the population of Parahthodes camtschaticus in Womens Bay, Alaska 105 ?’GF = rPGF’ where r^p = the rate of crab loss due to ghost fishing; r = the overall loss rate calculated above; and PGF - the proportion of crabs in this study that died from ghost fishing. We assumed that tagged crabs with a condition clas- sified as “unknown” had either died from causes other than ghost fishing (e.g., fishing mortality and poaching) or had molted. Additionally, we assumed that crab with tags classified as “derelict” had died from causes other than ghost fishing, had molted, or had been killed in ghostfishing gear. Therefore, we divided the crabs with unknown conditions or derelict tags into categories on the basis of the relative observed frequency of each cause of mortality. Tags classified as lost were kept as a separate category because no data on what happened to the tag were available. Divers found 20 tagged crabs and hundreds of closely aggregated crabs alive in intact ghost pots and released them. In most cases, it was not possible to record the numbers of untagged live crabs released by divers because of the large numbers of those crabs and because visibility was greatly limited by silt disturbed from pot handling; therefore, only rough estimates were made. Two of the tagged crabs that were released later died in another crab pot. Given the length of time the crabs were likely in the pots, it is unlikely that many of these crabs would have escaped alive (High and Worlund, 1979). To estimate an upper and lower limit for the mor- tality rate of crabs from ghost fishing in the absence of diver interference, we calculated the loss rates with the assumption that none of the crabs released from pots would have died (conservative estimate) and with the assumption that all of them would have died (up- per estimate). For the conservative estimate, we used the date the crab died, molted, or was lost after having been released as the date of the final condition. For the upper estimate, we classified the crab as a “ghostfish- ing mortality” and used the date the crab was released from the pot as the date of the final condition of the crab, as if the crab had been found dead in the pot. We performed a logistic regression, using sex as a factor and CL as a covariate, to examine whether size or sex of a crab made it more likely that it would be caught or killed in ghostfishing gear. Results The size of tagged crabs ranged from 42 to 162 mm CL (mean=100 mm CL). Of the 192 crabs tagged over the course of multiple studies, 90 were female and 102 were male. The number of crabs tagged each year ranged from 2 (in 2007 and 2008) to 20 (in 1996) and averaged 1 1 crabs per year. Crabs were tracked for an average of 147 days and a maximum of 468. The total number of observations on tagged crabs was 3300. Molting was Table 1 Final conditions determined in this study for the 192 red king crabs ( Paralithodes camtschaticus) or their tags that were tracked by acoustic tagging in associa- tion with other projects over the period of 1991-2008 in Womens Bay, Kodiak Island, Alaska. “Molted” means the tagged crab molted. “Unknown” means the tag was recovered by divers but it could not be determined if the tagged crab had molted or died. “Lost” means the tag could not be located from the surface. “Derelict” means the tagged crab either molted or died as evidenced by the fact the tag stopped moving but the condition was not determined by diving. “Other mortality” means the tagged crab died from a cause other than ghost fishing. “Ghostfishing mortality” means the tagged crab died in ghostfishing gear. “Handling mortality” means the tagged crab died as a result of the tagging process. “Tag failure” means the tag failed to adhere to the tagged crab’s carapace. Final condition Number Percentage Molted 76 39.6 Unknown 48 25.0 Lost 22 11.5 Derelict 16 8.3 Other mortality 13 6.8 Ghostfishing mortality 13 6.8 Handling mortality 3 1.6 Tag failure 1 0.5 the most common final condition of a tagged crab, fol- lowed by unknown condition (Table 1). Three crabs mi- grated out of Womens Bay during the project, but they were tracked and their final condition was determined as was done for all other tagged crabs. Thirteen crabs died in ghostfishing gear, and 13 more crabs had other sources of mortality. Known sources of mortality included predation by sea otters (3 crabs) and an octopus (1 crab) and likely poaching by humans (the tags of 2 crabs below the legal-size limit were returned to researchers with im- plausible stories about how the tags were obtained). Legal fishing did not account for a single mortality of a tagged crab during this study. Of the 13 crabs that died in ghostfishing gear, 12 crabs were caught in ghost pots and 1 crab was found in a ghost gill net. As for all the crabs caught in ghostfishing gear and released alive by divers, they all were found in the same type of gear: ghost pot. Crabs that died in ghostfishing gear ranged from 69 to 160 mm CL and included 4 ovigerous females. Crabs caught and released by divers ranged from 66 to 141 mm CL and included 5 ovigerous females. One difference between the crabs that died and the crabs that were released was the number of days between the time when the crab was caught in the trap (as es- timated from surface tracking) and the time when a 106 Fishery Bulletin 112(2-3) Table 2 Types of ghost pots found in Womens Bay, Kodiak Island, Alaska, by divers over the period of 1991-2008 and classification according to whether a pot was still intact and whether it was found upright or upside down. “Unknown” indicates that the type of pot could not be assessed. “Un- known condition” indicates that the divers could not determine if the pot was intact. Type Number found Number intact Percentage intact (%) No. found upside-down Unknown condition Dungeness 70 46 66 8 2 Webbed 42 30 71 2 2 Home made 20 10 50 2 1 Store bought 7 3 43 1 2 Unknown 4 0 0 0 2 Total 143 89 62 13 9 dive was performed. The amount of time crabs spent in the trap averaged 38 (standard deviation [SD] 23 days) for crabs that died and 10 (SD 8 days) for crabs that were alive when released. The red king crab ac- counted for the majority of the organisms found in crab pots. Rarely found species included sculpins ( Myoxocephalus spp. ), southern Tanner crab, and Dungeness crab ( Metacarcinus magister\ sensu Schram and Ng, 2012). Over the study period in Womens Bay, divers located 143 pots, of which 60 were found during the tracking of tagged crabs and 83 were found during other projects (Table 2). Of these pots, about half were Dungeness crab pots, which have the shape of a squat cylinder and a frame of steel covered with a mesh of stainless steel. Steel-frame pots, which are large, commercial- size pots with steel frames construct- ed in the shape of a pyramid, cone, or rectangle and which are covered in webbing that is usually made of nylon, were the next most frequently encoun- tered type of pot, followed by home- made pots. Only a few sport pots, which are smaller, light-weight pots that are commercially produced and easily re- trievable by hand, and pots of unknown type were found. Of the crab pots that were encountered, 62% were intact, in- dicating that those pots lacked a biode- gradable release and were capable of ghost fishing. Additionally, other than the Dungeness crab pots, most of the pots lacked escape rings. Likely reasons for pot loss, determined on the basis of field observations, included release of a pot after a line was cut by boat pro- pellers; entanglement of a pot in lines that were dragged by commercial barge towing bridals, sinking of a float due to biofouling on the lines, and breakage of a line by ice. The conservative estimate of the overall rate of loss of tagged crabs from all sources, including sources of mortality, was about 10% less than the upper estimate; however, the upper estimate of the predicted mortality rate from ghost fishing was nearly 300% higher than the conservative estimate of the predicted mortality rate from ghost fishing (Table 3). Other sources of loss did not vary substantially between the 2 estimates. Overall annual mortality estimated from tagging data ranged from 40% to 56% for the conservative and upper estimates (Fig. 3). Using the calculated rate of mortal- ity of tagged crabs in Womens Bay and applying it to the whole population of red king crab in this bay, we estimated that ghost fishing killed between 16% and 37% of the population per year, according to our con- servative and upper estimates (Fig. 3). Table 3 Instantaneous loss rates for tagged red king crabs ( Paralithodes camts- cliaticus) and tags in Womens Bay, Kodiak Island, Alaska, during the period of 1991-2008. The conservative estimate for ghostfishing mortal- ity was determined on the basis of counts of crabs that died in ghost- fishing gear. For the upper estimate, we assumed that all crabs caught in ghost-fishing gear and released by divers would have died. The estimate for experimental error includes handling mortality and tags that fell off crabs. The “All” category represents all sources of crab loss and is the parameter fitted by the exponential decay model (see the model in the Materials and methods section of the text for details). Standard errors of the mean are presented in parentheses. Source of loss Loss rate (days-1) Conservative Upper All Molting Ghostfishing mortality Other mortality Lost tag Experimental error 0.00709 (0.00003) 0.00476 (0.00017) 0.00056 (0.00007) 0.00081 (0.00017) 0.00081 (0.00016) 0.00015 (0.00024) 0.00782 (0.00004) 0.00466 (0.0003) 0.00147 (0.00023) 0.00075 (0.00017) 0.00077 (0.00017) 0.00016 (0.00008) Long et al.: Effects of ghost fishing on the population of Paivlithodes camtschaticus in Womens Bay, Alaska 107 Logistic regression revealed that neither size nor sex affected the probability of a tagged crab being killed or caught in ghostfishing gear, although there was a nonsignificant trend (P=0.055; Table 4) for larger (>140 mm CL) tagged crabs than for the smallest (40-60 mm CL) crabs to be killed in ghostfishing gear; however, this trend may be driven by the low number of crabs tagged in the smallest and largest size categories (Ta- ble 4, Fig. 4). Because we did not observe any ghostfish- ing mortality for crabs with CL <60 mm, we restrict our inference about the effect of ghost fishing at the population level to crabs with CL >60 mm. Although our data suggest that the ghostfishing mortality rate for the largest crabs (>140 mm CL; this size category includes legal-size crabs) is higher than our average rate for crabs of all sizes, we take the conservative approach by applying a constant rate of ghostfishing mortality to all crabs >60 mm CL because of our lower sample size for crabs in the largest size category. Discussion The 17-year data set used in our study allowed us to estimate the effect of ghost fishing on a local popula- tion of red king crab. The results of this study indicate that ghost fishing in Womens Bay was responsible for more mortality of red king crabs with CL >60 mm than any other single observed cause of mortality observed during our study. Indeed, our data provide evidence that the rate of mortality from ghost fishing may be almost double the rate from all other sources of mor- tality combined (for caveats to this assertion, see the discussion later in this section on mortality associated with molting). These results indicate that ghost fishing has a large, negative effect on the population of red king crab in Womens Bay and that changes in regu- lations designed to minimize ghost fishing or in their enforcement may be warranted. Although many studies have quantified the effects of ghost fishing, most estimate the number of animals killed per unit of time (e.g., Breen, 1987; Hebert et ah, 2001), the number killed per pot per unit of time (e.g., Bullimore et al., 2001; Al-Masroori et ah, 2004; Camp- bell and Sumpton, 2009), or simply the number of crabs caught per pot (e.g., Stevens et ah, 2000; Havens et ah, 2008). These studies focus on following or recov- ering lost gear and examining catches and mortality over time on a local population. However, in scaling up to estimate effects on larger or commercially targeted populations, assumptions or estimates of the number of pots lost and the population size are required. By following the fates of individuals in the Womens Bay population, we could estimate with precision the ac- tual mortality rate at the population level, assuming tags do not alter crab behavior. A drawback of this ap- proach, however, is that it applies only to Womens Bay and is not easily extrapolated to other areas. Still, this problem is one shared by all studies of ghost fishing. 0 50 100 150 200 250 300 350 Time (days) Figure 3 Predicted percent mortality over a year for red king crab (Paralithodes camtschaticus) in Womens Bay, Ko- diak Island, Alaska. “Overall mortality” is the total mortality from all sources and is the sum of ghostfish- ing mortality and other mortality. “Ghost fishing” is the percentage of crabs that are killed in ghostfishing gear. “Other mortality” includes all other sources of mortal- ity, including predation and fishing. (A) Conservative estimates of mortality; only crabs that died in ghost- fishing gear were included in estimates of ghostfishing mortality. (B) Upper estimates of mortality; all crabs caught in intact ghostfishing gear were included in these estimates of ghostfishing mortality. Our conservative estimate of ghostfishing mortal- ity is precise, and because of the narrow definition for crabs considered to have died in pots, it represents the absolute minimum effect that ghost fishing had on the population of red king crab in Womens Bay dur- ing our study. However, it is likely a large underesti- mate. Because this study was not originally intended to document the effects of ghost fishing, divers active- ly decreased the effects of ghost fishing by releasing trapped crabs and disabling ghost pots. How many of the crabs caught in pots would have died is difficult to determine. Although escape rates for red king crabs from intact commercial red king crab pots may be up 108 Fishery Bulletin 112(2-3) labile 4 Results of logistic regression used to examine effects of size (carapace length) and sex on the likelihood of death or capture in ghostfishing gear for red king crabs ( Paralithodes camtschaticus ) in Womens Bay, Kodiak Island, Alaska, during the period of 1991-2008. Z=Z- score; improbability. “Size” is the carapace length of the crab. Parameter Estimate Standard error Z P Crabs that died in ghostfishing gear Constant -4.69 1.30 -3.61 0.000 Size 0.02 0.01 1.92 0.055 Sex (female) -0.22 0.61 -0.37 0.714 Crabs that were caught in ghostfishing gear Constant -2.57 0.82 -3.12 0.002 Size 0.01 0.01 0.89 0.373 Sex (female) 0.63 0.40 1.58 0.114 0.30 0.25 ■ 0.20 c 0 1 0.15 Q. O a. 0.10 0 05 • 0.00 Killed Caught 42 16 53 10 1 33 36 1 1 e.0 ifr30 60 yOO V2.0 X 6,0* Size range (mm carapace length) Figure 4 Proportion of tagged red king crabs ( Paralithodes camtschaticus) that died (black bars) or were caught (gray bars) in ghostfishing gear in Womens Bay, Ko- diak Island, Alaska, during the period of 1991-2008. Note that no crabs with carapace lengths between 40 and 60 mm were caught or killed. The number above each set of bars represents the total number of crabs in each size category. The line with the long dashes in- dicates the overall proportion of tagged crabs killed in ghostfishing gear, and the line with the short dashes indicates the overall proportion caught in ghostfishing gear. to 90% (High and Worlund, 1979) and mortality of crabs trapped in such pots may be no higher than 17% (Godpy et al., 2003), these rates of escape and mortality do not reflect rates for most of the pots in our study be- cause many of the pots observed in our study targeted the Dungeness crab. Current regulations in Alaska require escape rings on Dungeness crab pots to be 121 mm, smaller than the 159 mm required for red king crab pots (ADFG4), but most of the pots found during our study that were of a type other than the Dunge- ness crab pot did not have escape rings. Therefore, we would expect much lower escape rates for red king crabs caught in the pots observed in this study. Estimates for mortal- ity rates have been much lower for crabs trapped in red king crab pots than for crabs caught in many other pot types: 31-46% mortality of blue swimmer crabs ( Portunus pelagicus) in blue swimmer crab pots (Camp- bell and Sumpton, 2009), 52% of Dungeness crabs in Dungeness crab pots (Breen, 1987), and 95% mortali- ty of snow crabs ( Chionoecetes opilio) in snow crab pots (Hebert et al., 2001). Given the relatively small escape rings of pots de- signed for much smaller crab species and the absence of escape rings in the pots observed in this study, the majority of red king crabs found trapped in pots likely would have been unable to escape and eventually would have died. Results from our study suggest that red king crabs <60 mm CL are less likely to be caught or killed in pots than crabs of other sizes and that crabs >140 mm CL are more likely than crabs <140 mm to become trapped or to die in pots; however, these findings were not statistically significant and may have been driven to some extent by small sample sizes for crabs in the largest and smallest size categories. Although small- er crabs probably are more likely to escape from pots (High and Worlund, 1979), our data indicate that crabs >60 mm CL are vulnerable to the types of ghost pots observed in this study. Escape of red king crabs from pots is asymptotic over time, and the number of crabs escaping levels off at about 8 days (High and Worlund, 1979). We esti- mated that only 8 of the 20 crabs released from pots had been trapped for less than 8 days. If the remaining 12 crabs had been able to escape, then they probably would have done so before they were rescued. Addi- tionally, divers in this study actively disabled 89 intact pots that would otherwise have continued to ghost fish, likely substantially lowering the effect of ghost fishing in Womens Bay. Given these observations, we believe that the true mortality rate from ghostfishing gear is somewhere between our conservative and upper esti- mates and is most likely closer to the upper estimate. Although our estimate of the rate of mortality of red king crabs from sources other than ghost fishing is rea- sonably precise, it is almost certainly an underestimate because it does not account for mortalities that may occur during molting, which is physiologically stressful Long et al.: Effects of ghost fishing on the population of Parcilithodes camtschaticus in Womens Bay, Alaska 109 for crustaceans (Leffler, 1972), or shortly after molting, when they are particularly vulnerable to predation (Shirley et ah, 1990; Ryer et ah, 1997; Marshall et ah, 2005). However, for the intermolt period of red king crab, our estimate is likely an accurate picture of mor- tality. Roughly a third of all crabs that died suffered predation, a sixth of them were poached, and the cause of death could not be determined for the remaining half. Ironically, no tagged male crabs of legal size were taken by fishermen in this study; no doubt, zero fishing mortality was due in part to the fact that few legal-size crabs were tagged and to the low fishing mortality for this species in the study area (Fig. 2). During another study in Bristol Bay, instantaneous mortality rates es- timated for the red king crab ranged from 0.02 to 1.75 yearn1 and most estimates ranged between 0.02 and 1.00 year-1 (reviewed in Zheng, 2005). Our estimates for rates of mortality from sources other than ghost fishing were 0.30 and 0.27 year-1, values that fall well within that range. The estimated mortality rate from ghost fishing is high enough to have a devastating effect on the popu- lation of red king crab in Womens Bay. A 60-mm-CL crab is at least 2 years from attaining reproductive maturity (Weber, 1967), and females must brood their eggs for a year after attaining maturity before they can reproduce successfully for the first time (Stevens and Swiney, 2007). Therefore, on the basis of our conser- vative and upper estimates of ghostfishing mortality (16-37% per year), we estimate, with the assumption that crabs become vulnerable to ghost fishing at a size of 60 mm CL, that 29-60% of male crabs and 41-75% of female crabs were killed in ghostfishing gear before they were able to reproduce for the first time during our study. To put those high mortality rates in context, the tar- get rate for fishing mortality designed to maintain a healthy stock size in Bristol Bay is <15% of the mature male biomass and only negligible numbers of nontar- geted mature female and immature crabs are allowed to be taken as bycatch (Zheng and Siddeek5). Because the fecundity of females increases more than an order of magnitude with crab size (Swiney et ah, 2012), ghost fishing keeps many females from reaching their full reproductive potential by killing them when they are small and have relatively low fecundity; moreover, be- cause ghost fishing indiscriminately removes both im- mature and mature crabs, including ovigerous females, it compounds its effects as it reduces the reproductive capacity of the local population, as well as the size of the local population itself. 5 Zheng , J., and M. S. M. Siddeek. 2010. Bristol Bay red king crab stock assessment in spring 2010. In Stock assess- ment and fishery evaluation report for the king and Tanner crab fisheries of the Bering Sea and Aleutian Islands regions. 2010 Crab SAFE, p. 135-246. North Pacific Fishery Man- agement Council, Anchorage, AK. [Available from http:// www.npfmc.org/wp-content/PDFdocuments/resources/SAFE/ CrabSAFE/CRABSAFE2010.pdf.l How typical Womens Bay may be among areas in the Gulf of Alaska is unknown. Its proximity to the city of Kodiak makes it a popular site for sport, sub- sistence, and commercial fisheries. The greater fishing effort and boat traffic in this bay, compared with such activities in other areas, has likely led to a higher rate of fishing gear loss. Additionally, participants in the sport and subsistence fisheries may be less likely to know about or comply with the requirements for escape mechanisms on pots. Supporting this premise, home- made pots found in our study were almost certainly not used for commercial purposes, and their structure was frequently noncompliant with established require- ments for escape rings and biodegradable release. En- forcement of regulations in the sport and subsistence fisheries also probably is less stringent than it is in the commercial fishery because there are far fewer com- mercial fishermen to monitor (who fish with numerous pots per boat) than subsistence fishermen (who fish with few pots per boat). It is likely that other coastal bodies of water with high densities of crabs near pop- ulation centers in Alaska have similar rates of ghost fishing and that bodies of water farther from human population centers have a lower rate. If ghost fishing does have the profound effect that is indicated by our data on the population of red king crab in Womens Bay, measures to reduce ghost fish- ing are warranted. Of the different types of gear, crab pots were the major cause of ghostfishing mortality; in contrast, gill nets were responsible for only one death. Therefore, efforts to reduce ghost fishing on the red king crab should focus on pots (although remov- ing nets may be a priority for other species, such as marine birds [Good et al., 2009; 2010]). Existing ghost pots can be located by side-scan sonar and removed by grappling (Stevens et al., 2000) and their threat elimi- nated as a result. Removal of ghost pots would be most effective in shallow areas of high fishing intensity, such as Womens Bay. The observed effects of ice deserve a special note. Womens Bay frequently had a fresh water lens from various freshwater sources (Long, 1972) that can freeze during the winter. Ice embedded pot floats and when the ice broke up, pots were dragged into deeper water, where their float lines were not long enough to reach the surface, or they were dragged into shallow water, where pot owners were not likely to look for them. The strain of being dragged across the bottom of the bay could have caused lines to break. Additionally, tides or wind moved thin sheets of ice that abraded floats and lines and caused them to sink. Ice and boats dragged pots across the bottom of the bay, flipping some of them upside down or partially burying them — outcomes that, in the case of Dungeness crab pots, rendered the es- cape mechanism, if present, ineffective. If areas af- fected by ice were closed during months when ice was a concern, the rate of pot loss could be reduced. This closure would not substantially affect fishing because crab pots cannot be checked during times of ice cover. no Fishery Bulletin 112(2-3) Such closures also would reduce the number of pots abandoned and left out long enough to have their floats sunk by marine-fouling organisms. Current regulations allow the tie-down of a pot lid to be equipped with cotton twine that will decay and release the lid after a length of time (ADFG4). How- ever, we observed that pots can be flipped upside down when lost, rendering this escape mechanism unwork- able. If Dungeness crab pots were required to have a sidewall release as an escape mechanism, this require- ment would greatly reduce this problem. To reveal fur- ther means of reducing ghost fishing, more experimen- tal work is needed to examine the effectiveness of vari- ous release mechanisms for allowing target species and other species to escape pots (Matsuoka et al., 2005). Conclusions Ghost fishing is an entirely wasteful source of mortal- ity in aquatic systems. Although lost fishing gear is an inevitable consequence of fishing, it can become a substantial drain on both fished and nonfished species. In Womens Bay, ghost fishing, primarily by crab pots, is a source of mortality that may have a strong nega- tive effect upon the population by killing 16-37% of the population per year. This negative effect could be decreased through implementation of measures to re- duce the loss rate of crab pots and to ensure that pots do not continue to fish long after they are lost. Acknowledgments We thank S. Persselin, B. Dew, S. Payne, and B. Ste- vens for help in tracking and diving on tagged crabs. This article was improved through comments by B. Knoth, J. Long, F. Morado, R. Foy, and 3 anonymous reviewers. Literature cited Al-Masroori, IT., H. Al-Oufi, J. L. Mcllwain, and E. McLean. 2004. Catches of lost fish traps (ghost fishing) from fish- ing grounds near Muscat, Sultanate of Oman. Fish. Res. 69:407-414. Baeta, F., M. J. Costa, and H. Cabral. 2009. Trammel nets’ ghost fishing off the Portuguese central coast. Fish. Res. 98:33-39. Breen, P. A. 1987. Mortality of Dungeness crabs caused by lost traps in the Fraser River Estuary, British Columbia. N. Am. •J. Fish. Manage. 7:429-435. 1990. A review of ghost fishing by traps and gillnets. In Proceedings of the second international conference on marine debris; Honolulu, 2-7 April 1989 (R. A. Shomu- ra and M. L. Godfrey, eds.), p. 571-599. NOAA Tech. Memo. NOAA-TMSWFSC-154. Bullimore, B. A., P. B. Newman, M. J. Kaiser, S. E. Gilbert, and K. M. Lock. 2001. A study of catches in a fleet of “ghost-fishing” pots. Fish. Bull. 99:247-253. Campbell, M. J., and W. D. Sumpton. 2009. Ghost fishing in the pot fishery for blue swim- mer crabs Portunus pelagicus in Queensland, Austra- lia. Fish. Res. 95:246-253. Dew, C. B. 1990. Behavioral ecology of podding red king crab, Paralithodes camtschatica. Can. J. Fish. Aquat. Sci. 47:1944-1958. 2010. Podding behavior of adult king crab and its effect on abundance-estimate precision. In Biology and man- agement of exploited crab populations under climate change (Kruse, G. H., G. L. Eckert, R. J. Foy, R. N. Lip- cius, B. Sainte-Marie, D. L. Stram, and D. Woodby, eds.), p. 129-151. Alaska Sea Grant College Program, Univ. Alaska Fairbanks, Fairbanks, AK. Dew, C. B., and R. A. McConnaughey. 2005. Did trawling on the brood stock contribute to the collapse of Alaska’s king crab? Ecol. Appl. 15:919-941. Dew, C. B., P A. Cummiskey, and J. E. Munk. 1992. The behavioral ecology and spatial distribution of red king crab and other target species: implications for sampling design and data treatment. In Proceedings of the international crab rehabilitation and enhancement symposium; Kodiak, AK (L. White and C. Nielsen, eds.), p. 39-67. Fisheries Rehabilitation, Enhancement and Development Division, Alaska Department of Fish and Game, Juneau, AK. Donaldson, W., and S. Byersdorfer. 2005. Biological field techniques for lithodid crabs. Alas- ka Sea Grant College Program Report AK-SG-05-03, 82 p. Univ. Alaska Fairbanks, Fairbanks, AK. Erzini, K., L. Bentes, R. Coelho, P. G. Lino, P. Monteiro, J. Ribeiro, and J. M. S. Goncalves. 2008. Catches in ghost-fishing octopus and fish traps in the northeastern Atlantic Ocean (Algarve, Portu- gal). Fish. Bull. 106:321-327. Gerrodette, T., B. K. Choy, and L. M. Hiruki. 1990. An experimental study of derelict gillnet frag- ments in the central Pacific Ocean. In Proceedings of the second international conference on marine de- bris; Honolulu, 2-7 April 1989 (R. A. Shomura and M. L. Godfrey, eds.), p. 600-614. NOAA Tech. Memo. NOAA-TMSWFSC-154. Godoy, H., D. M. Furevik, and S. Stiansen. 2003. Unaccounted mortality of red king crab (Para- lithodes camtschaticus) in deliberately lost pots off northern Norway. Fish. Res. 64:171-177. Good, T. P, J. A. June, M. A. Etnier, and G. Broadhurst. 2009. Ghosts of the Salish Sea: threats to marine birds in Puget Sound and the Northwest Straits from derelict fishing gear. Mar. Ornithol. 37:67-76. 2010. Derelict fishing nets in Puget Sound and the Northwest Straits: patterns and threats to marine fau- na. Mar. Pollut. Bull. 60:39-50. Havens, K. J., D. M. Bilkovic, D. Stanhope, K. Angstadt, and C. IJershner. 2008. The effects of derelict blue crab traps on marine organisms in the lower York River, Virginia. N. Am. J. Fish. Manage. 28:1194-1200. Long et al.: Effects of ghost fishing on the population of Paralithodes camtschaticus in Womens Bay, Alaska 111 Hebert, M., G. Miron, M. Moriyasu, R. Vienneau, and P. DeGrace. 2001. Efficiency and ghost fishing of snow crab ( Chion - oecetes opilio ) traps in the Gulf of St. Lawrence. Fish. Res. 52:143-153. High, W. L., and D. D. Worlund. 1979. Escape of king crab, Paralithodes camtschatica, from derelict pots. NOAA Tech. Rep. NMFS SSRF-734, 11 p. Kaiser, M. J., B. Bullimore, P. Newman, K. Lock, and S. Gilbert. 1996. Catches in ‘ghost fishing’ set nets. Mar. Ecol. Prog. Ser. 145:11-16. Laist, D. W. 1996. Marine debris entanglement and ghost fishing: a cryptic and significant type of bycatch? In Solving bycatch: considerations for today and tomorrow, p. 33- 39. Proceedings of the solving bycatch workshop; Se- attle, 25-27 September 1995. Alaska Sea Grant College Program Report AK-SG-96-03. Univ. Alaska Fairbanks, Fairbanks, AK. Lambert, D. M., J. M. Hoenig, and R. N. Lipcius. 2006. Tag return estimation of annual and semiannual survival rates of adult female blue crabs. Trans. Am. Fish. Soc. 135:1592-1603. Leffler, C. 1972. Some effects of temperature on the growth and metabolic rate of juvenile blue crabs, Callinectes sapi- dus, in the laboratory. Mar. Biol. 14:104-110. Long, E. R. 1972. Studies of marine fouling and boring off Kodiak Island, Alaska. Mar. Biol. 14:52-57. Marshall, S., K. Warburton, B. Paterson, and D. Mann. 2005. Cannibalism in juvenile blue-swimmer crabs Por- tunus pelagicus (Linnaeus, 1766): effects of body size, moult stage and refuge availability. Appl. Anim. Behav. Sci. 90:65-82. Matsuoka, T., T. Nakashima, and N. Nagasawa. 2005. A review of ghost, fishing: scientific approaches to evaluation and solutions. Fish. Sci. 71:691-702. Orensanz, J., J. Armstrong, D. Armstrong, and R. Hilborn. 1998. Crustacean resources are vulnerable to serial depletion — the multifaceted decline of crab and shrimp fisheries in the Greater Gulf of Alaska. Rev. Fish Biol. Fish. 8:117-176. Paul, J., A. Paul, and A. Kimker. 1994. Compensatory feeding capacity of 2 brachyuran crabs, Tanner and Dungeness, after starvation periods like those encountered in pots. Alaska Fish. Res. Bull. 1:184-187. Pollock, K. H., J. M. Hoenig, and C. M. Jones. 1991. Estimation of fishing and natural mortality when a tagging study is combined with a creel survey or port sampling. In Creel and angler surveys in fisheries management (D. Guthrie, J. M. Hoenig, M. Holliday, C. M. Jones, M. J. Mills, S. A. Moberly, K. H. Pollock, and D. R. Talhelm, eds.), p. 423-434. Am. Fish. Soc. Symp. 12, Bethesda, MD. Ramirez-Rodriguez, M., and F. Arreguin-Sanchez. 2008. Fishing time and trap ghost fishing for Cancer johngarthi along the Baja California Peninsula’s south- western coast, Mexico. J. Shellfish Res. 27:1265-1269. Ryer, C. H., J. van Montfrans, and K. E. Moody. 1997. Cannibalism, refugia and the molting blue crab. Mar. Ecol. Prog. Ser. 147:77-85. Santos, M. N., H. J. Saldanha, M. B. Caspar, and C. C. Monteiro. 2003. Hake ( Merluccius merluccius L., 1758) ghost fishing by gill nets off the Algarve (southern Portu- gal). Fish. Res. 64:119-128. Schram, F. R., and P. K. L. Ng. 2012. What is Cancer ? J. Crust. Biol. 32:665-672. Shirley, M. A., A. H. Hines, and T. G. Wolcott. 1990. Adaptive significance of habitat selection by molt- ing adult blue crabs Callinectes sapidus (Rathbun) within a subestuary of central Chesapeake Bay. J. Exp. Mar. Biol. Ecol. 140:107-119. Smolowitz, R. 1978. Trap design and ghost fishing: an overview. Mar. Fish. Rev. 40:2-8. Stevens, B. G., and K. M. Swiney. 2007. Hatch timing, incubation period, and reproduc- tive cycle for captive primiparous and multiparous red king crab, Paralithodes camtschaticus. J. Crust. Biol. 27:37-48. Stevens, B. G., I. Vining, S. Byersdorfer, and W. Donaldson. 2000. Ghost fishing by Tanner crab (Chionoecetes bairdi) pots off Kodiak, Alaska: pot density and catch per trap as determined from sidescan sonar and pot recovery data. Fish. Bull. 98:389-399. Swiney, K. M., W. C. Long, G. L. Eckert, and G. H. Kruse. 2012. Red king crab, Paralithodes camtschaticus, size- fecundity relationship, and interannual and seasonal variability in fecundity. J. Shellfish Res. 31:925-933. Weber, D. D. 1967. Growth of the immature king crab Paralithodes camtschatica (Tilesius). Bull. Int. North Pac. Fish. Comm. 21:21-53. Zheng, J. 2005. A review of natural mortality estimation for crab stocks: data-limited for every stock? In Fisheries as- sessment and management in data-limited situations (G. H. Kruse, V. F. Gallucci, D. E. Hay, R. I. Perry, R. M. Peterman, T. C. Shirley, P. D. Spencer, B. Wilson, and D. Woodby, eds.), p. 595-612. Alaska Sea Grant College Program Report AK-SG-05-02, Univ. Alaska Fairbanks, Fairbanks, AK. 112 Abstract— Chinook Salmon ( Oncorhyn - chus tshawytsclia) is an economically and ecologically important species, and populations from the west coast of North America are a major compo- nent of fisheries in the North Pacific Ocean. The anadromous life history strategy of this species generates populations (or stocks) that typically are differentiated from neighboring populations. In many cases, it is de- sirable to discern the stock of origin of an individual fish or the stock com- position of a mixed sample to monitor the stock-specific effects of anthropo- genic impacts and alter management strategies accordingly. Genetic stock identification (GSI) provides such dis- crimination, and we describe here a novel GSI baseline composed of geno- types from more than 8000 individual fish from 69 distinct populations at 96 single nucleotide polymorphism (SNP) loci. The populations included in this baseline represent the likely sources for more than 99% of the salmon en- countered in ocean fisheries of Cali- fornia and Oregon. This new genetic baseline permits GSI with the use of rapid and cost-effective SNP genotyp- ing, and power analyses indicate that it provides very accurate identifica- tion of important stocks of Chinook Salmon. In an ocean fishery sample, GST assignments of more than 1000 fish, with our baseline, were highly concordant (98.95%) at the reporting unit level with information from the physical tags recovered from the same fish. This SNP baseline represents an important advance in the technologies available to managers and researchers of this species. Manuscript submitted 13 February 2013. Manuscript accepted 10 February 2014. Fish. Bull. 112:112-130 (2014). doi:10.7755/FB. 112.2-3.2 The views and opinions expressed or implied in this article are those of the author (or authors) and do not necessarily reflect the position of the National Marine Fisheries Service, NOAA. Evaluation of a single nucleotide polymorphism baseline for genetic stock identification of Chinook Salmon ( Oncorhynchus tshawytscha ) in the California Current large marine ecosystem Anthony J. Clemento Eric D. Crandall John Carlos Garza Eric C. Anderson (contact author) Email address for contact author: eric.anderson@noaa.gov Fisheries Ecology Division Southwest Fisheries Science Center National Marine Fisheries Service, NOAA 1 10 Shaffer Road Santa Cruz, California 95060 Institute of Marine Sciences University of California, Santa Cruz 1 10 Shaffer Road Santa Cruz, California 95060 Chinook Salmon ( Oncorhynchus tshawytscha) are found in rivers from central California around the North Pacific Rim and the Bering Sea to Russia and are the target of valuable commercial and recreation- al fisheries. A key aspect of the life history of Chinook Salmon is natal homing, whereby each fish of this anadromous species typically returns to spawn in the same river in which it was born. This homing generates populations (or stocks) that may be genetically differentiated from neigh- boring populations and can exhibit local adaption (Utter et ah, 1989; Taylor, 1991). Recent population de- clines, particularly at the southern end of the native range of this spe- cies, have resulted in the listing of many stocks under the U.S. Endan- gered Species Act (ESA; Federal Reg- ister, 1990, 1999) and have highlight- ed the need to refine the manage- ment and conservation of Chinook Salmon. However, such refinements are challenging because the migra- tory life history of salmon means that the many effects from anthropo- genic sources that occur in rivers or in the ocean (e.g., fisheries, water di- version, or turbine entrainment) may affect multiple, intermingled stocks. In such cases, it may be necessary to discern the stock of origin of affected fish to monitor stock-specific impacts and design management strategies accordingly. The use of pre-existing biological markers to distinguish salmon stocks has a long history. The traits used in these efforts have included mor- phometric and meristic characters (Fournier et ah, 1984; Claytor and MacCrimmon, 1988), scale patterns (Cook, 1982), parasite assemblages (Boyce et al., 1985), and stable iso- tope ratios (Barnett-Johnson et ah, 2008). However, the most universally applicable methods have involved the use of genetic markers because every fish has a unique genetic makeup. The first genetic markers widely used for identification in salmon were electrophoretically detectable protein polymorphisms known as al- lozymes (Milner et ah, 1985; Shak- lee and Phelps, 1990; Tessier et ah, 1995; Allendorf and Seeb, 2000). With the advent of polymerase chain reaction (PCR), many more types of genetic markers became available to Clemento et al.: Evaluation of a single nucleotide polymorphism baseline for genetic stock identification of Oncorhynchus tshawytscha 113 discriminate salmon populations, including mitochon- drial DNA polymorphisms (Cronin et ah, 1993), minis- atellites (Beacham et ah, 1996; Miller et ah, 1996), microsatellites (Seeb et ah, 2007; Moran et ah, 2013), amplified-fragment length polymorphisms (Flannery et ah, 2007) and, most recently, single nucleotide polymor- phisms (SNPs; Smith et ah, 2005a, 2005b; Aguilar and Garza, 2008; Narum et ah 2008; Abach'a-Cardoso et ah, 2011; Clemento et ah, 2011). Genetic stock identification (GSI) typically proceeds in 2 steps. First, samples are collected from potential source populations and genotyped with a set of genetic markers in order to estimate population allele frequen- cies. These genotypes are called the “baseline.” Then, data from individuals sampled from a mixed-stock col- lection (called a “mixture”) and genotyped with the same set of genetic markers are compared with the baseline to estimate the relative proportions of indi- viduals that came from each of the represented source populations. Single individuals of unknown origin also can be assigned to specific populations. Maximum like- lihood or Bayesian methods typically are used to carry out GSI inference (Smouse et ah, 1990; Pella and Ma- suda, 2000). For the first large-scale baseline for GSI of Chinook Salmon allozyme markers were used (Teel et ah1), but technical and logistical issues limited their future ap- peal. The allozyme database was supplanted in Canada by a microsatellite baseline developed by the Depart- ment of Fisheries and Oceans (Beacham et ah, 2006) and more broadly by a microsatellite baseline database developed through a large, international collaboration (Seeb et ah, 2007). This collaboration required enor- mous effort to standardize data across laboratories because microsatellite allele names and sizes usually are not consistent between different laboratories and genotyping equipment. The Seeb et ah (2007) microsatellite baseline has been an effective tool for GSI but has a number of dis- advantages: genotyping and scoring of microsatellites is labor-intensive; genotyping error rates can be relatively high, making the 13 microsatellites in that baseline in- adequate for applications such as pedigree reconstruc- tion (Anderson and Garza, 2006; Garza and Anderson2, Abadia-Cardoso et ah, 2013); missing data rates also 1 Teel, D. J., P. A. Crane, C. M. Guthrie III, A. R. Marshall, D. M. Van Doornik, W. Templin, N. V. Varnavskaya, and L. W. Seeb. 1999. Comprehensive allozyme database discrimi- nates Chinook salmon around the Pacific Rim. NPAFC docu- ment 440, 25 p. [Available from Alaska Department of Fish and Game, Division of Commercial Fisheries, 333 Raspberry Rd., Anchorage, AK 99518.] 2 Garza, J. C., and E. C. Anderson. 2007. Large scale parent- age inference as an alternative to coded-wire tags for salmon fishery management. In PSC genetic stock identification workshop: Logistics Workgroup final report and recom- mendations; Portland, OR, 15-17 May 2007 and Vancouver, Canada, 11-13 September 2007, p. 48-55 p. [Available from Pacific Salmon Commission, 600-1155 Robson St., Vancouver, BC V6E 1B5, Canada.] can be quite high; and, finally, any new laboratory that wishes to use that baseline must undertake a costly standardization process. Additionally, it now has been demonstrated that SNPs, despite typically having only 2 alleles per locus, do have sufficient power to be em- ployed successfully in a GSI context with a modest num- ber of genetic markers (Smith et ah, 2007; Narum et ah, 2008; Templin et ah, 2011; Larson et ah, 2013). Early simulation studies indicated that the bi-allel- ic nature of SNPs would make them less useful than highly polymorphic microsatellites for population dis- crimination (Bernatchez and Duchesne, 2000; Kalin- owski, 2004). However, SNPs are located throughout the genome and may be discovered in genetic regions with higher than average divergence (Nosil et ah, 2009), increasing their utility for GSI. Moreover, SNPs do not suffer from many of the disadvantages of us- ing microsatellites: SNP markers are amenable to the automated, high-throughput genotyping required for large projects; SNP genotyping error rates are very low, making them suitable for pedigree reconstruction; and, importantly, SNP assays typically do not require stan- dardization between labs and, therefore, a SNP base- line is immediately useful to any group or agency that genotypes a mixture sample with the markers used in that baseline (Seeb et ah, 2011). Here, we describe the development and evaluation of a new baseline of SNP marker data for Chinook Salm- on in the southern part of their native range for use in ecological investigation in the California Current large marine ecosystem (and its tributaries) and in fisheries managed by the Pacific Fishery Management Council (PFMC). We introduce a panel of 96 SNP markers and a baseline of more than 8000 salmon from 68 populations of Chinook Salmon ranging from California to Alaska and a single collection of Coho Salmon (Oncorhynchus kisutch) from California. We describe the procedures used to select these SNP markers from among a larger number of candidates and document the resulting pat- terns of genetic differentiation between various popu- lations. We evaluate the power of this new baseline for GSI by both self-assignment (genetic identification of the most likely population of origin) and simulated mixture analyses, focusing on stocks commonly encoun- tered in PFMC fisheries. Finally, we analyze 2090 fish sampled in 2010 from the sport and commercial fisher- ies off the coast of California and compare the results of these analyses with the coded wire tag (CWT) data from these fish to demonstrate the effectiveness of this baseline for classifying individuals to specific manage- ment units. Materials and methods Baseline populations Populations were selected for inclusion in the new baseline to provide broad geographic coverage across 114 Fishery Bulletin 112(2-3) the range of Chinook Salmon in the the United States from Washington to California, while also allowing for the identification of fish from elsewhere in the geo- graphic range of this species. Adult fish were sampled on spawning grounds, in terminal fisheries, or at hatch- eries during the period of 2003-13 and were provided by numerous contributors (see the Acknowledgments section and Warheit et al.3). We included populations expected to be encountered in ocean fisheries off Cali- fornia and Oregon, as well as populations with special management status (e.g., ESA-listed populations). Ac- cordingly, the major lineages of Chinook Salmon from California and Oregon were emphasized in this base- line, as were populations distinguished by life histo- ry strategy (e.g., spring-run, fall-run, and winter-run strategies), but representatives of the major lineages from farther north also were included. DNA was extracted from samples for California populations with DNeasy Blood & Tissue Kits on a BioRobot 30004 platform (QIAGEN, Inc., Valencia, CA) according to the manufacturer’s protocols, and DNA from populations in Oregon, Washington, Canada, and Alaska was extracted by contributors ( see Acknowledg- ments section) who used various methods. Sample sizes ranged from 44 to 1409 individuals per population and averaged 116 individuals per population. The 1409 fish from the population in the Trinity River Hatchery ini- tially were genotyped with our SNP panel for another purpose, but they were included here in total to provide a comprehensive reference sample for identification of this important group. Excluding this disproportionately large sample, the average number of individuals per population was 97. In total, the new baseline includ- ed 7984 Chinook Salmon from 68 distinct populations (Table 1). Each population in this baseline belongs to a single reporting unit, a designation established in previous GSI research that reflects a combination of “genetic similarity, geographic features, and management appli- cations” (Seeb et al., 2007). Reporting units generally are composed of multiple populations that share ge- netic similarity or are subject to similar management regimes. The 68 populations of Chinook Salmon in our baseline fall into 38 distinct reporting units (Table 1), and some reporting units in Alaska and Canada are represented by only a single population. Coho Salmon occasionally are misidentified as Chi- nook Salmon in ocean fisheries and in ecological sam- pling. We included a collection of 47 Coho Salmon from California as the 69th population in our baseline to 3 Warheit, K. I., L. W. Seeb, W. D. Templin, and J. E. Seeb. 2013. Moving GSI into the next decade: SNP coordination for Pacific Salmon Treaty fisheries. FPT 13-09, 47 p. lAvail- able from Washington Department of Fish and Wildlife, 600 Capitol Way N., Olympia, WA 98501-1091.] 4 Mention of trade names or commercial companies is for iden- tification purposes only and does not imply endorsement by the National Marine Fisheries Service, NOAA. help us to identify Coho Salmon that have been identi- fied incorrectly as Chinook Salmon. Markers and genotyping We compiled a list of 192 TaqMan (Life Technologies Corp., Carlsbad, CA), or 5’-nuclease, SNP genotyp- ing assays from previously published discovery stud- ies (Smith et al., 2005a, 2005b; Campbell and Narum, 2008; Narum et al., 2008; Clemento et al., 2011) to test their scorability and power for GSI. TaqMan technol- ogy combines standard PCR primers that target the genomic region around a SNP with 2 different fluores- cent probes that identify the 2 nucleotide bases present at the SNP. As recommended by the manufacturer, we used a multiplex preamplification reaction to increase the copy number of targeted genomic regions. Multi- plex PCR products were diluted with 15 pL of 2 mM Tris buffer and were frozen. Samples then were genotyped on 96.96 Dynamic Ar- rays with an EP1 System (Fluidigm Corp., South San Francisco, CA) according to the manufacturer’s proto- cols. Fluidigm Dynamic Arrays use integrated nanoflu- idic circuitry to simultaneously determine the genotype at 96 SNP loci for 96 samples (2 of which are no-DNA template controls). Genotypes were determined with the Fluidigm Genotyping Analysis software (vers. 2.1.1). The use of quantitative PCR methods for geno- type determination involves discerning, on a 2-D graph, clusters of fluorescence intensity of the probes for the 2 alleles; the 2 homozygote clusters have fluorescence primarily from only 1 probe, but a heterozygote cluster has similar intensities from both probes. Marker selection We selected a panel of 95 SNP markers from among the 192 candidates, reserving 1 marker for a species identification assay (see final paragraph of this sec- tion). The risk of “high-grading bias” (i.e., wrongly in- flating the apparent resolving power of a group of loci for GSI) is particularly great when selecting a panel of markers to distinguish between populations that are closely related, as many of the populations in our baseline are. To avoid high-grading bias, we employed the “training-holdout-leave-one-out” (THL) procedure of Anderson (2010); this procedure requires that data be split into training and holdout sets. Training-set geno- types are used to select the loci included in a baseline and can be included in the eventual baseline, but they are not used to evaluate its performance. Rather, per- formance of a baseline is determined with simulation and self-assignment with only the holdout set, which was not used in any way to select baseline loci. We chose a training set of 372 individuals drawn from 22 populations ( 14 from California, 3 from Oregon, 3 from Washington, 1 from British Columbia, and 1 from Alas- ka) for initial genotyping with all 192 loci. Clemento et al.: Evaluation of a single nucleotide polymorphism baseline for genetic stock identification of Oncorhynchus tshawytscha 115 116 Fishery Bulletin 112(2-3) Clemento et al.: Evaluation of a single nucleotide polymorphism baseline for genetic stock identification of Oncorhynchus tshawytscha 117 118 Fishery Bulletin 112(2-3) For each locus, k, the observed relative frequencies, Pik and qqk, of the 2 SNP alleles were calculated for each population, i, in the training set. These values then were used to compute the expected probability of misassignment, P(MiSjjk), between every pair of popu- lations i and y with only a single locus k: P(Misjjk) = 0.5 [§(p!kSpjk)pik2 + S(Pik<7ik^Pjk<7jk)2pik9ik + §(qjk)0.2 would be required to achieve the necessary statistical power to assign parentage with sufficiently low false-negative and false-positive rates (Anderson and Garza, 2006). However, the observed MAFs for many loci were in fact >0.2 (and as high as 0.5), indicating that the desired statistical power could be achieved with fewer loci. Therefore, we selected the 70 loci with the highest MAF in the Feather River population, the primary target for subsequent parent- age investigations. We then used the P(Misjjk) rank- ings to select 25 additional loci that were useful for distinguishing between difficult-to-resolve populations and reporting units. Finally, an assay to discriminate between Chinook and Coho salmon was included as the 96th assay for genotyping with the Fluidigm 96.96 Dy- namic Arrays. Population genetics analyses The 7669 samples that were not in the training set for locus selection were genotyped with the final panel of 96 SNPs and used as the holdout set in subsequent power analyses (see the next section). This holdout set also was used for standard population genetics analyses. We tested each locus-population pair for de- viations from Hardy-Weinberg equilibrium (HWE) with the complete enumeration method (Louis and Demp- ster, 1987) in GENEPOP software, vers. 4.0 (Rousset, 2008). Similarly, in each population, all pairwise locus combinations were investigated for LD. Default Markov chain parameters were used, except for the number of batches, which was increased to 500 to reduce the stan- dard error to acceptable levels (<0.02; Rousset, 2008). Genetic differentiation (Fst) was estimated (with 0 of Weir and Cockerham, 1984) between all pairs of populations with the software package GENETIX, vers. 4.05 (Belkhir5). The data set was permuted 1000 times to determine the significance of F sx estimates. Phylo- geographic trees were constructed with the chord dis- tance (DCE) of Cavalli-Sforza and Edwards (1967) and the neighbor-joining algorithm in the software package PHYLIP, vers. 3.69 (Felsenstein6) and were visualized with Dendroscope software, vers. 3.2.10 (Huson et ah, 2007). Majority-rule consensus values were calculated from 10,000 bootstrap samples of the data through the use of the PHYLIP component CONSENSE. The Fqj values and genetic distances computed are expected to provide an inflated estimate of the isolation between populations because the SNP loci used in our analyses were not a random sample from the genome; some SNP loci were chosen for their power in resolving specific population pairs in our baseline. Nonetheless, these es- timates are useful for assessment of the relative genetic differentiation among the populations described here. Power analyses We used 3 different methods to assess the power of the SNP baseline for GSI. First, we performed a self- assignment analysis, and subsequently we generated and analyzed simulated mixtures with 2 different procedures. In self-assignment analysis, allele frequencies for each potential source population generally are esti- mated from the samples. Then, for each individual, the probability that its genotype would occur in each population (assuming Hardy-Weinberg and linkage equilibria) is calculated, and the individual is assigned to the population for which its genotype probability is highest. We used the likelihood method of Rannala and Mountain (1997), implemented in the software gsi_sim7 (Anderson et ah, 2008), to compute the genotype prob- 5 Belkhir, K., P. Borsa, L. Chikhi, N. Raufaste, and F. Bonhom- me. 1996-2004. GENETIX 4.05, logiciel sous WindowsTM pour le genetique des populations. Laboratoire Genome, Populations, Interactions, CNRS UMR 5000, LTniversite de Montpellier II, Montpellier, France. [Available from http:// kimura.univ-montp2.fr/genetix.] 6 Felsenstein, J. 2005. PHYLIP (Phylogeny Inference Pack- age), vers. 3.6. Department of Genome Sciences, Univ. Wash- ington, Seattle. [Available from http://evolution. genetics. washington.edu/phylip.html.] 7 Available from http://swfsc.noaa.gov/textblock.aspx7Division =FED&ParentMenuID=54&id= 12964. Clemento et al.: Evaluation of a single nucleotide polymorphism baseline for genetic stock identification of Oncorhynchus tshawytscha 119 Table 2 List of the 96 single nucleotide polymorphism (SNP) loci used to construct the baseline for genetic stock identification of Chinook Salmon ( Oncorhynchus tshawytscha ) from the west coast of North America, with dbSNP accession numbers (from the National Center for Biotechnology Information online repository for short genetic variations; https://www.ncbi.nlm.nih. gov/snp) and source reference (Sr) where available: l=Clemento et al., 2011; 2=Smith et al., 2005a; 3=Campbell and Narum, 2008; 4=Smith et al., 2005b. Locus dbSNP Sr Locus dbSNP Sr Locus dbSNP Sr Ots_94857-232 ss2755 18685 1 Ots_l 10495-380 ss275518741 1 Ots_131906-141 ss275518787 1 Ots_96222-525 ss275518688 1 Ots_110551-64 ss275518742 1 Ots_AldBl-122 ss275518788 1 0ts_96500-180 ss2755 18689 1 OkiOts_120255-113 unpubl. - Ots_AldoB4-183 ss275518789 1 Ots^97077-179 ss275518691 1 Ots_l 11.312-435 ss275518746 1 Ots Myc-366 ss275518795 1 Ots_99550-204 ss275518695 1 Ots_l 11666-408 ss275518747 1 Ots_ALDBINTl-SNPl ss275518796 1 Ots 100884-287 ss275518696 1 Ots„l 11681-657 ss275518748 1 Ots_NAML12-SNPl ss275518798 1 Ots_101119-381 ss275518697 1 Ots_112208-722 ss275518749 1 Ots_ ARNT-195 unpubl. - Ots_101704-143 ss275518699 1 Ots_112301-43 ss275518750 1 Ots_RAG3 unpubl. - Ots_102213-210 ss275518702 1 Ots_l 12419-131 ss275518751 1 Ots_AsnRS-60 ss48398657 2 Ots_102414-395 ss275518703 1 Ots_l 12820-284 ss275518752 1 Ots_aspat-196 ss65917744 3 Ots_102420-494 ss275518704 1 Ots_l 12876-371 ss275518753 1 Ots_CD59-2 unpubl - Ots_102457-132 ss275518705 1 Ots_l 13242-216 ss275518754 1 Ots_CD63 unpubl. - 0ts._102801-308 ss275518706 1 Ots_113457-40 ss275518755 1 Ots_EP-529 unpubl. - Ots. 102867-609 ss275518707 1 Ots_l 17043-255 ss275518757 1 Ots_GDH-81x ss65917741 3 Ots. 103041-52 ss275518708 1 OtsH 17242-136 ss275518759 1 Ots„HSP90B-385 ss65713207 2 OtsJ 04063-132 ss275518711 1 Ots_l 17432-409 ss275518762 1 Ots_MHCl ss49851328 4 Ots 104569-86 ss275518714 1 Ots_l 18175-479 ss275518763 1 Ots_mybp-85 unpubl - Ots_105105-613 ss275518715 1 Ots_ 118205-61 ss275518764 1 Ots_myoD-364 ss65917726 3 Ots^ 105132-200 ss275518716 1 Ots_118938-325 ss275518765 1 Ots_Ots311-101x ss65917748 3 Ots_105401-325 ss275518718 1 Ots.l 22414-56 ss275518767 1 Ots_PGK-54 unpubl. - Ots_105407-117 ss275518719 1 Ots_123048-521 ss275518768 1 Ots_Prl2 ss49851322 4 Ots_106499-70 ss275518724 1 Ots_123921-lll ss275518770 1 Ots_.RFC2-558 ss48398670 2 Ots_106747-239 ss275518725 1 Ots_124774-477 ss275518771 1 Ots_SClkF2R2-135 ss48398694 2 Ots_107074-284 ss275518726 1 Ots„127236-62 ss275518773 1 Ots_SWSlop-182 ss48398635 2 Ots_107285-93 ss275518728 1 Ots„128302-57 ss275518775 1 Ots_TAPBP unpubl. - Ots_107806-821 ss275518730 1 Ots_128693-461 ss275518777 1 Ots_u07-07.161 unpubl. - 0ts_108007-208 ss275518731 1 Ots_128757-61 ss275518778 1 Ots_u07-49.290 unpubl. - Ots_108390-329 ss275518732 1 Ots_129144-472 ss275518779 1 Ots_u4-92 ss48398636 2 Ots_108735-302 ss275518733 1 Ots_129170-683 ss275518780 1 Ots_BMP2-SNPl ss275518800 1 Ots_109693-392 ss275518737 1 Ots_129458-451 ss275518782 1 Ots_TFl-SNPl ss275518802 1 Ots_l 10064-383 ss275518738 1 Ots_130720-99 ss275518784 1 Ots_S71-336 unpubl. - Ots_l 10201-363 ss275518739 1 Ots_131460-584 ss275518785 1 Ots_unk_526 unpubl. abilities, employing a leave-one-out procedure that ex- cludes the gene copies of the individual being assigned and recalculates population allele frequencies before assignment. Analogous to the THL procedure of An- derson (2010), both the training and holdout sets were included for estimation of population allele frequencies. However, assignments of individuals in the training set were excluded from the results to avoid any high-grad- ing bias of assignment accuracy (Anderson, 2010). Analysis of simulated mixed fisheries is a common method for evaluation of the resolving power of a base- line for stock identification (Fournier et al., 1984; Wood et al., 1987; Kalinowski, 2004; Beacham et al., 2006). In many studies, samples from simulated fisheries that consist entirely of fish from one population are ana- lyzed in so called “100% simulations.” However, such simulations typically do not assess how well the base- line will perform on samples from fisheries that exploit more than one stock. Therefore, we conducted simula- tions with 20 different mixing proportion vectors. The population composition of these mixtures was deter- mined by using the baseline to estimate the relative proportions of reporting units present in 20 different month-by-area strata sampled from commercial fisher- ies off the coast of California and Oregon in 2010 and 2011 (E. Crandall et al., unpubl. data). These vectors reflect mixing proportions that are expected to be encountered in PFMC fisheries. For a given value of the mixing proportion vector of all pop- ulations, a replicate simulation consisted of 1) simu- lating the number of fish from each population in a sample size of 200 by drawing a multinomial ran- dom variable with cell probabilities equal to the mix- ing proportion vector; 2) simulating the genotypes of 120 Fishery Bulletin 112(2-3) the individuals from each population in the mixture sample with 2 different techniques (“cross-validation over gene copies” [CV-GC] and K-fold cross valida- tion [K-fold], see next paragraph); 3) calculating the maximum likelihood estimator (MLE) of the mixture proportions for all the populations from the simulated sample through use of the baseline, which contains all training and holdout individuals; and 4) estimat- ing the mixing proportion of each reporting unit by summing the mixing proportion estimates of its con- stituent populations. For each of the 20 values of the mixing proportion vectors, 20,000 replicates were conducted with CV-GC, and 1000 replicates were con- ducted with K-fold. For both methods, the 5% and 95% quantiles of the distribution of the MLE of reporting- unit proportions were calculated from the replicates for each mixing proportion vector. Simulations were undertaken in 2 different ways. With CV-GC, genotypes were simulated by randomly sampling gene copies from the holdout set (to avoid high-grading bias), and those same gene copies were removed from the baseline when calculating the likeli- hood of population origin for the simulated individual (see Anderson et al., 2008). With K-fold, genotypes were simulated by drawing entire individuals with- out replacement (a technique commonly referred to as “jackknifing”) from the holdout set to form the mixture sample. Those sampled individuals were not included in the baseline, but all unsampled individuals from the holdout set were included in the baseline for estima- tion of the mixing proportions. Mixed fishery samples Samples from 2090 salmon landed in fisheries in 2010 were collected by the California Department of Fish and Game (now Wildlife) at California ports. Just over half of these fish carried CWTs that identified their population of origin. All samples were genotyped with our panel of 96 loci. Individuals successfully genotyped at fewer than 60 loci were removed from further analy- sis. Failed genotypes were ones that either clustered with negative controls during scoring or fell outside of defined heterozygote and homozygote clusters, likely indicating sample contamination (Smith et al., 2011; Larson et al., 2013). We also used an individual het- erozygosity (iHz; the proportion of heterozygous loci for each fish) criterion of iHz >0.56 to identify and ex- clude potentially contaminated samples. Simulations of contaminated genotypes determined by using observed allele frequencies, indicated little overlap in the dis- tribution of iHz for contaminated and uncontaminated samples (data not shown) and that uncontaminated samples rarely had iHz >0.56. We used the maximum likelihood framework in gsi_ sim to estimate the mixing proportion of different pop- ulations among the 2090 fish, and then used that MLE as the prior for calculation of the posterior probability of population of origin for each fish. Posterior probabili- ties of origination from different reporting units were obtained through summation of the population-specific probabilities over all populations in a reporting unit. Individuals were then assigned to the reporting unit with the highest posterior probability. Because all fish would be assigned to a maximum a posteriori (MAP) population regardless of true origin, we employed a simulation method similar to that in Cornuet et al. (1999), but which was modified to ac- count for missing data, to detect fish that might have originated from a population that was not in the base- line or that had an otherwise aberrant genotype. Brief- ly, for each fish from the fishery assigned to a popula- tion, the allele frequencies from the MAP population were used to simulate 10,000 genotypes with an identi- cal pattern of missing data (if any) to that of the fish that was assigned. The log-probability of each simulated genotype was computed, given that it came from the population it was simulated from, and then the distribution of those values was compared with the log-probability, La, of the actual assigned fish’s genotype, given the allele frequencies in the MAP population, on the basis of a z-score (La minus the mean of the simulated values, all divided by the standard deviation of the simulated val- ues). The z-score calculation was done conditional on the exact pattern of missing data and was implemented in the C programming language as part of the gsi_sim software. A low-confidence assignment was defined to be one that had a z-score <3.0 and had either a report- ing unit posterior probability <0.9 or had fewer than 90 loci successfully genotyped. Fish with low confidence assignments were left in an “unassigned” category. Results Genotyping and basic population genetics We successfully genotyped 8031 samples from 69 pop- ulations for the baseline and submitted the data to the Dryad Digital Repository (http://doi.org/10.5061/ dryad. 5745sv). All individuals were retained in the baseline, regardless of missing data because we de- sired a realistic representation of missing data pat- terns for subsequent power analyses. One locus failed to amplify entirely in the Copper River population, and 3 loci failed in the Coho Salmon sample. Unbi- ased estimates of heterozygosity (Nei, 1978) ranged from 0.194 in the Rapid River Hatchery stock of the Snake River reporting unit to 0.381 in the Smith Riv- er population. The Coho Salmon in the baseline had very low heterozygosity (0.094). Observed heterozygos- ity and mean number of alleles generally were lower for populations from north of the Columbia River (Ta- ble 1), likely due to an ascertainment bias resulting from the selection of SNPs with high MAFs in Califor- nia and Oregon populations. Clemento et al.: Evaluation of a single nucleotide polymorphism baseline for genetic stock identification of Oncorhynchus tshawytscha 121 Significant deviations from HWE (P<0.0001) were ob- served at various loci in 17 populations but represented <0.3% of all observations. Only the Butte Creek spring- run, Trinity River Hatchery spring-run, and Smith Riv- er populations were not in HWE at more than 2 loci, with 5, 5, and 4 significant tests, respectively. Similarly, only 3 loci deviated from HWE in more than 2 popu- lations: Ots_u07_07.161 in 3 populations, Ots_111312- 435 in 6 populations, and Ots_111666-408 in 4 popula- tions. Only 1 population (Trinity River Hatchery spring run) displayed significant LD (P<0.001) at more than 1% of locus comparisons (1.14%), and, over all popu- lations, the percentage of significant comparisons was 0.16%. Only 2 locus pairs were significant in more than 5 populations: Ots_AldBl-122 and Ots_AldoB4-183, known to be in the same gene complex, were in LD in 42 populations, and Ots_Myc-366 and Ots_unk-526 displayed LD in 8 populations. A large range in the degree of differentiation be- tween populations was observed (Table 1). Mean F st across all populations (excluding Coho Salmon) was 0.183, indicating that approximately 18% of genetic variation was partitioned between population samples. Within reporting units that contained more than one population (N=18), pairwise Fst was between 0.000 and 0.152 and had a mean value of 0.018. Ten pair- wise comparisons, all within reporting units, were not significantly different from zero (P<0.01). Between re- porting units, Fgj values ranged from 0.005 to 0.411 and had a mean value of 0.188. The least differentiated populations were the fall-run populations from Califor- nia’s Central Valley, as has been observed with other genetic data sets (Williamson and May, 2005; Seeb et ah, 2007). Genetic structure of the Chinook Salmon popula- tions in the baseline is displayed in an unrooted neigh- bor-joining dendrogram (Fig. 1). Relationships are in strong agreement with expectations that were based on geography and previous studies (Waples et al., 2004; Beacham et ah, 2006; Templin et ah, 2011; Moran et ah, 2013); populations generally are organized north to south along the main branch, and populations from within the same drainage usually cluster together. Populations from California’s Central Valley are monophyletic in relation to the remainder of the popu- lations but are characterized by short branch lengths, small distances between nodes, and low bootstrap sup- port. Central Valley spring-run and fall-run popula- tions also are monophyletic, with the exception of the Feather River Hatchery spring run, which is included in the fall-run reporting unit because of a history of substantial mtrogression between the runs and the consequent difficulty of genetically distinguishing this stock from fall-run fish (Garza et al.8). Sacramento Garza, J. C., S. M. Blankenship, C. Lemaire, and G. Char- rier. 2008. Genetic population structure of Chinook Salm- on ( Oncorhynchus tshawytscha) in California's Central Val- ley. Final report for CalFed project “Comprehensive evalua- tion of population structure and diversity for Central Valley River winter-run fish are quite distinct as a result of a well-documented recent bottleneck (Hedrick et al., 1995) and have one of the longest branches on the tree, with bootstrap support of 100%. Fish from rivers in northern California and coastal Oregon also form a monophyletic group. Columbia River populations are dispersed throughout the tree, although populations from the same reporting unit generally share a com- mon branch, as do populations from Alaska. Accuracy of assignment and mixture estimations The 7669 individuals that remained after removal of training-set fish were subjected to self-assignment with gsi_sim (Table 1). Correct assignment to popu- lation ranged from 13% for the Butte Creek fall-run population to 100% for 5 different populations. The following reporting units had the lowest correct as- signment rates to population: Central Valley fall run, Upper Columbia River summer/fall run, and Western Alaska, Lower Kuskokwim River, averaging 28%, 36%, and 40%, respectively. The lowest rate of correct as- signment to reporting unit was for the Siuslaw River population from the Mid Oregon Coast reporting unit, with over half of the individuals assigning to popula- tions in the North Oregon coast reporting unit. The largest change in correct assignment percentage from population to reporting unit was for the Central Valley fall run, which increased to 91%. The results of the mixture simulations for the 9 re- porting units most frequently found in California and Oregon fisheries appear in Figure 2. Results for the remaining reporting units are not shown because they are relatively uninformative as a result of the rar- ity with which populations from north of the Colum- bia River are encountered at the southern end of the California Current marine ecosystem, an observation corroborated by historical CWT data: in the 3 decades since 1983, only 0.5% of all CWTs recovered from Chinook Salmon in California ocean fisheries were from stocks outside of California or Oregon (Regional Mark Information System, Regional Mark Process- ing Center, http://www.rmpc.org). Accurate estimates of the mixing proportions were obtained for fishery samples simulated either by CV-GC or by K-fold. The mean maximum likelihood estimate of the proportion of each reporting unit was generally highly correlated with the true proportion, indicating that any bias was very small. For 6 reporting units (Central Valley fall run, Sac- ramento River winter run, Klamath River, California Coast, Rogue River, and North Oregon Coast), the 5% and 95% quantiles for reporting-unit mixing propor- tions corresponded closely with the quantiles one would obtain with perfect identification of all fish (see the gray regions in Fig. 2). The somewhat wider GSI quan- Salmon,” 54 p. [Available from http://swfsc.noaa.gov/publia- tions/FED/OlllO.pdf.J 122 Fishery Bulletin 112(2-3) a S •"J Q (D cd O "5 * S .2 3 Q 03 ^ A O £ £ 03 "> Q £ co a c -2 (D -££ cc Q o o •-H (D >-. C/2 O 3 03 Qh (X) CD ,03 O -G XI (XI o cp ^ O 03 (X) •2 .5 o S S 3 W ^ £ be m d) .S 03 rC CO Q ^ CD o cd C/3 ’’3 S ft ^ C 0) M O £ G >s 0) £> be

^ 0 r£ CD O CD +-> > _ Sh a s 1! ; ■ G rC 03 ? Clemento et al.: Evaluation of a single nucleotide polymorphism baseline for genetic stock identification of Oncorhynchus tshciwytscha 123 o J3 -d Cl O CL TJ 0) TO E True proportion True proportion True proportion Figure 2 Estimates of mixing proportions from cross-validation over gene copies (CV-GC) and K-fold sim- ulations for the 9 most abundant reporting units of Chinook Salmon (Oncorhynchus tshawyts- cha) encountered in California fisheries: Central Valley (A) spring, (B) fall, and (C) winter; (D) California Coast; (E) Klamath River; (F) North California/South Oregon Coast; (G) Rogue River; and (H) Mid Oregon Coast; and (I) North Oregon Coast. The x-axis gives the true proportion of fish from each reporting unit, and the y-axis gives the estimated proportion. The dashed line is the y=x line. Gray shaded regions give the range between the 5% and 95% quantiles of estimates that would be achieved with perfect assignment of fish to a reporting unit (i.e., they represent the uncertainty due to the fact that fishery proportions are estimated with a finite sample; in our simulations, a sample of 200 fish). The 5% and 95% quantiles of the estimates derived from the CV-GC and the K-fold replicates are shown with vertical line segments and open diamonds, respectively. Reporting units for which these bars and diamonds coincide with the gray region had estimated proportions as accurate as one would expect given unambiguous identification of fish to reporting unit. Filled circles and open triangles indicate the mean over 20,000 CV-GC and 1000 K-fold replicates, respectively. These points fall along the dotted line when the estimator is unbiased. 124 Fishery Bulletin 112(2-3) Table 3 Genetic stock identification (GSI) results from assignment of samples of Chinook Salmon ( Oncorhyn - chus tshawytscha) collected in 2010 from the California fishery to their source populations through the use of a single nucleotide polymorphism baseline, as well as concordance with recoveries of coded wire tags (CWTs). N.=North; S.=South. Stock Number from GSI Number with CWT Number of GSI-CWT matches GSI-CWT agreement (%) California Coast 30 1 0 0.00% Central Valley fall 1581 958 957 99.90% Central Valley spring 7 1 0 0.00% Klamath River 108 50 49 98.00% Lower Columbia spring 1 0 0 - Mid Columbia Tule fall 7 2 2 100.00% Mid Oregon Coast 14 1 0 0.00% N. California/S. Oregon Coast 58 25 25 100.00% Rogue River 154 11 5 45.45% Snake River fall 1 1 1 100.00% Upper Columbia summer/fall 8 2 2 100.00% Total 1969 1052 1041 98.95% tile intervals observed for the Central Valley spring- run reporting unit were likely due to its similarity to the Central Valley fall-run reporting unit, combined with the fact that the spring run is typically at much lower abundance than is the fall run. Likewise, the genetic similarity of fish from the Mid Oregon Coast reporting unit and the Northern California/Southern Oregon Coast reporting unit made it difficult to ac- curately estimate mixing proportions for these report- ing units; however, the estimates were still quite good and largely unbiased. Therefore, despite the enlarged quantile intervals for Central Valley spring-run and the Mid Oregon Coast reporting units versus Northern California-Southern Oregon reporting unit, the results from both simulation methods indicated that the SNP baseline is capable of providing estimates of the true mixing proportions for most reporting units that are nearly as accurate as one would expect given perfect identification of each fish. Fishery samples Of the 2090 samples collected from California fisher- ies in 2010, 85 samples were excluded because they did not yield acceptable genotypes (<60 successfully genotyped loci) and 2 samples were excluded because they were duplicates of 2 other samples in the data set. Eight fish exceeded the iHz threshold of 0.56 and were removed because of potential contamination. Sev- en fish were identified as Coho Salmon through both GSI assignment and with the species-diagnostic assay. Another 18 samples did not meet assignment confi- dence criteria (mean 2-score of -3.99 and a mean of 75 successfully genotyped loci) and were also excluded. For the remaining 1969 fish, assignment probabilities to reporting unit ranged from 36.4% to 100% (mean 98.5%) and z-scores ranged from -4.12 to 2.68 (mean -0.04). Centra] Valley fall-run fish dominated the stock composition, accounting for more than 80% of sampled fish, followed by the Rogue River (7.79%) and Klam- ath River (5.46%) reporting units and 8 other stocks with <5% (Table 3). Of the assigned fish, 1052 retained CWTs that were recovered. Genetic assignment to re- porting unit disagreed with CWT origin for only 11 fish (1.05%), and, of these mismatches, 6 were fish with Klamath or Smith River tags that were assigned to the genetically similar Rogue River reporting unit. Discussion Here we describe one of the first large-scale SNP base- lines for genetic stock identification of Chinook Salmon and the first designed for use with fisheries in the Cali- fornia Current large marine ecosystem off the contigu- ous United States. Chinook Salmon are an economi- cally and ecologically important species and are a ma- jor component of fisheries in the North Pacific Ocean. We genotyped more than 8000 individual fish from 69 distinct populations at 96 SNP loci to construct the baseline. The reporting units included in the baseline represent the likely sources for more than 99% of the fish typically encountered in PFMC fisheries off Cali- fornia and Oregon. Furthermore, results from mixture analyses and self-assignment indicate that the baseline has near maximum possible power for discrimination of Chinook Salmon stocks at the reporting unit level. Estimates of Clemento et al.: Evaluation of a single nucleotide polymorphism baseline for genetic stock identification of Oncorhynchus tshawytscha 125 the mixture proportions of Central Valley fall, Central Valley winter, California Coast, Klamath River, and Rogue River reporting units (Fig. 2) were no more vari- able than were estimates that would have been obtained if every fish had carried an unambiguous reporting-unit tag. Estimates of mixing proportions for Central Valley spring, North California/South Oregon, and Mid Oregon Coast reporting units were somewhat more variable but appeared to be nearly unbiased. In the ocean fishery sample, assignments of more than 1000 individuals to reporting unit, determined with our baseline, were high- ly concordant (98.95%) with the CWTs recovered from the same fish. This SNP baseline, therefore, represents an important addition to the technologies available to managers and researchers. Methodological considerations Management of salmon fisheries in the Pacific Ocean off North America can be roughly divided into 3 fisher- ies by region: California and Oregon fisheries, managed by the Pacific Fishery Management Council (PFMC); Washington, British Columbia, Canada, and southeast- ern Alaska fisheries, subject to the international Pacific Salmon Treaty, reporting to and regulated by the Pa- cific Salmon Commission; and fisheries farther north and west in Alaska that are managed by the state, with salmon bycatch under the purview of the North Pacific Fishery Management Council. The genetic base- line described here was designed primarily to identify fish caught in PFMC ocean fisheries and in ecological investigations in the southern portion of the California Current ecosystem and its associated tributary rivers and streams. We have shown that it performs well in this area but, because of an ascertainment strategy during SNP discovery that included individuals from the Columbia River and British Columbia (Clemento et ah, 2011), the baseline also has sufficient statisti- cal power to identify the source of some fish from else- where in the North American range of this species. We observed high rates of self-assignment to report- ing unit for all regions represented in the baseline, although some reporting units clearly were composed of populations with minimal differentiation from each other. Moreover, the utility of our baseline could be ex- tended effectively by simply genotyping the same panel of SNPs on additional populations in those regions, de- spite the reduced heterozygosity and mean number of alleles (Table 1), and presumably statistical power in our baseline, for populations from Canada and Alaska. Other SNP baselines for Chinook Salmon also have been described or are being constructed. Templin et al. (2011) described a 45 SNP locus baseline for popula- tions in the northern and western parts of the Chinook Salmon range, designed primarily for GSI of popula- tions from western and southcentral Alaska. This same baseline was used also to probe the seasonal distribu- tion and migration pattern of Chinook Salmon in the Bering Sea and North Pacific Ocean (Larson et al., 2013). Despite the presence of 14 populations from California, Oregon, and Washington in that baseline, Larson et al. (2013) appropriately emphasized that resolution of those southern populations is sufficient only for broad-scale assignments. Similarly, Warheit et al.3 described the marker selection for eventual devel- opment of a SNP baseline for application to fisheries managed by the Pacific Salmon Commission. Although the existence of multiple regional base- lines is likely to expand, it still will benefit the entire community of fishery managers and scientists to care- fully design marker panels with as much overlap as possible. It is conceivable that 2 or 3 panels of 96 SNPs could provide the level of resolution needed for identi- fication throughout the range of Chinook Salmon. Al- ternatively, as next-generation sequencing techniques mature, genotyping-by-sequencing (GBS) approaches might yield data for GSI at a lower cost than that with current genotyping techniques. A GBS approach could be used to simultaneously genotype all of the SNPs in each of the regional baselines, allowing mixed-stock analysis throughout the range of this species. Inclusion of the species-diagnostic marker and Coho Salmon sample in the baseline provided insight into the prevalence of misidentification of Coho Salmon in ocean fisheries. In the 2010 fishery off California, 7 fish sampled as Chinook Salmon were found to be Coho Salmon. Without methods to identify Coho Salmon, the baseline would assign them with erroneously high confidence to a northern, low-heterozygosity Chinook Salmon population (data not shown). This problem is characteristic of most statistical methods for perform- ing GSI: if an individual’s true population of origin is not included in the baseline, then even if all the pop- ulations in the baseline are very poor candidates for that fish’s origin, that fish might still be assigned with high posterior probability to one of the populations. This situation occurs when one population is much more likely to be the population of origin, than any of the other incorrect populations, even if it is not a likely origin for that individual on an absolute scale. We introduced a simulation-based z-score method, implemented in gsi_sim, to identify fish that likely have not originated from populations in the baseline. An alternative, Bayesian nonparametric approach to dealing with fish from populations not in the baseline identifies those fish and estimates the allele frequen- cies in their (unrepresented) source population (Pella and Masuda, 2000). That approach is appropriate particularly when large numbers of fish are sampled from each of the populations that are not included in the baseline and when the unrepresented populations are quite divergent from all of the populations in the baseline. We chose the z-score approach over the Bayesian nonparametric approach for 3 main reasons: 1) it is computationally fast and simple (there are no conver- gence problems that might be difficult to detect); 2) our baseline was sufficiently comprehensive for stocks that 126 Fishery Bulletin 112(2-3) contribute to PFMC fisheries, and therefore it was un- likely that large numbers of fish would originate from any single unrepresented population, let alone a highly divergent one; and 3) our approach is more appropriate for identification of fish whose genotypes are aberrant because of genotyping complications or sample contam- ination. Regardless of which method is used, all GSI estimation should include some analysis to identify fish that are either from populations not included in the baseline or that have aberrant genotypes for another reason. GSI is highly dependent on source populations be- ing genetically differentiated enough from one another for discrimination. In situations where hatchery brood- stock transfers, supplementation, or other processes in- crease straying and gene flow between fish populations, genetic differentiation decreases and it can become more difficult to use GSI. Such is the case in the Cen- tral Valley of California, where average Fgx between populations in the fall-run reporting unit was 0.006 and in the spring-run reporting unit was 0.013. In the dendrogram (Fig. 1), this region was characterized by extremely short branch lengths, small internodal dif- ferences, and weak bootstrap support. Extensive stray- ing of hatchery salmon due to off-site juvenile releases (California Hatchery Scientific Review Group9) and water operations (Fisher, 1994) has eliminated his- torical differentiation between populations of fall-run Chinook Salmon (Williamson and May, 2005). Intro- gression between fall-run and spring-run fish at the Feather River Hatchery, and likely elsewhere within the basin, has reduced differentiation between these 2 phenotypes, with mean F gx of 0.025 between fall-run and naturally spawning spring-run populations. Sampling of different stocks for baseline construc- tion in the presence of high stray rates is not entirely straightforward, particularly when populations are largely sympatric and not visually distinguishable. For example, there is clearly a single fish from the Central Valley fall-run reporting unit that was sam- pled as a winter-run fish in our baseline. These types of occurrences are almost inevitable given the high degree of disturbance and hatchery supplementation over much of the range of Chinook Salmon. One ap- proach is to move fish with discrepant genotypes from the baseline populations in which they were sampled to the ones to which they are assigned with GSI (e.g., Banks et al., 2000). However, such a procedure can introduce an upward bias in the predicted accuracy of the baseline, if, in fact, the removed fish actually do belong to the populations from which they were sam- pled and simply have unlikely genotypes at the genet- ic markers used for baseline construction. We chose to 9 California Hatchery Scientific Review Group. 2012. Cali- fornia Hatchery Review Report, 102 p. Prepared for the U.S. Fish and Wildlife Service and Pacific States Marine Faisheries Commission. [Available from http://swfsc.noaa. gov/publications/FED/01067.pdf and (appendices) http://ca- hatcheryreview.com/reports.] be conservative by both 1) accepting a slightly lower rate of predicted resolution obtained by not removing miscategorized fish and 2) avoiding an upward bias in predicted GSI accuracy if the fish removed are not miscategorized. Implications for management Accurately estimating the proportion of fish from dif- ferent populations in mixed-stock ocean fisheries has important applications for harvest management and conservation. Stocks that comingle in ocean fisheries can vary widely in productivity and abundance. With- out precise information on their ocean distribution, as can be provided by GSI, managers have few options for protection of depressed or at-risk stocks from fish- ery impacts other than that of shutting down or cur- tailing fisheries over broad areas, as is currently done (Beacham et al., 2008). For example, in 2008 and 2009, the largest closures on record of fisheries in Califor- nia and Oregon were enacted to protect the severely reduced Central Valley fall-run stock (Lindley et al., 2009). The economic effects of fishery closures are substantial, resulting in millions of dollars of lost in- come for fishermen, coastal communities, and retailers (Michael 10). Management of Chinook Salmon in California, Or- egon, and Washington and in fisheries managed by the Pacific Salmon Commission depends heavily on information generated by an elaborate CWT program (Hankin et al., 2005). Tiny wire tags are mechanically implanted into the heads of juvenile fish, with each tag bearing a code that identifies the release group and source hatchery (or stock) of that fish. Tagging of natu- rally spawned juvenile fish has generally proven un- successful (Beacham et al., 1996), and, for that reason, tagged hatchery stocks are used as proxies to estimate fishery impacts for groups of natural stocks. Aside from the largely unvalidated assumption that such proxies accurately reflect fishery impacts on associated natu- ral stocks (Hankin et al., 2005), the physical effects of tagging fish and removing their nerve-rich adipose fin (Buckland-Nicks et al., 2012) as an associated ex- ternal mark can increase disease transmission (Elliott and Pascho, 2001), interfere with homing (Morrison and Zajac, 1987; Habicht et al., 1998) and swimming- ability (Reimchen and Temple, 2004) and may affect size-at-return for adult salmon (Vander Haegen et al., 2005). Moreover, extremely low recovery rates mean that CWT data are often quite limited and great un- certainty is frequently associated with the estimates derived from them (Hankin et al., 2005). GSI has been advanced as an alternative to CWTs in fishery management for several decades. Our direct 10Michael, J. 2010. Employment impacts of California salm- on fishery closures in 2008 and 2009. Business Forecasting Center, Univ. of the Pacific, Stockton, CA. [Available from http://forecast.pacific.edu/BFC%20salmon%20jobs.pdf.] Clemento et al.: Evaluation of a single nucleotide polymorphism baseline for genetic stock identification of Oncorhynchus tshawytscha 127 comparison of CWT with genetic assignments demon- strates that our baseline is capable of identifying fish to reporting unit with accuracy comparable to that of CWTs. Furthermore, the use of GSI can identify consid- erably more fish to reporting unit, including fish from natural stocks. Confident genetic assignments were ob- tained for -94% of fish from the 2010 fishery sample, but only 1052 of those fish carried CWTs and this num- ber is inflated partially because of oversampling of fish believed to carry CWTs. Fishery management decisions rely heavily on co- hort-based ocean harvest models (cf., O’Farrell et al., 2012), which require information on both stock of ori- gin and age of fish impacted by fisheries. Because GSI does not provide the age of individuals, it is not by itself an adequate alternative to CWTs. Nonetheless, new statistical methods capable of integrating GSI, length data, and scale- or otolith-based age data have been developed recently, allowing managers to draw important inference about PFMC fisheries that are not possible with CWTs alone (Satterthwaite et ah, 2014). Moreover, pedigree-based genetic tagging does supply age for salmon (Anderson and Garza, 2006; Garza and Anderson2). This method, termed “parentage-based tag- ging” (PBT), can identify the actual parents of a geno- typed individual through parentage analysis if they have been genotypecl with the same genetic markers. If the parents’ date of spawning is known, as it typically is in a hatchery, then the reconstructed pedigrees yield the offspring’s precise age and any associated parental spawning information. Importantly, both PBT and GSI can be undertaken with the same SNP genotypes, and the SNPs used in our GSI baseline are sufficiently powerful for PBT with Chinook Salmon from California to Washington (Ander- son, 2012). This interoperability of genotype data en- ables an integrated program that uses both GSI and PBT simultaneously, providing identification for all fish in a fishery or ecological sample and yielding signifi- cantly greater inference than either method alone. For example, GSI cannot distinguish between spring-run and fall-run fish from the Feather River Hatchery in California, but PBT distinguishes them, almost with- out error, from any mixture. Likewise, although it is difficult to implement PBT in natural populations, the same SNP genotypes used in a PBT analysis permit accurate identification (by GSI) of fish from the natu- rally spawning, ESA-listed “California Coastal Chinook Salmon Evolutionarily Significant Unit.” Conclusions The advent of high-throughput SNP genotyping al- ready has revolutionized human genetics (Jenkins and Gibson, 2002), providing previously unattainable resolution (e.g., Novembre et ah, 2008) and is poised to do the same for fisheries biology and management. As described here, we used a careful and statistically valid power analysis of SNP genotypes from a large number of Chinook Salmon populations concentrated at the southern end of the native range of this spe- cies to show that SNPs can provide a powerful baseline for genetic stock identification (see also Larson et ah, 2013) in fisheries and ecological investigation in the California Current large marine ecosystem and its trib- utaries in California and Oregon. We predict that these advances in genetic resources and methods will foster fundamental improvements in the way salmon popula- tions are studied, monitored, and managed. Acknowledgments The authors would like to thank the entire Molecular Ecology and Genetic Analysis Team in the Fisheries Ecology Division of the Southwest Fisheries Science Center (SWFSC) for their invaluable assistance with genotyping and analyses. Of critical importance to the successful completion of this project were the baseline samples provided to us by the California Department of Fish and Game (now Wildlife; S. Harris), Hoopa Valley Tribal Fisheries Department (G. Kautsky), Oregon De- partment of Fish and Wildlife, Oregon State University Department of Fisheries and Wildlife (M. Banks), Ida- ho Department of Fish and Game (M. Campbell), Co- lumbia River Inter-Tribal Fish Commission (S. Narum), NOAA Northwest Fisheries Science Center (P. Moran), U.S. Fish and Wildlife Service (M. Brown, D. Hawkins, and C. Smith), Washington Department of Fish and Wildlife (S. Blankenship and K. Warheit), University of Washington School of Aquatic and Fishery Science (L. Seeb), Department of Fisheries and Oceans, Canada (T. Beacham), and Alaska Department of Fish and Game (W. Templin). Fishery samples were collected by the California Department of Fish and Game (now Wild- life) and provided to us by M. Heisdorf and M. Palmer- Zwahlen. We also thank T. Beacham and 2 anonymous referees for comments that improved the manuscript. This project received funding from NOAA’s Coopera- tive Fisheries Research Program and the SWFSC. A. Clemento also received support from a California Bay Delta Science Fellowship and the University of Califor- nia Coastal Environmental Quality Initiative. Many of the baseline samples were collected and DNA extracted with funds from the Pacific Salmon Commission. Literature cited Abadia-Cardoso, A., A. J. Clemento, and J. C. Garza. 2011. Discovery and characterization of single nucleotide polymorphisms in steelhead/rainbow trout, Oncorhyn- chus mykiss. Mol. Ecol. Resour. 11 (suppl. sl):31-49. Abadia-Cardoso, A., E. C. Anderson, D. E. Pearse, and J C. Garza 2013. Large-scale parentage analysis reveals repro- ductive patterns and heritability of spawn timing in 128 Fishery Bulletin 112(2-3) a hatchery population of steelhead ( Oncorhynchus mykiss). Mol. Ecol. 22:4733-4746. Aguilar, A., and J. C. Garza. 2008. Isolation of 15 single nucleotide polymorphisms from coastal steelhead, Oncorhynchus mykiss (Salmoni- dae). Mol. Ecol. Resour. 8:659-662. Allendorf, F., and L. W. Seeb. 2000. Concordance of genetic divergence among sockeye salmon populations at allozyme, nuclear DNA, and mi- tochondrial DNA markers. Evolution 54:640-51. Anderson, E. C. 2010. Assessing the power of informative subsets of loci for population assignment: standard methods are up- wardly biased. Mol. Ecol. Resour. 10:701-710. 2012. Large-scale parentage inference with SNPs: an ef- ficient algorithm for statistical confidence of parent pair allocations. Stat. Appl. Genet. Mol. Biol. 11:12. doi: 10.1515/1544-6115.1833 Anderson, E. C., and J. C. Garza. 2006. The power of single-nucleotide polymorphisms for large-scale parentage inference. Genetics 172:2567-2582. Anderson, E. C., R. S. Waples, and S. T. Kalinowski. 2008. An improved method for predicting the accuracy of genetic stock identification. Can. J. Fish. Aquat. Sci. 65:1475-1486. Banks, M. A., V. K. Rashbrook, M. Calavetta, C. Dean, and D. Hedgecock. 2000. Analysis of microsatellite DNA resolves genetic structure and diversity of Chinook salmon ( Oncorhyn- chus tshawytscha) in California’s Central Valley. Can. J. Fish. Aquat. Sci. 57:915-927. Barnett-Johnson, R., T. E. Pearson, F. C. Ramos, C. Grimes, and R. B. MacFarlane. 2008. Tracking natal origins of salmon using isotopes, otoliths, and landscape geology. Limnol. Oceanogr. 53:1633-1642. Beacham, T. D., R. E. Withler, and T. Stevens. 1996. Stock identification of chinook salmon ( Oncorhyn- chus tshawytscha) using minisatellite DNA varia- tion. Can. J. Fish. Aquat. Sci. 53:380-394. Beacham, T. D., J. R. Candy, K. L. -Jonsen, -J. Supernault, M. Wetklo, L. Deng, K. M. Miller, R. E. Withler, and N. Varnavskaya. 2006. Estimation of stock composition and individual identification of Chinook salmon across the Pacific Rim by use of microsatellite variation. Trans. Am. Fish. Soc. 135:861-888. Beacham, T. D., I. Winther, K. L. Jonsen, M. Wetklo, L. Deng, and J. R. Candy. 2008. The application of rapid microsatellite-based stock identification to management of a Chinook Salmon troll fishery off the Queen Charlotte Islands, British Colum- bia. N. Am. J. Fish. Manage. 28:849-855. Bernatchez, L., and P. Duchesne. 2000. Individual-based genotype analysis in studies of parentage and population assignment: how many loci, how many alleles? Can. J. Fish. Aquat. Sci. 57:1-12. Boyce, N. P., Z. Kabata, and L. Margolis. 1985. Investigations of the distribution, detection, and biology of Henneguya salminicola (Protozoa, Myxozoa), a parasite of the flesh of Pacific salmon. Can. Tech. Rep. Fish. Aquat. Sci. 1405, 55 p. Buckland-Nicks, J. A., M. Gillis, and T. E. Reimchen. 2012. Neural network detected in a presumed vestigial trait: ultrastructure of the salmonid adipose fin. Proc. R. Soc. Lond., Ser. B: Biol. Sci. 279:553-563 Campbell, N., and S. R. Narum. 2008. Identification of novel single-nucleotide polymor- phisms in Chinook salmon and variation among life his- tory types. Trans. Am. Fish. Soc. 137:96-106. Cavalli-Sforza, L. L., and A. W. F. Edwards. 1967. Phylogenetic analysis: models and estimation pro- cedures. Am. J. Hum. Genet. 19:233-257. Claytor, R., and H. MacCrimmon. 1988. Morphometric and meristic variability among North American Atlantic salmon (Salmo salar). Can. J. Zool. 66:310-317. Clemento, A. J., A. Abadia-Cardoso, H. A. Starks, and J. C. Garza. 2011. Discovery and characterization of single nucleo- tide polymorphisms in Chinook salmon, Oncorhynchus tshawytscha . Mol. Ecol. Resour. 11 (suppl. sl):50-66. Cook, R. C. 1982. Stock identification of sockeye salmon ( Oncorhyn- chus nerka ) with scale pattern recognition. Can. J. Fish. Aquat. Sci. 39:611-617. Cornuet, J., S. Piry, G. Luikart, A. Estoup, and M. Solignac. 1999. New methods employing multilocus genotypes to select or exclude populations as origins of individuals. Genetics 153:1989-2000. Cronin, M. A., W. J. Spearman, and R. L. Wilmot. 1993. Mitochondrial DNA variation in Chinook ( On- corhynchus tshawytscha ) and chum salmon (O. keta ) detected by restriction enzyme analysis of polymerase chain reaction (PCR) products. Can. J. Fish. Aquat. Sci. 50:708-715. Elliott, D. G., and R. J. Pascho. 2001. Evidence that coded-wire-tagging procedures can enhance transmission of Renibacterium salmoninarum in Chinook salmon. J. Aquat. Anim. Health 13:181-193. Federal Register. 1990. Endangered and threatened wildlife and plants; listing of the Sacramento River winter-run Chinook salmon as threatened. Vol. 55(231):4962. GPO, Wash- ington, DC. 1999. Endangered and threatened species; threatened status for three Chinook Salmon Evolutionarily Signifi- cant Units (ESUs) in Washington and Oregon, and en- dangered status for one Chinook Salmon ESU in Wash- ington. Vol. 64(56):14308-14328. GPO, Washington, DC. Fisher, F. W. 1994. Past and present status of Central Valley Chinook salmon. Conserv. Biol. 8:870-873. Flannery, B. G., J. K. Wenburg, and A. J. Gharrett. 2007. Variation of amplified fragment length polymor- phisms in Yukon River chum salmon: population struc- ture and application to mixed-stock analysis. Trans. Am. Fish. Soc. 136:911-925. Fournier, D. A., T. D. Beacham, B. E. Riddell, and C. A. Busack. 1984. Estimating stock composition in mixed stock fish- eries using morphometric, meristic, and electrophoretic characteristics. Can. J. Fish. Aquat. Sci. 41:400-408. Habicht, C., S. Sharr, D. Evans, and J. E. Seeb. 1998. Coded wire tag placement affects homing ability of pink salmon. Trans. Am. Fish. Soc. 127:652-657. Clemento et al.: Evaluation of a single nucleotide polymorphism baseline for genetic stock identification of Oncorhynchus tshawytscha 129 Hankin, D. G., J. H. Clark, R. B. Deriso, J. C. Garza, G. S. Mor- ishima, B. E. Riddell, C. Schwarz, and J. B. Scott. 2005. Report of the expert panel on the future of the cod- ed wire tag recovery program for Pacific salmon. Pacif- ic Salmon Commission Tech. Rep. 18, 230 p. [Available from http://www.psc.org/pubs/psctrl8.pdfl Hedrick, P. W., D. Hedgecock, and S. Hamelberg. 1995. Effective population size in winter-run Chinook salmon. Conserv. Biol. 9:615-624. Iluson, D. H., D. C. Richter, C. Rausch, T. Dezulian, M. Franz, and R. Rupp. 2007. Dendroscope: an interactive viewer for large phylogenetic trees. BMC Bioinformatics 8:460. doi: 10.1186/1471-2105-8-460 Jenkins, S., and N. Gibson. 2002. High-throughput SNP genotyping. Comp. Funct. Genomics 3:57-66. Kalinowski, S. T. 2004. Genetic polymorphism and mixed-stock fisheries analysis. Can. J. Fish. Aquat. Sci. 61:1075-1082. Larson, W. A., F. M. Utter, K. W. Myers, W. D. Templin, J. E. Seeb, C. M. Guthrie III, A. V. Bugaev, and L. W. Seeb. 2013. Single-nucleotide polymorphisms reveal distri- bution and migration of Chinook salmon ( Oncorhyn- chus tshawytscha ) in the Bering Sea and North Pacific Ocean. Can. J. Fish. Aquat. Sci. 70:128-141. Lindley, S. T., C. B. Grimes, M. S. Mohr, W. Peterson, J. Stein, J. T. Anderson, L. W. Botsford, D. L. Bottom, C. A Busack, T. K. Collier, J. Ferguson, J. C. Garza, A. M. Grover, D. G. Hankin, R. G. Kope, P. W. Lawson, A. Low, R. B. MacFarlane, K. Moore, M. Palmer-Zwahlen, F. B. Schwing, J. Smith, C. Tracy, R. Webb. B. K. Wells, and T. H. Williams. 2009. What caused the Sacramento River fall Chinook stock collapse? NOAA Tech. Memo. NMFS-SWF- SC-447, 121 p. Louis, E. J., and E. R. Dempster. 1987. An exact test for Hardy- Weinberg and multiple al- leles. Biometrics 43:805-811. Miller, K. M., R. E. Withler, and T. D. Beacham. 1996. Stock identification of coho salmon ( Oncorhynchus kisutch ) using minisatellite DNA variation. Can. J. Fish. Aquat. Sci. 53:181-195. Milner, G. B., D. J. Teel, F. M. Utter, and G. A. Winans. 1985. A genetic method of stock identification in mixed populations of Pacific salmon, Oncorhynchus spp. Mar. Fish. Rev. 47:1-8. Moran, P, D. J. Teel, M. A. Banks, T. D. Beacham, M. R. Bell- inger, S. M. Blankenship, J. R. Candy, J. C. Garza, J. E. Hess, S. R. Narum, L. W. Seeb, W. D. Templin, C. G. Wallace, and C. T. Smith. 2013. Divergent life-history races do not represent Chi- nook salmon coast-wide: the importance of scale in Quaternary biogeography. Can. J. Fish. Aquat. Sci. 70: 415-435. Morrison, J., and D. Zajac. 1987. Histologic effect of coded wire tagging in chum salmon. N. Am. J. Fish. Manage. 7:439-441. Narum, S. R., M. A. Banks, T. D. Beacham, M. R. Bellinger, M. R. Campbell, J. Dekoning, A. Elz, C. M. Guthrie, C. Kozfkay, K. M. Miller, P. Moran, R. Phillips, L.W. Seeb, C. T. Smith, K. Warheit, S. F. Young, and J. C. Garza. 2008. Differentiating salmon populations at broad and fine geographical scales with microsatellites and single nucleotide polymorphisms. Mol. Ecol. 17:3464-3477. Nei, M. 1978. Estimation of average heterozygosity and genetic distance from a small number of individuals. Genetics 89:583-590. Nosil, P, D. Funk, and D. Ortiz-Barrientos. 2009. Divergent selection and heterogeneous genomic divergence. Mol. Ecol. 18:375-402. Novembre, J., T. Johnson, K. Bryc, Z. Kutalik, A. Boyko, A. Auton, A. Indap, K. King, S. Bergmann, M. Nelson, M. Ste- phens, and C. D. Bustamante. 2008. Genes mirror geography within Europe. Nature 456:98-101. O’Farrell, M. R., M. S. Mohr, A. M. Grover, and W. H. Satterthwaite. 2012. Sacramento River winter Chinook cohort recon- struction: analysis of ocean fishery impacts. NOAA Tech. Memo. NOAA-TM-NMFS-SWFSC-491, 74 p. Pella, J., and M. Masuda. 2000. Bayesian methods for analysis of stock mixtures from genetic characters. Fish. Bull. 99:151-167. Rannala, B., and J. L. Mountain. 1997. Detecting immigration by using multilocus geno- types. Proc. Natl. Acad. Sci. USA 94:9197-9201. Reimchen, T. E., and N. F. Temple. 2004. Hydrodynamic and phylogenetic aspects of the adipose fin in fishes. Can. J. Zool. 82:910-916 Rousset, F. 2008. Genepop’007: a complete reimplementation of the Genepop software for Windows and Linux. Mol. Ecol. Resour. 8:103-106. Satterthwaite, W., M. S. Mohr, M. R. O’Farrell, E. C. Anderson, M. A. Banks, S. J. Bates, M. R. Bellinger, L. A. Borgerson, E. D. Crandall, J. C. Garza, B. J. Kormos, P. W. Lawson, and M. L, Palmer-Zwahlen. 2014. Use of genetic stock identification data for com- parison of the ocean spatial distribution, size-at-age, and fishery exposure of an untagged stock and its in- dicator: California Coastal versus Klamath River Chi- nook. Trans. Am. Fish. Soc. 143:117-133. Seeb, L. W., A. Antonovich, M. A. Banks, T. D. Beacham, M. R. Bellinger, S. M Blankenship, M. R. Campbell, N. A. De- covich, J. C. Garza, C. M. Guthrie III, T. A. Lundrigan, P. Moran, S. R. Narum, J. J. Stephenson, K. J. Supernault, D. J. Teel, W. D. Templin, J. K. Wenburg, S. F. Young, and C. T. Smith. 2007. Development of a standardized DNA database for Chinook salmon. Fisheries 32:540-552. Seeb, J. E., G. Carvalho, L. Hauser, K. Naish, S. Roberts, and L. W. Seeb. 2011. Single-nucleotide polymorphism (SNP) discovery and applications of SNP genotyping in nonmodel organ- isms. Mol. Ecol. Resour. 11 (suppl. sl):l-8. Shaklee, J. B., and S. R. Phelps. 1990. Operation of a large-scale, multi-agency program for genetic stock identification. Am. Fish. Soc. Symp. 7:817-830. Smith, C. T., A. Antonovich, W. D. Templin, C. M. Elfstrom, S. R. Narum, and L. W. Seeb. 2007. Impacts of marker class bias relative to lo- cus-specific variability on population inferences in Chinook salmon: a comparison of single-nucleotide polymorphisms with short tandem repeats and allo- zymes. Trans. Am. Fish. Soc. 136:1674-1687. 130 Fishery Bulletin 112(2-3) Smith, C. T., C. M. Elfstrom, L. W. Seeb, and J. E. Seeb. 2005a. Use of sequence data from rainbow trout and At- lantic salmon for SNP detection in Pacific salmon. Mol. Ecol. 14:4193-4203. Smith, C. T., W. D. Templin, J. E. Seeb, and L. W. Seeb. 2005b. Use of the 5'-nuclease reaction for single nu- cleotide polymorphism genotyping in Chinook salm- on. Trans. Am. Fish. Soc. 134:207-217. Smith, M. J., C. E. Pascal, Z. Grauvogel, C. Habicht, J. E. Seeb, and L. W. Seeb. 2011. Multiplex preamplification PCR and microsatellite validation enables accurate single nucleotide polymor- phism genotyping of historical fish scales. Mol. Ecol. Resour. 11 (suppl. sl):268-277. Smouse, P. E., R. S. Waples, and J. A. Tworek. 1990. A genetic mixture analysis for use with incom- plete source population data. Can. J. Fish. Aquat. Sci. 47:620-634. Taylor, E. B. 1991. A review of local adaptation in Salmonidae, with particular reference to Pacific and Atlantic salm- on. Aquaculture 98:185-207. Templin, W. D., J. E. Seeb, J. R. Jasper, A. W. Barclay, and L W. Seeb. 2011. Genetic differentiation of Alaska Chinook salmon: the missing link for migratory studies. Mol. Ecol. Re- sour. 11 (suppl. sl):226-246. Tessier, N., L. Bernatchez, P. Presa, and B. Angers. 1995. Gene diversity analysis of mitochondrial DNA, microsatellites and allozymes in landlocked Atlantic salmon. J. Fish Biol. 47:156-163. Utter, F. M., G. B. Milner, G. Stahl, and D. J. Teel. 1989. Genetic population structure of Chinook salm- on, Oncorhynchus tshawytscha, in the Pacific North- west. Fish. Bull. 87:239-264. Vander Haegen, G. E., H. L. Blankenship, A. Hoffmann, and D. A. Thompson. 2005. The effects of adipose fin clipping and coded wire tagging on the survival and growth of spring Chinook salmon. N. Am. J. Fish. Manage. 25:1161-1170. Williamson, K. S., and B. May 2005. Homogenization of fall-run Chinook salmon gene pools in the Central Valley of California, USA. N. Am. J. Fish. Manage. 25:993-1009. Waples, R. S., D. J. Teel, J. M. Myers, and A. R. Marshall 2004. Life-history divergence in Chinook salmon: his- toric contingency and parallel evolution. Evolution 58:386-403. Weir, B. S., and C. C. Cockerham. 1984. Estimating F-statistics for the analysis of popula- tion structure. Evolution 38:1358-1370. Wood, C. C., S. McKinnell, T. Mulligan, and D. Fournier. 1987. Stock Identification with the maximum-likelihood mixture model: sensitivity analysis and application to complex problems. Can. J. Fish. Aquat. Sci. 44:866-881. 131 Effect of demography on spatial distribution: movement patterns of the Albemarle Sound- Roanoke River stock of Striped Bass (Aforone saxatilis ) in relation to their recovery Abstract— We analyzed tag returns from a long-term tagging program to evaluate the movement patterns of the Albemarle Sound-Roanoke River (AR) stock of Striped Bass (Morone saxati- lis) during a period of stock recovery in 1991-2008. The AR stock was found to increase its movement outside the Albemarle Sound estuary (from <4% to 15-31%) as it recovered from 1991 to 2008. Analysis with multinomial lo- gistic regression where recapture area was modeled as a function of fish size and stock abundance indicated that Striped Bass from the AR stock exhibit a strong size-dependent emigration pattern. Larger (older) adults >600 mm in total length (TL) were much more likely to emigrate to ocean habitats (after spawning) than were smaller adults (350-600 mm TL), which mostly remained in inshore estuarine habi- tats. Smaller adults showed evidence of density-dependent movement and were recaptured only in adjacent es- tuarine systems, the Pamlico Sound and lower Chesapeake Bay, during periods of increased stock abundance. Assessment and management strate- gies for the AR stock of Striped Bass could be improved by accounting for movement (and hence harvest) outside the currently assumed stock bound- ary. More broadly, this study illustrates that changes in the demographics, such as size structure and total abundance, within fish populations can result in major shifts in their distribution and that long-term tagging data are useful in detection of such population-level changes in movement patterns. Manuscript submitted 18 April 2013. Manuscript accepted 12 February 2014. Fish. Bull. 112:131-143 (2014). doi:10.7755/FB. 112.2-3.3 The views and opinions expressed or implied in this article are those of the author (or authors) and do not necessarily reflect the position of the National Marine Fisheries Service, NOAA. Jody L. Callihan (contact author)1 Charlton H. Godwin2 Jeffrey A. Buckel3 Email address for contact author: ilcallih@ncsu.edu 1 Department of Applied Ecology North Carolina State University Campus Box 7617 Raleigh, North Carolina 27695 2 Northern District Office North Carolina Division of Marine Fisheries 1367 US 17 South Elizabeth City, North Carolina 27909 3 Center for Marine Sciences and Technology Department of Applied Ecology North Carolina State University 303 College Circle Morehead City, North Carolina 28557 The demographics of fish popula- tions can be important in shaping their movement patterns. Numerous species have been shown to increase their distributional range or move- ment distances as population abun- dance increases (Swain and Wade, 1993; Brodie et al., 1998; Overholtz, 2002; Abesamis and Russ, 2005; Dun- ning et al., 2006), a response pre- sumably due to density-dependent mechanisms (e.g., intraspecific com- petition for food or the saturation of optimal habitats) (MacCall, 1990). In addition, changes in movement patterns with ontogenetic changes in fish are common because habitat requirements change as species age (Werner and Gilliam, 1984; Dahlgren and Eggleston, 2000). The demographics of fish popu- lations are continually shifting for reasons that include changes in fish- ing pressure and the natural envi- ronment (e.g., recruitment varia- tion) that can alter age structure and abundance (Longhurst, 2002; Berkeley et al., 2004; Hutchings and Baum, 2005) and in turn cause pop- ulation-level changes in movement patterns. Understanding if and how population-level movements (and distribution) change over time is of particular importance for exploited fishery species because such changes can pose challenges for assessment and management techniques, for which stock boundaries are often as- sumed to be static and not dynamic (Winters and Wheeler, 1985; Ham- mer and Zimmermann, 2005; Link et al., 2011). Striped Bass ( Morone saxatilis) occur throughout the East Coast of the United States and have sup- ported important fisheries there for centuries (Merriman, 1941). Tag- ging studies clearly have shown that spawning populations (or stocks) of Striped Bass in the mid-Atlantic re- gion, which includes the Hudson Riv- er, Delaware River, and Chesapeake Bay, generally exhibit an anadro- mous life-history strategy and un- dergo extensive seasonal migrations. After spawning in the freshwater portion of their respective estuaries, many adults emigrate to Atlantic 132 Ocean waters from New Jersey to Maine in early sum- mer, move south in the fall to overwintering habitats in coastal waters from New Jersey to Cape Lookout in North Carolina, then return to their natal estuary in subsequent springs to spawn (Boreman and Lewis, 1987; Waldman et ah, 1990; Dorazio et al., 1994; Welsh et ah, 2007). In contrast, the Albemarle Sound-Roanoke River (AR) stock of Striped Bass, hereafter referred to as the “AR stock,” has historically been viewed as a nonmigratory stock, and most fish are believed to re- main in their natal estuarine system, the Albemarle Sound estuary, throughout their lives (Merriman, 1941; Hassler et al.1). Indeed, in the most extensive tagging study to date on the AR stock by Hassler et al.1, virtu- ally all (99%) of the 2428 returns of the 9220 adults tagged in the Roanoke River during the springs of 1959-77, occurred within the Albemarle Sound estu- ary. The few returns that occurred outside Albemarle Sound (<1% of the total) were from an adjacent estu- ary (Pamlico Sound); remarkably, no returns were from ocean waters (Hassler et al.1). These differences in migration patterns may have been due to differences in life-history strategies (non- anadromous vs. anadromous) between the AR stock and more northerly stocks, or it could have been a re- sult of a historic lack of larger, older fish (>600 mm in total length [TL] ) in the AR stock because of high har- vest levels. Differences in life-history strategy would be perplexing given that these stocks occur in the same zoogeographic province (mid-Atlantic coast of the Unit- ed States) and given that some of them are in close latitudinal proximity (e.g., the AR and Chesapeake Bay stocks). In 1988, the North Carolina Division of Ma- rine Fisheries (NCDMF) began a cooperative tagging program with the North Carolina Wildlife Resources Commission (NCWRC) to address this question and to further investigate the migration dynamics of the AR stock of Striped Bass. Much of the past work of tagging individuals from the AR stock was done when Striped Bass were at low levels of abundance and overfished (NCDMF and NCWRC2). In more recent years (1991-2008), the AR stock, as well as the Chesapeake Bay stock (Richards and Rago, 1999), made a dramatic recovery from their 1 Hassler, W. W., N. L. Hill, and J. T. Brown. 1981. The sta- tus and abundance of striped bass, Morone saxatilis, in the Roanoke River and Albemarle Sound, North Carolina, 1956- 1980. North Carolina Department of Natural Resources and Community Development, Division of Marine Fisheries, Special Scientific Report 38, 156 p. [Available from the Di- vision of Marine Fisheries, 3441 Arendell St., Morehead City, NC 28557. 2 NCDMF (North Carolina Division of Marine Fisheries) and NCWRC (North Carolina Wildlife Resources Commis- sion). 2013. Amendment I to the North Carolina Estuarine Striped Bass Fishery Management Plan, 420 p + appendices. North Carolina Division of Marine Fisheries, North Carolina Department of Environment and Natural Resources, More- head City, NC. [Available from http://p0rtal.ncdenr.0rg/c/ document_library/get_file?uuid=d3fdf967-82d5-4653-8b79- 20247c5ed5ad&groupld=38337, accessed January 2014.] Fishery Bulletin 112(2-3) depleted state in the late 1970s and 1980s. The esti- mated total abundance of the AR stock nearly doubled during the 1990s, increasing from 1.0 to 1.9 million fish, and remained at high levels (>1.8 million fish) throughout the 2000s (NCDMF and NCWRC2). In ad- dition, the age and size structure of the stock expanded as larger (>600 mm TL) and older (age 9+) fish became more prevalent as the stock recovered (NCDMF and NCWRC2). The recovery of the AR stock was a result of a combination of factors, namely more stringent fish- ing regulations that increased development to older age classes and improvements in environmental condi- tions that enhanced spawning habitat and recruitment of young Striped Bass (e.g., regulated river flows that were more conducive for the transport and survival of eggs and larvae) (Rulifson and Manooch, 1990; NCDMF and NCWRC2). For this study, we first addressed the following ques- tion: Have Striped Bass of the AR stock increased their movement outside of the Albemarle Sound estuary since population rebuilding in the 1990s? After showing that the movement of the AR stock out of the estuary has indeed increased, we then related recapture locations of tagged individuals to both fish size and total annual stock abundance (density) in an effort to explain this increase in emigration over the past 2 decades (1991- 2008). Lastly, we discuss the management implications of this increased movement given the stock is currently considered to be resident. Materials and methods j Fish tagging During the springs of 1991-2008, 42,534 adult Striped Bass from the AR stock (mostly >350 mm TL; Fig. 1) were tagged and released on their well-described spawning grounds (Hassler et al.1) -200 km upstream of the mouth of the Roanoke River in North Carolina (Fig. 2A). During weekly sampling events throughout April and May, Striped Bass were collected with an electrofishing boat and transported to a tagging vessel, where they were held in a “live well” until processing. Fish in good condition were measured (TL to the near- est millimeter), weighed (to the nearest gram), and sex was determined by expression of gonadal products. The fish were then tagged just above the posterior tip of the pelvic fin with a Floy (model FM-843) internal anchor tag (Floy Tag, Inc., Seattle, WA). Fish were immedi- ately released after tagging. The streamer of the tags indicated a “reward” (US $5 or a baseball cap) would be offered for reporting information on recaptured Striped Bass (e.g., recovery date and location, and tag number) 3 Mention of trade names of commercial companies is for iden- tification purposes only and does not imply endorsement by the National Marine Fisheries Service, NOAA. Call Ihan et al: Effect of demography on the spatial distribution of the Albemarle Sound-Roanoke River stock of Morone saxatilis 133 Females Males 50 40 30 20 10 0 Total length (mm) Figure 1 Size distributions of tagged Striped Bass ( Morone saxatilis) by time period and sex. Fish were collected by electrofishing during spring in the Roanoke River. Note that sex was determined for nearly all (>99%) tagged fish. Data from 1994 were excluded because few fish (n= 9) were tagged that year. 1991-93 (/?= 1331) 50 1 40 30 20 10 il 1991-93 (n=4981) to the NCDMF, whose contact information was printed on the tag. Data analysis We used multinomial logistic regression to evaluate the effects of fish size and stock abundance on the recapture location (i.e., to evaluate the movements) of Striped Bass of the AR stock. For this analysis, recapture locations of tagged fish were assigned to 1 of 4 broad geographic areas: 1) the Albemarle Sound estuary, 2) the Pamlico Sound estuary, 3) ocean wa- ters of North Carolina, or 4) northern coastal waters from Virginia to Massachusetts (Fig. 2B). Therefore, recapture area constituted a multicategory response variable. Explanatory variables were fish size (TL at tagging) and total annual abundance of the AR stock (1991-2008). Annual abundance estimates (of age 1+ fish) were obtained from a statistical catch-at-age model from the most recent AR stock assessment (NCDMF and NCWRC2) and served as a proxy for the annual densities of conspecifics (AR stock only) with which tagged Striped Bass were expected to interact each year. Sex was not included as an explanatory variable because it was confounded with fish size be- yond 800 mm TL because all but 4 tag returns from this size range were from females. However, across smaller sizes (400—800 mm TL), over which sexes were more equally represented, similar-size males and fe- males were generally recaptured in the same areas, indicating that movements differed little between sexes. For the purpose of our analyses, we included only tag returns that occurred after the first 2 weeks but within the first calendar year at liberty. By restrict- ing returns to those returns that occurred within the first calendar year at liberty (on or before 31 Decem- ber), movement between tagging and recapture loca- 134 Fishery Bulletin 112(2-3) (A) Capture and release location (represented by the star in the upper Roanoke River) of tagged Striped Bass ( Morone saxa- tilis) during the period of 1991-2008 and reference map for waterbodies in coastal North Carolina. (B) Geographic areas of recapture used in data analyses: 1) Albemarle Sound estuary (area shaded in gray), 2) Pamlico Sound estuary (area shaded in black), 3) North Carolina ocean waters (box 3), and 4) northern coastal waters (box 4). tions could be known to occur during a given year. This restriction allowed movements (recapture area) to be directly related to stock abundance, which was estimated on an annual basis (i.e., each calendar year) from 1991 to 2008, the terminal year in the as- sessment. In addition, restriction of returns to a rela- tively short time period at liberty (<9 months) mini- mized the opportunity for growth between tagging and recapture, thereby ensuring that fish lengths at tagging (the size variable used in our analyses) were representative of the size of fish when movement occurred. To reach another recapture area (outside the Albe- marle Sound estuary), tagged fish would have had to travel a considerable distance (>300 km) from their release site in the upper Roanoke River. Therefore, to reduce the likelihood of underestimation of fish move- ment, we excluded tag returns from the first 14 days at liberty, affording tagged fish a more realistic period of time to complete movement or migration to another system. Indeed, the earliest tag return from outside the Albemarle Sound estuary (in North Carolina ocean waters) occurred at 16 days after tagging, providing justification for our 14-day exclusion window. Finally, data from 1994 were excluded from analyses because of reduced tagging efforts in that year (only 9 fish were tagged and 1 returned). To determine which explanatory variables affected movements of Striped Bass and to assess their rela- tive importance, we used an information-theoretic ap- proach. A multinomial logistic regression model was run for each of the 5 possible combinations of explana- tory variables: 1) length, abundance, lengthxabundance (interaction model), 2) length and abundance, 3) length only, 4) abundance only, and 5) intercept only (no ef- fects model). Akaike’s information criterion (AIC) val- ues were obtained for each candidate model. We con- sidered the model with the lowest AIC value as the most parsimonious or “best,” but we also computed ad- ditional diagnostics, Akaike differences (A;) and Akaike weights (w{), to assess how other models performed in comparison to this single best model (Burnham and Anderson, 2002). The first of these other diagnostics was calculated as Callihan et al.: Effect of demography on the spatial distribution of the Albemarle Sound-Roanoke River stock of Morone saxatilis 135 A; = AIC ; - AICmin , where AIC \ - the AIC value of a given model (/); and AlCmin = the AIC value of the best model (mini- mum AIC). As a general guideline, models with A; close to zero have considerable empirical support, models with A; of 4-7 have much less support, and models with A; of 9-14 have little support (Anderson, 2008). The follow- ing equation was used to calculate uq values: expl-^Ai) wi = r— i v Note that R refers to the set of models being evaluated. Values of uq can be interpreted as the probability that a particular model (i) is the best model for the data set given that one of the models must be selected as the best (Anderson, 2008). We based our inferences on parameter estimates from the model (i.e., on the combination of explanatory variables) deemed most parsimonious from AIC diag- nostics. For example, if the third model (length effect only) was determined to be the best model, values of the parameters (i.e., regression coefficients) that repre- sented the effect of fish length were used to calculate the predicted relative probability of Striped Bass being recovered in each recapture area as a function of their size at tagging. We assessed the fit of the best model through the use of both Pearson and deviance good- ness-of-fit tests. Because explanatory variables were continuous, it was necessary to group data for these tests (Agresti, 1996). For this purpose, we used 100-mm bins and abundance bins of 0.1 million fish. All statisti- cal analyses were performed in SAS, vers. 9.1.3 (SAS Institute, Inc., Cary, NC) with a. =0.05. Results Tag return summary From 1991 to 2008, 1197 tagged Striped Bass were re- ported as having been recaptured within their first 9 months at liberty (late April-December); analyses con- ducted for this study were based on data from these individuals. Hook-and-line (recreational) anglers ac- counted for a majority (84%) of tag returns. Although most returns (80%) were from fish 400-600 mm TL (at tagging), fish lengths ranged from 287 to 1105 mm TL. Moreover, nearly all tag returns (154 of 156) of larger Striped Bass (>600 mm TL) were from years in which stock abundance exceeded 1.5 million fish (Table 1). Temporal recapture trends The AR stock of Striped Bass increased their move- ment outside of the Albemarle Sound estuary as the TabSe 1 Number of tag returns of Striped Bass ( Morone saxatilis) per combination of total annual stock abundance (millions of fish) and interval of total length (TL) at tagging. Annual abundance estimates (1991-2008) of Albemarle Sound-Roanoke River Striped Bass were obtained from a statistical catch-at-age model (NCDMF and NCWRC2). Only those tag returns occurring after the first 2 weeks but within the first calendar year at liberty were included in data analyses and are enu- merated here. “-”=no tag returns for that year. Abundance (millions) Year Number of returns per size (mm TL) interval <400 400-499 500-599 600-699 700-799 800-899 900-999 >1000 1.035 1993 — 37 15 — — _ _ _ 1.101 1991 10 17 5 - - - - 1.104 1992 2 48 7 1 1 - - - 1.388 1995 - 4 17 - - - - _ 1.518 2005 1 68 36 7 2 - 6 1.569 1996 4 7 13 1 - - — - 1.673 2004 - 17 8 - 1 3 1 1.752 1997 6 38 28 4 1 - - - 1.803 2006 - 54 80 6 5 3 9 3 1.828 2003 7 20 36 7 2 1 3 1.829 2001 1 35 38 12 2 2 — _ 1.836 2000 1 29 16 4 2 - - - 1.860 2002 2 27 39 4 7 1 1 1.877 1998 4 41 23 7 3 - - - 1.895 2008 38 42 13 3 - 3 5 6 1.907 1999 1 29 17 5 1 - — — 2.051 2007 — 19 40 11 3 1 5 2 136 Fishery Bulletin 112(2-3) Figure 3 Time series for the period of 1991-2008 of the following trends of the Albemarle Sound-Roanoke River stock of Striped Bass ( Morone saxatilis ): 1) total annual abundance (millions of fish; gray bars) of the stock (age 1+ fish) estimated from a statistical catch-at-age model (NCDMF and NCWRC2), 2) annual percentage of tag returns (solid line) that occurred outside the Albemarle Sound estuary from Striped Bass tagged and released on the spawning grounds in the upper Roanoke River, and 3) catch per unit of effort for fish (number h-1) age 9+ (>700 mm in total length) (dashed line) in annual spring electrofishing surveys on the Roanoke River spawning grounds. population rebuilt over the past 2 decades (1991-2008). In the early 1990s, few tag returns occurred outside the Albemarle Sound estuary: <4% annually across the years of 1991-96, with the exception of 1995 (Fig. 3). However, as the stock increased in abundance and its age structure expanded, returns from regions outside the Albemarle Sound estuary increased considerably and ranged from 15% to 31% annually during the years of 1997-2008 (Fig. 3). Effects of fish size and stock abundance on recapture area Fish size and stock abundance affected recapture area. The best multinomial logistic regression model in- cluded the main effects of both fish length and stock abundance but not their interaction (Table 2). Good- ness-of-fit tests indicated this model fitted the sample data well (Pearson goodness of fit, x2=H7, degrees of freedom=126, P=0.70; deviance goodness of fit, %2=104, degrees of freedom=126, P=0.93). Although the best model showed that recapture area depended on both fish size and stock abundance, AIC diagnostics across the suite of models indicated that fish length exerted a much stronger effect than abundance. Specifically, the model that included only fish length had moderate empirical support (A;=3.8, mi=0.12), but the model that included stock abundance alone had very little support (Ai=462.2, h^O) (Table 2). Striped Bass of the AR stock exhibit a strong size- dependent migration pattern, whereby both the inci- dence of emigration and the distance emigrants move increase with fish size. The best model predicted that the probability of emigration from (i.e., recapture outside) the Albemarle Sound estuary increased dra- matically with fish size. Specifically, the probability of recapture within Albemarle Sound declined sharp- ly (from values >90%) beyond 600 mm TL, the size at which recapture probabilities began to increase in other areas, such as Pamlico Sound and ocean waters (Fig. 4). The model predicted that Striped Bass 700- 800 mm TL in length were most likely to be recap- tured in ocean waters of North Carolina (Fig. 40 and that the largest fish (>850 mm TL) were most likely to be recaptured in the northern coastal region (Fig. 4D). Empirical tag return data supported the move- ment pattern indicated by the best model. Nearly all (92%) of the tag returns of smaller fish (<600 mm TL; Callihan et al.: Effect of demography on the spatial distribution of the Albemarle Sound-Roanoke River stock of Morone saxatilis 137 Table 2 Diagnostics with Akaike’s information criterion (AIC) for candidate multinomial lo- gistic regression models that relate the recapture area (Albemarle Sound estuary, Pamlico Sound estuary. North Carolina ocean waters, or northern coastal waters from Virginia to Massachusetts) of tagged Striped Bass (Morone saxatilis) to fish length and total annual stock abundance for the years 1991-2008. Each model rep- resents a different combination of these explanatory variables. Note that A;=Akaike’s differences and tr^Akaike’s weights, where lower values of A; and higher values of W[ indicate greater relative empirical support for a model. Model AIC A; Wi Length + abundance 1003.2 0.0 0.80 Length only 1007.0 3.8 0.12 Length + abundance + length x abundance 1007.8 4.6 0.08 Abundance only 1465.4 462.2 0.0 Intercept only 1496.1 492.9 0.0 71 = 1040) occurred within the Albemarle Sound estu- ary (Fig. 5A). Yet, only 47% of returns of fish 600-799 mm TL (?i= 102) and 2% of returns of fish >800 mm TL (ii=55) occurred in Albemarle Sound; most tag returns of these larger fish occurred in ocean waters (Fig. 5, B and C). Interestingly, the majority (78%) of tag re- turns of the largest fish in this study (800-1105 mm TL) occurred in distant coastal waters from New Jer- sey to Cape Cod, 780 to 1250 km from the release site (Fig. 50. Stock abundance also affected the areas in which Striped Bass were recaptured. The best model pre- dicted a slight increase ( ~ 5%) in recapture of small Striped Bass (<600 mm TL) in the Pamlico Sound re- gion as stock abundance increased from 1 to 2 million fish (Fig. 4B). This trend also was evident in empirical tag return data. Returns from the Pamlico Sound es- tuary, -6% of all returns, occurred only during years in which stock abundance exceeded 1.4 million fish. There were no returns from the Pamlico Sound estu- ary during years of lower abundance (1. 0-1.1 million fish) (Fig. 6). Discussion Continuous tagging over a 20-year period, a length of effort that is rare in most fisheries, allowed us to determine the strong effect of fish size and relative- ly smaller effect of stock abundance on a fish stock’s spatial distribution. Multiple stocks of Striped Bass co-occur along the East Coast of the United States during nonspawning periods. Therefore, by tagging fish on their natal spawning grounds (when stocks are separated), we were able to investigate stock-specific movements and spatial distribution — information that could otherwise not have been resolved with approach- es such as fisheries-independent surveys (e.g., trawl surveys). In this section, we provide further details on the effects of fish size and stock abundance on the spatial distribution of the AR stock of Striped Bass and on the implications for management of Striped Bass. Effects of fish size on recapture area The increase in tag returns of the AR stock from re- gions outside its natal estuary over the past 2 decades was largely due to expansion of the age and size struc- ture of the stock as it recovered. The majority of re- turns (67%) that occurred outside the Albemarle Sound estuary during the stock recovery period were from ocean waters. Model results and empirical data both showed the probability of Striped Bass being recap- tured in ocean waters increased dramatically with fish size beyond 600 mm TL, to the point where the larg- est individuals (>800 mm TL) were almost exclusively captured in ocean waters. Therefore, it is not surpris- ing that returns from ocean waters increased over the past 2 decades as more fish from this largest size class (which was the class most likely to emigrate to ocean habitats) became available for tagging and recapture as the age and size structure of the AR stock expanded. The strong size-dependent emigration pattern of Striped Bass revealed by this study helps explain the lack of recaptures in ocean waters by Hassler et al.1, who also focused on the AR stock. To collect fish for tagging, Hassler et al.1 primarily used small- mesh (<150 mm stretched) gill nets that likely se- lected for smaller fish. Indeed, of the 2428 returns in their study, most (86%) were from fish 400-550 mm TL at tagging, and only 2 returns (<0.1%) were from fish >800 mm TL at tagging. Moreover, the vast ma- jority (88%) of tag returns in their study occurred within the first year at liberty. Therefore, given the small sizes of tagged fish and short-term nature of returns ( i . e . , small tagged fish did not have time to grow into larger size categories because of high har- 138 Fishery Bulletin 112(2-3) _Q 03 n o TL (mm) B q 0.8 TL (mm) TL (mm) Figure 4 TL (mm) Predicted probabilities of tag returns in each recapture area as a function of total length (TL) and annual total stock abun- dance (millions of fish) of Striped Bass (Morone saxatilis ) during the period of 1991-2008. We used the following 4 recapture areas (A) Albemarle Sound estuary, (B) Pamlico Sound estuary, (C) North Carolina ocean waters, and (D) northern coastal waters (for locations, see the map in Fig. 2B). Probabilities are based on parameter estimates from the most parsimonious multinomial logistic regression model that related the recapture area of Striped Bass to TL and stock abundance. Cooler and warmer colors represent low and high tag return probabilities, respectively, as follows: (0.0, H; 0.2, H; 0.4, 9; 0.6, D; 0.8, H; 1.0, ■). Note that tag return probabilities sum to 1.0 (across recapture areas) for a given combination of TL and stock abundance. vest), the lack of ocean recaptures by Hassler et al.1 is not surprising. Nearly all fish recaptured in their study (>99%) were smaller than the size at which ap- preciable ocean emigration occurs (>800 mm TL), as indicated in our study. Although other factors, such as prey availability and susceptibility to predation, may be involved, wa- ter temperature appears to be a salient factor in ex- planation of the size-dependent migration and distri- bution patterns of the AR stock. A change in tempera- ture preferences with fish size has been hypothesized to be the main driver of the size-dependent emigration pattern observed previously for other stocks of Striped Bass (Coutant, 1985), especially the Chesapeake stock (Dorazio et al., 1994; Secor and Piccoli, 2007). Decreases in temperature optima with fish size can be explained by bioenergetic principles. Specifically, the temperature threshold beyond which the increase in to- tal metabolic load starts to become stressful (i.e. , the point at which the scope for activity and growth begins Callihan et al.: Effect of demography on the spatial distribution of the Albemarle Sound-Roanoke River stock of Morone saxatilis 139 Tag return locations of Striped Bass ( Morone saxatilis ) along the eastern seaboard of the United States by length group (data pooled across years): (A) fish 287-599 mm in total length (TL) (n = 1020 returns), (B) fish 600-799 mm TL («=101 returns), and (C) fish 800-1105 mm TL (n= 55 returns). Bubble sizes represent the number of tag returns from each location (within each length group). The star in panel A denotes the location where Striped Bass were tagged and released during annual spring electrofishing surveys conducted in the Roanoke River in 1991-2008. Only those tag returns that occurred after the first 2 weeks but within the first calendar year at liberty were included in analyses and are shown. The location of 21 tag returns (of the 1197 total) could be assigned only to 1 of the 4 broad geographic recapture areas (shown in Fig. 2B) and are, therefore, not shown. to decline) occurs at progressively lower temperatures as fish size increases because larger individuals have a greater total metabolic demand than smaller indi- viduals on the basis of body size alone (Hartman and Brandt, 1995). Therefore, after spawning, most large Striped Bass may emigrate, as we found, to cooler northern ocean habitats, which would provide a met- abolic reprieve, rather than spend their summers in warm estuarine waters. Interestingly, Striped Bass of the AR stock in the in- termediate size range of 700-850 mm TL, were mainly recaptured in ocean waters off North Carolina, from the Oregon Inlet north to the border of North Carolina and Virginia. No Striped Bass were recaptured in ocean waters south of Cape Hatteras, where summer temper- atures (>26°C; http://www.ndbc.noaa.gov, Station#41036) are similar to summer temperatures in Albemarle Sound. Therefore, nearby ocean waters may provide an adequate thermal refuge (23-26°C; http://www.ndbc.noaa. gov, Station#44100) during summer for Striped Bass in the size range of 700-850 mm TL. One intriguing ques- tion is whether the size at which the onset of ocean emigration occurs will shift to a smaller size as inshore estuarine waters, which already approach 30°C in sum- mer (http://waterdata.usgs.gov, Gage#0208114150), are ex- pected to continue warming under current projections for climate change (IPCC, 2007). Continuation of the long-term tagging program on the AR stock of Striped Bass could help address this question. Previous research on northern stocks of Striped Bass has provided evidence for diverse lifetime migration patterns: some members of a given population reside in freshwater or estuarine environments throughout their life (resident contingent) and others are more explor- atory and engage in large-scale coastal migrations (mi- gratory contingent) (Clark, 1968; Secor, 1999). There is particularly strong evidence for this “contingent” be- havior in Striped Bass in the Hudson River (Secor and Piccoli, 1996; Secor et al., 2001; Zlokovitz et al., 2003). Our study, however, provides little indication of this phenomenon in the AR stock of Striped Bass. If con- tingent behavior had been prevalent, one would have expected that some large fish would have remained and been recaptured in the Albemarle Sound after spawn- ing. Yet, of the 50 fish exceeding 855 mm TL that were recovered in our study, none were recaptured within Albemarle Sound and, instead, all were taken in the ocean. It is possible that contingent behavior is not 140 Fishery Bulletin 112(2-3) Figure 6 Tag return locations of Striped Bass ( Morone saxatilis) <600 mm in total length in North Carolina and Virginia coastal waters by stock abundance in the year of release: (A) annual abundance values of 1.0-1. 1 million fish (n = 138 returns), (B) annual abun- dance values of 1.4-1. 7 million fish (n=169 returns), and (C) annual abundance values of 1. 8-2.0 million fish (n= 713 returns). Bubble sizes represent the number of tag returns from each location (within each abundance group) as indicated in the legend. The star in panel A denotes the location where Striped Bass were tagged and released during annual spring electrofishing surveys in the Roanoke River in 1991-2008. Only those tag returns that occurred after the first 2 weeks but within the end of the first calendar year at liberty were included in analyses and are shown. The location of 20 tag returns (of the 1040 total) could be as- signed to only 1 of the 4 broad geographic recapture areas (shown in Fig. 2B) and are, therefore, not shown. beneficial, and, therefore, it does not manifest in the AR stock because of high inshore water temperatures during summer that would be unsuitable for “resident” fish once they attain a large size. The possibility for latitudinal differences in the frequency of contingent behavior in Striped Bass and other fishes warrants fu- ture investigation. Effects of stock abundance on recapture area Stock abundance in the year of release was includ- ed in the best model explaining where Striped Bass were recaptured. This effect was primarily a result of smaller Striped Bass being recaptured in the ad- jacent estuarine systems of Pamlico Sound and lower Chesapeake Bay only in the years of highest abun- dance (Fig. 60. Also, evidence of recapture patterns within the Albemarle Sound estuary were indicative of a density effect. Namely, tag returns were much more common in the eastern portions of Albemarle Sound, particularly in Currituck Sound (6% vs. 1% of returns) and Croatan and Roanoke sounds (32% vs. 6%), during years in which stock abundance exceed- ed 1.4 million fish in contrast to years when it was below this level (Fig. 6). Therefore, although adults generally may remain inshore until they reach larg- er sizes (>600 mm TL), the distances they disperse within estuarine habitats, after spawning, tend to increase with the abundance of conspecifics, presum- ably because of density-dependent mechanisms. These movements likely are important ecologically to prey of Striped Bass because the smallest size groups (<600 mm TL) are the most numerous in this popu- lation (i.e., predation effects may change with stock abundance). Future research should investigate these possibilities and better isolate the effects of density by controlling for environmental covariates, such as the abundance of competitor species and changing habitat suitability, as suggested by Shepherd and Lit- vak (2004). Management implications Results from this study have important implications for the management of Striped Bass along the East Coast of the United States. With current assessment strategies, Striped Bass from the AR stock are assumed not to contribute to the Atlantic Ocean mixed stock Callihan et al.: Effect of demography on the spatial distribution of the Albemarle Sound-Roanoke River stock of Morone saxatilis 141 fishery (ASMFC4). However, this study revealed that some members of the AR stock, those fish surviving to sizes >800 mm TL, are indeed migratory and, there- fore, unequivocally contribute to (i.e., are harvested by) the mixed stock fishery of the Atlantic coast. Because management benchmarks for the mixed stock fishery, such as the threshold fishing mortality (Fmsy=0-41; ASMFC4), currently are based on data from Chesa- peake, Hudson, and Delaware stocks that are poten- tially more productive than the AR stock, it is possible that the mixed stock fishery could affect the AR stock disproportionately. Accordingly, future research should establish the productivity of the AR stock of Striped Bass in relation to other stocks. If the AR stock is found to be less productive, then future work also should de- termine the implementation costs of more stringent fishing regulations in the mixed stock fishery, namely the amount and value of harvest that would be lost from more productive stocks (Chaput, 2004; Crozier et al., 2004; Hilborn et al., 2004). Results from this study also have implications for the assessment and management of Striped Bass within North Carolina. Currently, landings of Striped Bass outside the Albemarle Sound estuary (region 1; Fig. 2B) are not included in the AR stock assessment (NCDMF and NCWRC2). Stock status is based on the estimate of fishing mortality (F,threshold=0-27) for fully recruited Striped Bass of age 4-6 and 400-600 mm TL, a size group for which fish were found in this study to increase their movement to adjacent estuarine sys- tems outside the stock boundary as they increased in abundance. Therefore, by not including fish that move to and are harvested in adjacent systems, the AR stock assessment underestimates fishing mortality. Accord- ingly, future research should examine the sensitivity of fishing mortality estimates from the AR stock assess- ment to additional landings of age-4-6 Striped Bass of AR origin outside the Albemarle Sound estuary. Caveats It is important to note that the analyses in this study indicate the probability of recapture location; move- ments are inferred from these data. Fishermen behav- ior (e.g., spatiotemporal differences in fishing effort or size targeting because of regulations and economic value) can affect and potentially bias tag returns and inferences about movement patterns (Hilborn, 1990; Gillanders et al., 2001). The size-dependent migration pattern that we observed could be due to differences in selectivity between ocean and estuarine fisheries; that is, small tagged fish could have migrated to the ocean but not been caught in the fishery. However, fisheries- 4 ASMFC (Atlantic States Marine Fisheries Commission). 2003. Amendment 6 to the Interstate Fishery Management Plan for Atlantic Striped Bass. Fishery Management Re- port No. 41 of the Atlantic States Marine Fisheries Commis- sion, 63 p. [Available from http://www.asmfc.org/uploads/ file/sbAmendment6.pdf.] independent data indicate that it is predominantly the large Striped Bass of the AR stock that migrate to ocean waters. In a mobile telemetry study, Haeseker et al. (1996) searched the Albemarle Sound during sum- mer (May-August) for the presence of 26 telemetered Striped Bass (all but 1 fish <600 mm TL) that partici- pated in the April Roanoke River spawning run. They relocated 25 (96%) of these fish in the Albemarle Sound at least 1 month after spawning, providing evidence that smaller Striped Bass mostly remain in the estuary after spawning. Furthermore, in an ongoing telemetry study, 163 Striped Bass ranging in length from 445 to 1146 mm TL (mean=580 mm TL) were telemetered in the Roanoke River during spring, beginning in 2011, by Harris and Hightower.5 Most large fish in their study (15 of thel8 individuals >900 mm TL at tagging) have been detected by coastal receiver arrays in Massachu- setts, New York, New Jersey, Delaware, and Virginia, but no smaller individuals have been detected in these northern ocean waters (Harris and Hightower5). Hence, results from these fisheries-independent telemetry studies corroborate the strong size-dependent emigra- tion pattern of the AR stock of Striped Bass that we inferred from tag recaptures in our study. A limitation of our study was that nearly all tag returns (99%) from larger fish (>600 mm TL) occurred during years of higher stock abundance (>1.5 million fish). Therefore, it is possible that the observed ocean emigration of larger fish was due in part to the higher abundance of similar size conspecifics (i.e., density- dependent mechanisms). However, ocean emigration of the AR stock of Striped Bass appears to be a size- dependent phenomenon related to bioenergetics as de- scribed and is probably largely independent of ambi- ent population density or abundance. Two lines of evi- dence support this notion. First, data on large Striped Bass (>600 mm TL) across the more restricted range of annual values of stock abundance (1. 5-2.0 million fish) indicate that density had little effect (an increase <3%) on the probability of large fish being recaptured in ocean waters. Second, just as we found in our study, Dorazio et al. (1994) found a strong size-dependent em- igration pattern for the Chesapeake Bay stock: most fish >800 mm TL were recovered in northern ocean wa- ters from New Jersey to Maine. Their study occurred in 1988-91, years when the Chesapeake Bay stock was at relatively low abundance levels and still rebuilding, demonstrating that substantial ocean emigration of large fish, albeit from a different stock, still occurs at low densities. 5 Harris, J. E., and J. E. Hightower. 2013. Unpubl. data. North Carolina Cooperative Fish and Wildlife Research Unit, U. S. Geological Survey, and Department of Applied Ecology, North Carolina State Univ., Raleigh, NC 27695. 142 Fishery Bulletin 112(2-3) Conclusions Our study revealed major changes in the movements and associated distribution of a fish stock as it recov- ered from a depleted state. During the early phases of rebuilding, the stock was largely confined to its na- tal estuary but dramatically expanded its distribution, and degree of anadromy, as recovery continued. This major shift in distribution was due to changes in the demographics — namely size structure and total abun- dance— of the stock as it recovered. Size structure has received little attention in the fisheries literature in regard to its effects on stock distribution but appears to be important. Although the recovery of Striped Bass often is re- garded as one of the few success stories in fisheries management (Richards and Rago, 1999), many global fish stocks are either currently experiencing rebuilding or have recently recovered, for example, nearly one- third of the 166 stocks examined worldwide by Worm et al. (2009). It is possible that the spatial dynamics of these and other rebuilding stocks will differ from their depleted state. For instance, as stocks recover and more individuals are allowed to reach larger sizes (e.g., through a reduction in fishing mortality; Berke- ley et al., 2004), the spatial distribution of stocks may shift or expand because larger, older fish often have different migratory behaviors and habitat preferences than smaller, younger individuals (Heifetz and Fujioka, 1991; Macpherson and Duarte, 1991; Shepherd et al., 2006; Griiss et al., 2011). Such changes in the move- ment and distribution of fish populations can have im- portant consequences for stock assessments, as argued previously, and also affect ecosystem dynamics (e.g., as predators move into new areas, they can exert top- down changes in community structure; Casini et al., 2012). Therefore, resource managers should be aware of potential changes in the movement and distribution of recovering fish stocks and account for them accordingly if they manifest. As indicated in our study, long-term tagging and monitoring data are useful for detection of population-level changes in the movement and distri- bution of fishes. Acknowledgments We thank the many individuals from the NCDMF and North Carolina Wildlife Resources Commission who collected and tagged Striped Bass and also compiled and processed tag return data. We also are grateful to the many fishermen who provided tag return informa- tion. Analyses for this study were completed while J. Callihan was a Marine Fisheries Management Fellow supported by NCDMF (Coastal Recreational Fisheries License fund award no. 3210) and North Carolina Sea Grant award E/GS-6. The manuscript was improved by comments from J. Harris, J. Finn, R McGrath, and R. Laney. Literature cited Abesamis, R. A., and G. R Russ. 2005. Density-dependent spillover from a marine re- serve: long-term evidence. Ecol. Appl. 15:1798-1812. Agresti, A. 1996. An introduction to categorical data analysis, 1st ed., 290 p. John Wiley & Sons, New York. Anderson, D. R. 2008. Model based inference in the life sciences, a prim- er on evidence, 2nd ed., 184 p. Springer, New York. Berkeley, S. A., M. A. Hixon, R. J. Larson, and M. S. Love. 2004. Fisheries sustainability via protection of age struc- ture and spatial distribution of fish populations. Fish- eries 29:23-31. Boreman, J., and R. R. Lewis. 1987. Atlantic coastal migration of striped bass. Am. Fish. Soc. Symp. 1:331-339. Brodie, W. B., S. J. Walsh, and D. B. Atkinson. 1998. The effect of stock abundance on range contraction of yellowtail flounder ( Pleuronectes ferruginea) on the Grand Bank of Newfoundland in the Northwest Atlantic from 1975 to 1995. J. Sea Res. 39:139-152. Burnham, K. R, and D. R. Anderson. 2002. Model selection and multimodel inference: a practical information-theoretic approach, 2nd ed., 496 p. Springer- Verlag, New York. Casini, M., T. Blenckner, C. Mollmann, A. Gardmark, M. Lin- degren, M. Llope, G. Kornilovs, M. Plikshs, and N. Stenseth Christian. 2012. Predator transitory spillover induces trophic cas- cades in ecological sinks. Proc. Natl. Acad. Sci. USA. 109:8185-8189. Chaput, G. 2004. Considerations for using spawner reference levels for managing single- and mixed-stock fisheries of Atlan- tic salmon. ICES J. Mar. Sci. 61:1379-1388. Clark, J. 1968. Seasonal movements of striped bass contingents of Long Island Sound and the New York Bight. Trans. Am. Fish. Soc. 97:320-343. Coutant, C. C. 1985. Striped bass, temperature, and dissolved oxygen: a speculative hypothesis for environmental risk. Trans. Am. Fish. Soc. 114:31-61. Crozier, W. W„ P-J. Schon, G. Chaput, E. C. E. Potter, N. O. Maoileidigh, and J. C. MacLean. 2004. Managing Atlantic salmon ( Salmo salar L.) in the mixed stock environment: challenges and consider- ation. ICES J. Mar. Sci. 61:1344-1358. Dahlgren, C. P, and D. B. Eggleston. 2000. Ecological processes underlying ontogenetic habi- tat shifts in a coral reef fish. Ecology 81:2227-2240. Dorazio, R. M., K. A. Hattala, C. B. McCollough, and J. E. Skjeveland. 1994. Tag recovery estimates of migration of striped bass from spawning areas of the Chesapeake Bay. Trans. Am. Fish. Soc. 123:950-963. Dunning, D. J., J. R. Waldman, Q. E. Ross, and M. T. Mattson. 2006. Dispersal of age-2+ striped bass out of the Hudson River. Am. Fish. Soc. Symp. 51:287-294. Gillanders, B. M., D. J. Ferrell, and N. L. Andrew. 2001. Estimates of movement and life-history param- eters of yellowtail kingfish ( Seriola lalandi ): how useful Callihan et al.: Effect of demography on the spatial distribution of the Albemarle Sound-Roanoke River stock of Morone saxatilis 143 are data from a cooperative tagging programme? Mar. Freshw. Res. 52:179-192. Griiss, A., D. M. Kaplan, S. Guenette, C. M. Roberts, and L. W. Botsford. 2011. Consequences of adult and juvenile movement for marine protected areas. Biol. Conserv. 144:692-702. Haeseker, S. L., J. T. Carmichael, J. T., and J. E. Hightower. 1996. Summer distribution and condition of striped bass within Albemarle Sound, North Carolina. Trans. Am. Fish. Soc. 125:690-704. Hammer, C., and C. Zimmermann. 2005. The role of stock identification in formulating fish- ery management advice. In Stock identification meth- ods: applications in fishery science (S. X. Cadrin, K. D. Friedland, and J. R. Waldman, eds.), p. 631-658. Else- vier Academic Press, Burlington, MA. Hartman, K. J., and S. B. Brandt. 1995. Comparative energetic and the development of bioenergetics models for sympatric estuarine pisci- vores. Can. J. Fish. Aquat. Sci. 52:1647-1666. Heifetz, J., and J. T. Fujioka. 1991. Movement dynamics of tagged sablefish in the northeastern Pacific. Fish. Res. 11:355-374. Hilborn, R. 1990. Determination of fish movement patterns from tag recoveries using maximum likelihood estimators. Can. J. Fish. Aquat. Sci. 47:635-643. Hilborn, R., A. E. Punt, and J. Orensanz. 2004. Beyond band-aids in fisheries management: fixing world fisheries. Bull. Mar. Sci. 74:493-507. Hutchings, J. A., and J. K. Baum. 2005. Measuring marine fish biodiversity: temporal changes in abundance, life history, and demography. Phi- los. Trans. R. Soc. Lond., Ser B: Biol. Sci. 360:315-338. IPCC (Intergovernmental Panel on Climate Change). 2007. Climate change 2007: synthesis report. Contri- bution of working groups I, II, and III to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change, 104 p. [Available from http://www. ipcc.ch/publications_and_data/ar4/syr/en/contents.html, accessed October 2012.] Link, J. S., J. A. Nye, and J. A. Hare. 2011. Guidelines for incorporating fish distribution shifts into a fisheries management context. Fish Fish. 12:461-469. Longhurst, A. 2002. Murphy’s law revisited: longevity as a factor in re- cruitment to fish populations. Fish. Res. 56:125-131. MacCall, A. D. 1990. Dynamic geography of marine fish populations, 155 p. Univ. Washington Press, Seattle. Macpherson, E., and C. M. Duarte. 1991. Bathymetric trends in demersal fish size: is there a general relationship? Mar. Ecol. Prog. Ser. 71:103-112. Merriman, D. 1941. Studies on the striped bass ( Roccus saxatilis ) of the Atlantic coast. Fish. Bull. 50:1-77. Overholtz, W. J. 2002. The Gulf of Maine-Georges Bank Atlantic her- ring ( Clupea harengus ): spatial pattern analysis of the collapse and recovery of a large marine fish com- plex. Fish. Res. 57:237-254. Richards, R. A., and P. J. Rago. 1999. A case history of effective fisheries management: Chesapeake Bay Striped Bass. N. Am. J. Fish. Manage. 19:356-375. Rulifson, R. A., and C. S. Manooch III. 1990. Recruitment of juvenile Striped Bass in the Roa- noke River, North Carolina, as related to reservoir dis- charge. N. Am. J. Fish. Manage. 10:397-407. Secor, D. H. 1999. Specifying divergent migrations in the concept of stock: the contingent hypothesis. Fish. Res. 43:13-34. Secor, D. H., and P. M. Piccoli. 1996. Age-and sex-dependent migrations of Striped Bass in the Hudson River as determined by chemical micro- analysis of otoliths. Estuaries 19:778-793. 2007. Oceanic migration rates of upper Chesapeake Bay striped bass ( Morone saxatilis ), determined by otolith microchemical analysis. Fish. Bull. 105:62-73. Secor, D. H., J. R. Rooker, E. Zlokovitz, and V. S. Zdanowicz. 2001. Identification of riverine, estuarine, and coastal contingents of Hudson River striped bass based upon otolith elemental fingerprints. Mar. Ecol. Prog. Ser. 211:245-253. Shepherd, G. R., J. Moser, D. Deuel, and P. Carlsen. 2006. The migration pattern of bluefish (Pomatomus saltatrix) along the Atlantic coast determined from tag recoveries. Fish. Bull. 104:559-570. Shepherd, T. D., and M. K. Litvak. 2004. Density-dependent habitat selection and the ideal free distribution in marine fish spatial dynamics: con- siderations and cautions. Fish Fish. 5:141-152. Swain, D. P, and E. J. Wade. 1993. Density-dependent geographic distribution of At- lantic cod ( Gadus rnorhua ) in the southern Gulf of St. Lawrence. Can. J. Fish. Aquat. Sci. 50:725-733. Waldman, J. R., D. J. Dunning, Q. E. Ross, and M. T. Mattson. 1990. Range dynamics of Hudson River striped bass along the Atlantic coast. Trans. Am. Fish. Soc. 119:910-919. Welsh, S. A., D. R. Smith, R. W. Laney, and R. C. Tipton. 2007. Tag-based estimates of annual fishing mortality of a mixed Atlantic coastal stock of striped bass. Trans. Am. Fish. Soc. 136:34-42. Werner, E. E., and J. F. Gilliam. 1984. The ontogenetic niche shift and species interac- tions in size-structured populations. Annu. Rev. Ecol. Syst. 15:393-426. Winters, G. H., and J. P. Wheeler. 1985. Interaction between stock area, stock abundance, and catchability coefficient. Can. J. Fish. Aquat. Sci. 42:989-998. Worm, B., R. Hilborn, J. K Baum, T. A. Branch, J. S. Collie, C. Costello, M. J. Fogarty, E. A. Fulton, J. A. Hutchings, S. Jennings, O. P. Jensen, H. K. Lotze, P. M. Mace, T. R. Mc- Clanahan, C. Minto, S. R. Palumbi, A. M. Parma, D. Ricard, A. Rosenberg, R. Watson, and D. Zeller. 2009. Rebuilding global fisheries. Science 325:578-585. Zlokovitz, E. R., D. H. Secor, and P. M. Piccoli. 2003. Patterns of migration in Hudson River striped bass as determined by otolith microchemistry. Fish. Res. 63:245-259. Residence time and habitat duration for predators in a small mid-Atlantic estuary Email address for the contact author: john.manderson@noaa.gov Abstract— Residence times of individ- ual fishes should reflect the durations over which habitat resources support survival, metabolic maintenance, and adequate growth. From May to Octo- ber in 2006 and 2007, we measured residencies of ultrasonically tagged age-la- Striped Bass ( Morone saxati- lis ; n= 46), age-0 and age-l+ Bluefish (. Pomatomus saltatrix\ n= 45 and 35) and age-l+ Weakfish (Cynoscion rega- lis ; w=41) in a small estuarine tribu- tary in New Jersey with 32 ultrasonic receivers to monitor movements and sensors to measure habitat resources. Striped Bass and age-l+ Bluefish used the estuary for medians of 9.5 days (d) (max=58 d) and 22 d (max=88 d), and age-0 Bluefish and Weakfish were resident for medians of 30 d (max=52 d) and 41 d (max=88 d), respectively Small individuals <500 mm TL were likely to remain in the estuary longer at warmer temperatures than were large individuals. Size-dependent temperature responses were similar to optimal temperatures for growth reported in previous studies. Freshwa- ter discharge also influenced residence time. All species were likely to remain in the estuary until freshwater dis- charge rates fell to a value associated with the transition of the estuarine state from a partially to fully mixed state. This transition weakens flows into the upstream salt front where prey concentrations usually are high. Time of estuarine residence appeared to be regulated by temperatures that controlled scopes for growth and the indirect effects of freshwater discharge on prey productivity and concentration. Changes in the seasonal phenology of temperature, precipitation, and human water use could alter the durations over which small estuarine tributar- ies serve as suitable habitats. Manuscript submitted 5 April 2013. Manuscript accepted 10 March 2014. Fish. Bull. 112:144-158 (2014). doi:10.7755/FB.112.2-3.4 The views and opinions expressed or implied in this article are those of the author (or authors) and do not necessarily reflect the position of the National Marine Fisheries Service, NOAA. John P. Manderson (contact author) Linda L. Stehlik Jeff Pessutti John Rosendale Beth Phelan Behavioral Ecology Branch Northeast Fisheries Science Center National Marine Fisheries Service, NOAA James J. Howard Marine Sciences Laboratory 74 Magruder Road Highlands, New Jersey 07732 Temperate estuaries serve as spawn- ing, nursery, and feeding habitats for many fishes and invertebrates dur- ing warmer months (Mann, 2000; Able, 2005; Able and Fahay, 2010). Warm temperatures, high nutrient- stimulated primary and secondary productivity, and abundant spatial or structural refuges from preda- tion enhance growth and survival. However, because estuaries are shal- low, semi-enclosed bodies of water along the land-sea boundary, high- frequency atmospheric variability is rapidly translated into variability in biophysical processes that regulate the vital rates of species (e.g., water temperature, freshwater discharge, nutrient inputs, circulation and re- tention, and dissolved oxygen). Estu- arine habitat suitability is, therefore, largely controlled by atmospheric and tidal forcing. As a result, estua- rine habitat suitability is dynamic, and suitable habitats have temporal dimensions of timing and duration that are as important as the spatial dimensions of location and volume (Livingston, 1987; Manderson et ah, 2002; Manderson et ah, 2003; Man- derson et ah, 2006; Peterson et al., 2007). Animals move in variable envi- ronments to fulfill requirements for survival, metabolic maintenance, growth, and reproduction and are believed to “climb” local fitness gra- dients that fall within their percep- tual ranges (Armsworth and Rough- garden, 2005). Individual animals should minimize movement costs by becoming resident in suitable habi- tats until more costly long-distance movements are required by changes in habitat resources, such as tem- perature, oxygen concentrations, and concentrations of predators or prey or by life history event schedules. Changes in atmospheric forcing (e.g., air temperature and precipitation) that change both the timing and per- sistence of suitable shallow coastal habitats should affect the movement costs and energy budgets of the in- dividual animals that use them. Be- cause changes in atmospheric forcing and hydrography are coherent over spatial scales of 100s to 1000s of ki- lometers (Hare and Able, 2007; Man- derson, 2008; Shearman and Lentz, 2010), effects on energy budgets of individual animals are likely to af- fect demographic rates at the popu- lation level. In this study, we used passive ul- trasonic biotelemetry and environ- mental monitoring to measure rela- tionships of residence and egress of 3 predators — Striped Bass (age-l+ Morone saxatilis), Bluefish (age-0 Manderson et al.: Residence time and habitat duration for predators in a small mid-Atlantic estuary 145 and age-l+ Pomatomus saltatrix) and Weakfish (age- 1+ Cynoscion regalis ) — to habitat conditions in a small mid-Atlantic estuarine tributary that serves as a sum- mer feeding and nursery ground. Individuals of these 3 species undertake broad-scale seasonal migrations of 100s to 1000s of kilometers along the Atlantic coast of the United States but can exhibit site fidelity in sum- mer feeding and nursery habitats (Ng et al., 2007; Tay- lor et al., 2007; Pautzke et al., 2010; Turnure, 2010). They occupy upper trophic levels in mid-Atlantic es- tuarine food webs and are responsible for the transfer of nutrients and energy between benthic and pelagic compartments within estuaries and between estuaries and the coastal ocean (Hagy, 2002; Krause et al., 2003; Johnson et al., 2009). We report on the seasonal and size-dependent pat- terns of residency of these predators in a small tribu- tary (surface area of -1000 ha) in New Jersey over 2 years. We use generalized additive mixed models to quantify size-dependent relationships of time of estu- arine residence to water temperature and freshwater discharge. We assume that residency and flux rates of individuals through the estuary reflect the timings and durations when habitat resources support survival, metabolic maintenance, and at least adequate growth, except when emigration is triggered by changes in re- quirements associated with life-history-event schedules (e.g., timing of spawning) (Charnov, 1976; Winkler et al., 1995; Belisle, 2005). Materials and methods Study area We performed acoustic biotelemetry in the Navesink River, New Jersey, a tributary of the Hudson-Raritan Estuary (Fig. 1), described in detail in other stud- ies (Shaheen et al., 2001; Stoner et al., 2001; Scharf et al., 2004; Manderson et al., 2006). The Navesink River is nearly 1.5 km wide and extends -12 km east from its primary freshwater source, the Swim- ming River, to the Shrewsbury River and then to the Hudson-Raritan Estuary where it connects to the At- lantic Ocean. Salinities range from as low as 0.08%c at the head of the Swimming River to -27 %c at the confluence of the Navesink and Shrewsbury rivers. The tidal range averages 1.4 m. Tidal currents are flood dominated and attenuate in the middle and up- per river, an area that is both deeper (mean depth [p D] = 1.5 m mean low water [MLW]; maximum of ~9 m) and has sediments of finer grains than the lower river (p D=1.0 m MLW; maximum of ~6 m) (Chant and Stoner, 2001; Fugate and Chant, 2005). The low- er river has a complex network of channels flanked by sandbars and vegetated coves. Infrastructure of the estuarine observatory Fishes tagged with ultrasonic transmitters were de- tected with an array of omnidirectional receivers (mod- el VR2, VEMCO1 *, Bedford, Canada) moored throughout the Navesink River from May 15 to October 3, 2006, and from April 18 to October 31, 2007 (Fig. 1). We at- tached receivers to anchored lines that had surface and subsurface floats. The subsurface floats suspended the receivers -80 cm above the bottom. In 2006, the array consisted of 27 receivers. In 2007, we moored 5 addition- al receivers in several marsh creeks and coves. Nearest neighbor distances between receivers in the river aver- aged 493 m (standard deviation 141 m, within a range of 216-788 m). On the basis of range tests, receivers moored in the middle and upper river had detection ranges of 350-600 m. Detection ranges were smaller and more variable in the lower river, which is topo- graphically complex. The estuarine volume monitored by the array of all moored receivers was ~1.397xl07 m3 (surface area=932 ha) at MLW. In 2006, the receivers were retrieved in September. We subsequently discov- ered that a few tagged fishes remained in the estuary after the receivers were retrieved. Therefore, in 2007, receivers were left in place for an additional month. We measured environmental variation with moored instruments and supplemental mobile surveys. The moored instruments provided measurements -12 cm above the bottom of the seafloor at 20-min intervals and included 3 Star-Oddi (Gardabaer, Iceland) tem- perature, salinity, and pressure sensors; 3 YSI, Inc. (Yellow Springs, Ohio) temperature, salinity, pressure, and dissolved oxygen sensors; and an Aanderaa RCM 9 (Aanderaa Data Instruments, Bergen, Norway) meter that measured current speed and direction, tempera- ture, salinity, pressure, and optical backscatter. Star- Oddi sensors were used throughout the system (Fig. 1). YSI sensors were deployed in the upper river where episodes of low dissolved oxygen occur. We moored the RCM 9 in the channel that connects the lower and middle rivers. Weekly hydrographic surveys were per- formed from a 6-m vessel through the use of a Hydro- lab DataSonde probe (Hach Hydromet, Loveland, CO) with temperature and salinity sensors mounted 0.5 m below the surface of the water and integrated with a GPS, and a Sea-Bird Electronics, Inc. (Bellevue, WA) SBE 25 Sealogger CTD with temperature, conductivity, pressure, dissolved oxygen, photosynthetically active radiation, turbidity, and fluorometer sensors. During each weekly survey, we performed cross-sectional tran- sects of the river that intercepted all receiver moor- ings. The Hydrolab DataSonde and GPS continuously recorded temperate, salinity, and geographic position at 1-s intervals. Vertical profiles of the water column at each mooring were measured with the conductiv- 1 Mention of trade names or commercial companies is for iden- tification purposes only and does not imply endorsement by the National Marine Fisheries Service, NOAA. 146 Fishery Bulletin 112(2-3) Figure 1 Map of the study area in the Navesink River, New Jersey, on the northeastern coast of the United States and the locations of the 32 moorings with ultrasonic receivers (white circles) and sensors (dark symbols) that measured the physical environment in the study area in which we captured, released, and monitored the movements of tagged Striped Bass (Morone saxatilis), Bluefish (Poma- tomus saltatrix), and Weakfish ( Cynoscion regalis) in 2006 and 2007 for a study of residence times and duration of habitat suitability for these 3 predators. The 5 moorings added in 2007 are indicated by asterisks. A, B and C labels indicate the locations referred to in the text and in the legend for Figure 2. Instruments deployed with receivers included temperature, salinity, pressure, and dis- solved oxygen sensors from YSI, Inc., temperature, salinity, and pressure sensors from Star-Oddi, and an RCM-9 meter from Aanderaa Data Instruments that measured current speed and direction, temperature, salinity, pressure, and optical backscatter. Measurements of freshwater discharge were made at the U.S. Geological Survey (USGS) stream flow station in the Swimming River. ity, temperature, and depth (CTD) sensor. In 2007, we performed additional hydrographic surveys associated with gillnet surveys of predators and prey in the up- per river. Measurements of freshwater discharge (in me- ters per second) from the Swimming River were made at the U.S. Geological Survey stream flow sta- tion (http://nwis.waterdata.usgs.gov/nj/nwis/uv/7site_ no=0 1407500&PARAmeter_cd= 00065,00060). Baro- metric pressures, wind, and air temperatures were measured 7.5 km from the study area at the NOAA weather station in Sandy Hook, New Jersey (http:// www.ndbc.noaa. gov/station_page.php?station=sdhn4). Ultrasonic tagging From May 14 to September 8, 2006 and from May 1 to October 2, 2007, we used hook and line to capture age- 1+ Striped Bass, age-0 and age-l+ Bluefish, and age-l-f- Weakfish as seasonally available within the footprint of the estuarine observatory (Table 1). On the basis of pub- lished age-length relationships, we divided Bluefish into age-0 and age-l+ age classes at a total length of 290 mm (Chiarella and Conover, 1990; Munch and Conover, 2000; Scharf et al., 2004). We transported fishes to the James J. Howard Marine Sciences Laboratory in Highlands, New Jersey, for internal tagging. Fishes were held <8 days (d) in tanks (2.5-m diameter, 0.35-m depth) sup- plied continuously with ambient estuarine water. We anaesthetized fishes with AquiS (AquiS New Zealand, Ltd., Lower Hutt, New Zealand) at a concentration of 54 mg/L. Duration of anesthesia averaged ~3 min. After a fish was anaesthetized, we made an incision 1-2 cm long on its ventral midline and inserted into the body cavity a sterilized, uniquely coded ultrasonic transmitter (V9-6L with a frequency of 69 kHz, rep- etition rate of 40-120 s, dimensions of 9 mmx20 mm, weight of 2 g in water, and minimum battery life of 110 d; VEMCO). We closed incisions with 2 or 3 nylon su- tures (Ethilon 30 and 40 with FS1 cutting needle, Ethi- Manderson et al.: Residence time and habitat duration for predators in a small mid-Atlantic estuary 147 Table 1 Median (Md) total lengths (TL) in millimeters, number of fish released, release dates, median number of detections, and median residence times (in days) of fishes released in 2006 and 2007with ultrasonic transmitters in the Navesink River, New Jersey, for a study of residence times and habitat duration of 3 predators: age-l+ Striped Bass (Morone saxatilis), age-0 and age-l+ Bluefish ( Pomatomus saltatrix), and age-l+ Weakfish ( Cynoscion regalis). Median days detected (i.e., residence time in days) and confidence limits (CL) were calculated with right-censored Kaplan-Meier survival analysis (see Fig. 3). Signifi- cant Spearman’s rank correlation coefficients (p) between release day and body length are shown in bold type. An asterisk (*) denotes that fishes caught by anglers before egress or detected by receivers on the last fall day of the experiment were censored (Striped Bass=l, age-0 Bluefish=4, age-l-i- Bluefish=l, and Weakfish=l). Species-and- age class Year Md TL (Range) Number released Release dates Release date vs. length Spearman’s p, P-value Md number of detections (range) Md days (95%CL) (range) Striped Bass 2006 465 34 15May-28Jun 0.41, 0.016 2475 16(9,28) (359-630) (343-20331) (2-58*) 2007 442 12 3May-19Jun 0.76, 0.004 1469 8(7, °°) (2-50) (342-510) (22-3440) Age-l-i- Bluefish 2006 335 14 5Jun-16Aug 0.45, 0.107 5428 19(16,42) (310-390) (311-17586) (10-48) 2007 455 21 lMay-19Jun -0.83, <0.001 3543 29(20,46) (310-610) (60-21174) (3-88) Age-0 Bluefish 2006 210 15 27Aug-9Sep 0.68, 0.005 2503 29(21,oo) (175-270) (291-7889) (5-37*) 2007 246 30 29Aug-21Sep 0.28, 0.140 1706 29(22,37) (222-275) (101-6777) (1-52*) Weakfish 2006 337 15 13Jul-16Aug 0.32, 0.244 4040 33(22, oo) (224-535) (41-16568) (4-64*) 2007 389 26 29Jun-90ct -0.42, 0.034 1708 47(35,70) (304-500) (31-11391) (6-88*) con, Somerville, NJ). We measured the total length (TL) of each fish in millimeters and inserted unique anchor tags into the dorsal muscle. Fishes recovered from an- aesthesia in <9 min and were monitored for 2-48 hours in flow-through laboratory tanks. We released fishes in good condition at randomly selected locations in the river. This random release approach was used to moni- tor initial patterns of habitat selection during the first 24-48 hours. We released <5 individuals of each age class for each species per week to observe movements over the broadest range of environmental conditions. Striped Bass, Weakfish, and age-l+ Bluefish (n>12 for all classes) implanted with replica transmitters survived >120 d in the laboratory (B. Phelan and J. Rosendale, unpubl. data). Several age-0 Bluefish <170 mm TL died after implantation of replica transmitters. We, therefore, released only Bluefish >175 mm TL with active transmitters in the field. Analyses In this investigation, we analyzed predator residence times in and egress from the estuarine tributary rather than movements within the tributary. We eliminated data from 2 tagged fishes whose movement trajectories indicated that they died shortly after release. Then, we aggregated data collected at all receivers to calculate the presence or absence of each fish in the estuary for each day of observation. Individuals detected in the lower estuary and subsequently not detected for 24 h were considered absent. Several fishes detected at the upstream receiver in the Swimming River disappeared for a short time and then were detected in the Swim- ming River or upper Navesink River. We assumed these fishes had spent that time upstream of the receiver ar- ray and, therefore, had remained in the estuary. We performed all analyses with R software (R Core Team, 2013). We estimated the number of days that tagged fishes used the estuary with right-censored Kaplan-Meier survival analysis (Bennetts et al., 2001). We censored 5 age-0 Bluefish and 1 Striped Bass detected in the estuary when receivers were removed in the fall, and 1 Weakfish caught by an angler in the Navesink River. In survival analysis, observations are censored when the study ends before the event response occurs (in this case, egress) or when an individual is removed from the study (e.g., dies) before the event response occurs. 148 Fishery Bulletin 112(2-3) ,a> 10 -L, i i . > May Jun Jul Aug Sep Oct 150 200 250 300 Day of the year Figure 2 For a study of residence times and habitat duration of Striped Bass (Morone saxatilis), Bluefish (Pomatomus saltatrix), and Weakfish ( Cynoscion regalis) in 2006 and 2007, mean daily (A) temperatures were measured at the RCM 9 sensor moor- ing in the Navesink River and (B) freshwater (FW) discharge rates (y-axis on log scale) were measured at a U.S. Geological Survey stream flow station in the Swimming River, For loca- tions of the mooring and flow station, see Figure 1. Because the presence-absence data were serially correlated in time for each individual fish, errors were modeled as a first-order autoregressive pro- cess nested within each individual fish. Release date and year also were considered as model covariates. Preliminary models were made with smoothing splines, and covariates were chosen through the use of manual backward selection, analysis of par- tial deviance, and Akaike’s information criterion (AIC) (Wood, 2006). To avoid over-fitting smooths, we set gamma to 1.4 and the basis dimension (k) to 5, limiting the maximum degrees of freedom of the smooths to 4. A covariate was removed from the model if its smoother was statistically insig- nificant, the change in AIC was >0 when the vari- able was removed, or 2 standard error confidence bands in the deviance plots included zero through- out variable domain. Covariates with equivalent degrees of freedom =1 were tested as linear ef- fects before they were eliminated on the basis of these criteria. We used tensor product smooths, which are appropriate when covariates are mea- sured on different scales, to test 2-way interac- tions between body sizes and other significant covariates (Wood, 2006). Because the response of age-l-i- Bluefish to temperature was strongly discontinuous across body sizes (i.e., lengths) at -500 mm TL, we pooled individuals into 2 body- size classes (300-500 mm TL, >500 mm TL) and treated body size as a factor covariate. We used log-rank tests for differences in “residency” curves between the species-and-age classes and years (Harrington and Fleming, 1982). We examined relationships between the presence of individuals in each age class of each tagged species in the estuary and environmental variation with logis- tic generalized additive mixed models (GAMM) in the gamm4 library in R (Aarts et ah, 2008; Wood, 2012). We limited final analyses to body size and the environ- mental variables of water temperature and freshwater discharge, which are important drivers of the estua- rine habitat suitability. Other measured environmen- tal variables were correlated with temperature and freshwater discharge and had lower explanatory power in preliminary models. In addition, the time series for salinity and oxygen in the estuary were incomplete. Fi- nally, complex preliminary GAMMs with more than a few variables also failed to converge. We analyzed water temperatures measured at the RCM 9 mooring and daily freshwater discharge (cubic meters per second) measured in the Swimming River because they were the most complete and accurate time series. We log transformed freshwater discharge values, which were strongly leptokurtic. Individual fish was considered as the random effect in all models. Results Patterns of temperature and freshwater discharge Spring and summer of 2006 were hotter and drier than those seasons in 2007 (Fig. 2, A and B). In 2006, spring warming rates were slightly higher and, in late July- early August, temperatures exceeded 30.0°C (2006 max=30.2°C; 2007 max=28.0°C). During the autumn, however, temperatures were cooler in 2006 than in 2007. Discharge in the Swimming River was high dur- ing the spring of both years. Freshwater discharge was low (<2 m3 s-1) and discharge events were relatively rare from mid-July through August 2006. In 2007, pe- riods of low discharge occurred briefly (2-3 d) once in July and twice in August. Discharge was low through- out much of the fall of 2007, in contrast to several epi- sodes of high river discharge that were produced by frequent rains during the autumn of 2006. Patterns of release The species-and-size classes were available for collec- tion and release during different periods of time (Table 1). Striped Bass were released in May and June. We released age-l+ Bluefish from May to July, but large Manderson et al.: Residence time and habitat duration for predators in a small mid-Atlantic estuary 149 Number of days in the estuary Figure 3 Kaplan-Meier analysis showing that (A) tagged Striped Bass ( Morone saxatilis) used the Navesink River for the fewest number of days, (B) Weakfish ( Cynoscion regalis) for the greatest number of days, and (C) age-l+ and (D) age-0 Bluefish ( Pomatomus saltatrix) were resident in the estuary for intermediate durations in 2006 and 2007 for a study of residence times and habitat duration of these 3 predators. Vertical lines crossing the horizontal line at 0.5 indicate the median number days (numbers above x-axis) each species used the small estuarine sys- tem in each year. individuals >500 mm TL were released in May and early June. We released age-0 Bluefish >175 mm TL in August and September. Weakfish were available for release from late June to mid-September. Release date and body size covaried for each age class of each species during at least one year (Table 1). Smaller fishes were generally available for ear- lier release. However, in 2007, large age-l+ Blue- fish and Weakfish were released earlier than smaller individuals. Patterns of egress Although most of the tagged fishes remained in the Navesink River until final egress, several individuals of all age classes of tagged species made temporary ex- cursions out of the estuary for a period >3 d. More than half of the Striped Bass that we released in 2006 left the estuary temporarily and returned after absences of 3-53 d (n=18; mean excursion [p] = 15.6 d). In 2007, only 25% of the tagged Striped Bass made temporary excur- sions (n=3, m=15.6 d, max=33 d). Several Striped Bass made 2 or more excursions (n=7, max=6 d). Three fish that left the estuary in June or July 2006 returned in late August or September after absences >50 d. Weakfish and Bluefish showed stronger fidelity to the estuarine tributary than Striped Bass. More than 74% of the Weakfish and age-l+ Bluefish that we released remained in the estuary until final egress. Temporary excursions of these fishes (Bluefish n: 2006=4, 2007=5; Weakfish n\ 2006=5, 2007= 6) lasted 2-52 d (p=~15 d). In 2006, Bluefish and Weakfish left the estuary tempo- rarily during the period of late July-early August when temperatures exceeded 28°C and freshwater discharge was low (Fig. 2). In 2007, age-l+ Bluefish made excur- sions outside the estuary in late June-early July, and Weakfish made them throughout the summer. Age-0 Bluefish rarely left the tributary before final egress (n: 2006=1, 2007=3; m=10 d; range: 4-19 d). Duration of estuarine habitat use The species and size classes remained in the estuary for different lengths of time 485 mm TL used the estuary less than 24 d. Many smaller fish had longer residencies (n= 24) and some of them (n= 4) used the estuary >50 d. Median residency periods for age-l+ Bluefish were 19 d in 2006 and 29 d in 2007, but the interannual difference was not significant (%2=1.3, df=l, P=0.248). Several age-l+ Bluefish <500 mm TL (n= 8) used the estuary >40 d. Age-0 Bluefish used the system for a median of 29 d, and distributions of residencies were nearly identical in the 2 years of this study (%2=0.2, df=l, P~ 0.651). Age-0 fish remained in the river for as long as 52 d (n=2). Residencies of tagged age-0 and age-l+ Bluefish were not statistically different (x2=1.8, df=3, P= 0.625). However, we were unable to tag age-0 fish <175 mm TL that occurred in the Navesink River as early as June (senior author, unpubl. data), and re- ceivers were removed before the final egress of several tagged age-0 individuals ( n : 2006=4, 2007=1). There- fore, age-0 Bluefish probably used the system much longer than age-l-i- fish. Weakfish remained in the estuary for a median of 33 d in 2006 and 47 d in 2007 (/2=5.6, df=l, P=0.02). Resi- dencies may have been longer in 2007 because, during that year, Weakfish were released earlier and the ob- servation period was longer. All Weakfish <400 mm TL (n=18) used the estuary >40 d and 10 individuals were resident >60 d. Effects of environmental variables and body size on residence and egress Smaller individuals of all 3 species tended to re- main in the estuary at warmer temperatures than those preferred by larger individuals (Table 3, Fig. 4). Size-dependent temperature responses were con- tinuous for Striped Bass and Weakfish but discontinu- ous for Bluefish. On average, Striped Bass were more likely to leave the system when temperatures ex- ceeded 23°C than when cooler temperatures occurred (Fig. 4A). However, temperature effects were greater for larger Striped Bass, which left rapidly as tem- peratures warmed in the early summer. In contrast, smaller fish were more likely to remain in the sys- tem into the summer when temperatures were rela- tively warm. Large Bluefish released in early summer were also likely to emigrate from the estuary when temperatures increased above 23°C (Fig. 40. In con- trast, smaller age-l-i- Bluefish were likely to be pres- ent when temperatures were warmer (Fig. 4D). Age- 0 Bluefish were likely to remain in the estuary when temperatures were warmest. It was unlikely for age-0 Bluefish to leave until autumn temperatures fell be- low ~20.5°C (Fig. 4E). Weakfish were resident in the estuary at the warmest temperatures and were likely to leave the estuary when temperatures cooled below 23°C (Fig. 4B). Larger Weakfish emigrated at slightly higher temperatures than smaller fish. All 4 species-and-age classes were more likely to leave the estuary when the Swimming River dis- Manderson et al.: Residence time and habitat duration for predators in a small mid-Atlantic estuary 151 Table 3 Results from the final generalized additive mixed models of effects of body size, estuarine temperature, and freshwater dis- charge (FW) on the residence time of ultrasonieally tagged Striped Bass (Morone saxatilis), Bluefish (Pomatomus saltatrix; age-0 and age-l+), and Weakfish ( Cynoscion regalis ) released into the Navesink River, New Jersey, in 2006 and 2007 (see Figs. 4 and 5 for deviance plots). Individual fish was included as a random effect (i.e. , intercept) in all models. Temporal autocorrelation in detections was considered as a first-order, autoregressive process that occurred within each fish The independent variables included in initial models were year as a factor, as well as release day, body length, temperature , and freshwater discharge , all of which were first considered with cubic smoothing splines (s) with a maximum of 4 degrees of freedom. Tensor product smooths (t2) were used to model interactions. Variables were included as linear effects if expected degrees of freedom (EDF) of splines were close to 1, and they were eliminated from models when they did not contribute to a reduction in Akaike’s information criterion (AIC). Body length (total length in millimeters) was considered as a class variable in modeling the temperature response of age-l+ Bluefish because the lengthx temperature interaction was strongly discontinuous. Species and Parametric coefficient Estimate SE Z-value P-value AIC Striped Bass Intercept -2.778 0.177 -15.73 <0.0001 4354 Approximate significance of nonparametric terms EDF X2 t2 ( Temperature , body length) 16.678 228.65 <0.0001 4111 s(log(FW Discharge + 1)) 3.898 201.90 <0.0001 3888 s(Release day) Coefficient of multiple determination [f?2]=0.15 2.925 29.66 <0.0001 3844 Age-1+ Bluefish Intercept -1.739 0.376 -4.626 <0.0001 4301 Year -1.055 0.473 -2.230 0.0257 4309 Approximate significance of nonparametric terms EDF X2 s (Temperature)'.Length <500 mm 3.927 286.79 <0.0001 3700 s (TemperatureY.Length >500 mm 2.970 193.74 <0.0001 s(log(FVV Discharge + 1)) 3.893 418.689 <0.0001 3122 s (Release day) R2= 0.264 1.958 9.362 0.0088 3120 Age-0 Bluefish Intercept 0.861 0.550 1.566 0.1170 1974 log(FVF Discharge + 1) 0.922 0.171 5.385 <0.0001 1900 Approximate significance of nonparametric terms EDF X2 s (Temperature) R2= 0.235 3.345 330.9 <0.0001 1128 Weakfish Intercept -16.980 4.714 -3.602 0.0179 3731 Year 2.484 1.049 2.367 0.0257 3729 Release day 0.079 0.022 3.555 0.0004 3727 Approximate significance of terms EDF X2 t2 (Temperature, Body length ) 12.413 580.5 <0.0001 2060 s(log(FW Discharge + 1)) 3.865 131.0 <0.0001 1932 R2=0.368 charge fell below ~2 m3 s"1 than when discharge was higher (Table 3, Fig. 5). This effect was evident when we included year as a factor and when we modeled years separately. As a result, the response to low dis- charge did not appear to be related to interannual differences in sample size or freshwater discharge. Striped Bass were also likely to leave the tributary during episodes of high freshwater discharge ( >50 m3 s_1; Fig. 5A). Age-0 Bluefish and Weakfish were best modeled with linear discharge terms, indicating that the animals were not likely to leave the estuary dur- ing periods when freshwater discharge was high (Fig. 5, B and C). There were significant differences in patterns of residency among individual fishes (random intercept; Table 3). Furthermore, the year effect was significant in GAMMs for Striped Bass, Weakfish, and age-l+ Bluefish, consistent with descriptions in the previous section, un- der Patterns of egress. Release date was significant in the models for Striped Bass, Weakfish, and age-l+ Bluefish. 152 Fishery Bulletin 1 12(2-3) E E O) c 0) "ca o E E cn c 0> A 600 550 500 450 400 350 B 500 450 400 350 300 250 15 20 25 30 Temperature (°C) Figure 4 Deviance plots from logistic generalized additive mixed models showing partial effects of temperature and body size on the residence of (A) Striped Bass ( Mo rone saxatilis) and (B) Weakfish ( Cynoscion regalis), a Bluefish ( Pomatomus saltatrix) with total lengths (C) <300 mm, (Di of 300-500 mm, and (E) >500 mm, all tagged in the Navesink River in 2006 and 2007 (see Table 3). The relationship of residence to temperature and body size was continu- ous for Striped Bass and Weakfish which were more likely to be resident over a broader temperature range at smaller body sizes than they were at larger body sizes. Vertical lines crossing the horizontal line at 0.0 indicate boundaries between positive and negative ef- fects, and shaded areas represent ±2 standard-error confidence bands. The fish that were released later in the season typically were small and generally had longer residencies. Discussion Our observations of estuarine residency for individu- al Striped Bass, Bluefish, and Weakfish indicate that small (-1000 ha), mid-Atlantic estuarine tributaries, such as the Navesink River, contain habitat resources necessary to support survival and adequate growth of age-0 and age-1 individuals for relatively long peri- ods of time during the summer months. Of the age-l+ Striped Bass that we tagged, 25% used the river for more than 26 d, and the same fractions of the Bluefish (age 1 and 0) and age-l+ Weakfish that we tagged used the system for more than 36 and 62 d. Earlier inves- tigations that used fortnightly gillnet surveys of the Navesink River and adjacent Sandy Hook Bay indicat- ed that these 3 predators are abundant in the system, which they use as a feeding habitat, nursery habitat, or both (Scharf et ah, 2004; Manderson et. al., 2006). Our Manderson et al.: Residence time and habitat duration for predators in a small mid-Atlantic estuary 153 D Freshwater discharge log (rp3 s"1 +1) Figure 5 Plots from logistic generalized additive mixed models showing partial deviance effects of freshwater discharge from the Swimming River on the residence of the 3 predator species in the Navesink River (see Table 3) tagged in our study in 2006 and 2007: (A) Striped Bass (Morone saxatilis), (B) Weakfish ( Cynoscion regalis), and (C) age-0+ and (D) age-1 Bluefish ( Pomatomus saltatrix). Vertical lines crossing the horizontal line at 0.0 indicate boundaries between positive and negative effects, and shaded areas represent ±2 standard-error confidence band. telemetry study indicates that high abundances reflect relatively long-term residence times for predators <500 mm TL rather than a rapid flux of many transient in- dividuals through the ecosystem. Long-term residences of individual fishes with little straying indicates that temperature, oxygen, and prey resources persist at suitable levels in the small tributary for relatively long periods. Predators with small body sizes had longer residen- cies and, therefore, appeared to be supported longer by habitat resources in the small ecosystem than were larger individuals. Large Striped Bass and Bluefish (>500 mm TL) used the tributary for a few days to a few weeks during the spring. Large Weakfish (>400 mm TL) released later in the summer were also relatively transient. Smaller age-l+ Bluefish remained in the es- tuary for intermediate lengths of time. Finally, age-0 Bluefish and small age-l-i- Weakfish (<400 mm TL) had the longest residence times (median residence=29 d and ~40 d) that were probably underestimated in our study. Although Weakfish were common in gill nets in May (L. Stehlik and senior author, unpubl. data), we were able to capture them only with hook and line in early July after their diets had shifted from invertebrate to fish prey. Small Bluefish (20-30 mm TL), which cannot be surgically tagged, are collected in Navesink River as early as June in beach seines and fine mesh gillnets (L. Stehlik, unpubl. data). Small Weakfish and Blue- fish were, therefore, resident in the Navesink River probably for much longer periods than those that we measured. Our observations of long residences of small predator cohorts in the Navesink River are consistent with observations made in larger estuarine ecosystems (Grothues and Able, 2007; Taylor et al., 2007; Wingate and Secor, 2007; Mather et al., 2009; Turnure, 2010). The size-dependent patterns of estuarine residence time for the 3 studied predators may have been relat- ed to size-dependent requirements for prey resources. 154 Fishery Bulletin 1 12(2-3) Metabolic rates of animals generally scale with body mass to approximately the % power (Anderson-Teixeira et ah, 2009; and references therein; note that in our study we were concerned with resource requirements of individual whole fish that influence residency, not with mass-specific metabolic rates). Therefore, larger or older individuals require more prey resources per unit of energy cost of prey acquisition (i.e. , search, capture, handling time, and digestion) than do smaller preda- tors to meet metabolic demand at a given temperature. Prey resources in the Navesink River include large numbers of small invertebrates and fishes, such as age-0 Atlantic Menhaden (Brevoortia tyrannus), Atlan- tic Silverside (Menidia menidia ), Bay Anchovy ( Anchoa mitchilli), mysids, and sevenspine bay shrimp ( Cran - gon septemspinosa ) (Scharf et ah, 2004; L. Stehlik and senior author, unpubl. data). Diet studies of small size classes (<500 mm TL) of predators indicate that age-0 Atlantic Menhaden (<200 mm TL) are preferred prey that, with other small prey, reside in the Navesink River throughout the warmer months. Larger (>200 mm TL), energy rich age-l+ At- lantic Menhaden, consumed by the largest Bluefish and Striped Bass, are abundant in the river during late spring, but they migrate out of the tributary in June and July before returning again in early autumn (Scharf et al., 2004). Early summer egress of the larg- est Striped Bass and Bluefish (>500 mm TL) from the Navesink River coincided with the typical timing of egress for age-l+ Atlantic Menhaden and other large prey (L. Stehlik and senior author, unpubl. data). These large prey may be required by large fishes, particularly when warm temperatures increase metabolic demand. Metabolic demand in ectotherms is regulated by environmental temperatures, as well as by body size (Hartman and Brandt, 1995; Brown, 2004; Sousa et al., 2010), and our GAMMs indicated that residence times of the 3 predators in the Navesink River were a func- tion of the interaction between body size and water temperature. For all species, threshold temperatures for egress and breadths of temperatures associated with estuarine residence decreased with increasing body size. The largest age-l+ Striped Bass and Blue- fish (>500 mm TL) released in the spring were likely to remain in the river only until temperatures exceeded 23°C in the early summer. Smaller Striped Bass were less sensitive than large fish and remained in the river over a broader range of warmer temperatures. Smaller age-l+ Bluefish were also more likely to be resident at warmer temperatures ranging from 23°C to 26°C. Age-0 Bluefish remained in the estuary at the warmest temperatures recorded and were unlikely to leave until temperatures declined below 19°C during autumn. Finally, Weakfish also remained in the estu- ary when temperatures were warmest and were more likely to leave the river when temperatures declined below 23°C in the autumn. Smaller Weakfish, however, remained in the river longer and over a broader range of temperatures. The relationships between estuarine residency time, body size, and environmental temperature that we ob- served are consistent with bioenergetic studies and metabolic theory (Gillooly et al., 2001; Brown, 2004; Harris et al., 2006; Sousa et al., 2010). The species- and size-specific temperatures of estuarine residence and egress that we measured were extremely similar to temperatures and size-dependent, scopes for growth reported by Steinberg (1994) and Hartman and Brandt (1995). In those studies, growth potential exceeded 2% of body weight per day at temperatures of 12-25°C (op- timal 15°C) for Striped Bass, 16 -26°C (optimal 20°C) for Bluefish and 20-29°C (optimal 23.5°C) for Weakfish. Smaller individuals generally had higher optimal temperatures for growth because metabolic demand and prey requirements are generally smaller for ani- mals with small body sizes. For example, the thermal optima for age-l+ Bluefish was ~20°C, but growth po- tential for age-0 Bluefish reached a maximum at tem- peratures of ~25-27°C (Steinberg, 1994; Hartman and Brandt, 1995; Scharf et al., 2006). Ranges of optimal temperatures for various performance measures are also generally broader for smaller, juvenile ectotherms (Freitas et al., 2010), and our GAMMs indicated that smaller fishes were more likely than larger fish to re- main in the Navesink River over a broader range of temperatures. Because metabolic demand increases with temperature as well as body size, prey supply shortages are more likely to occur during the warmest summer months for large animals in small estuarine tributaries like the Navesink River. Residence time and egress of the 3 studied preda- tors also were related to the rate of freshwater dis- charge from the Swimming River into the Navesink River. On the basis of the 4 GAMMs that we construct- ed independently for the predators, we determined that animals were more likely to leave the small estuarine system when average daily freshwater discharge rates from the Swimming River fell below -2 m3 s-1 than when discharge rates were higher. High discharge events ( >50 m3 s-1) also appeared to affect residencies of Striped Bass and, perhaps, age-l+ Bluefish. Howev- er, in contrast with this low discharge response, high discharge response thresholds varied by species. The predators that we tagged were euryhaline and probably did not respond behaviorally at the scale of the whole estuary to the direct physiological effects of increas- ing or high salinities. We hypothesize that the effects of low discharge on residence and egress were indirect through hydrographic processes that control the avail- ability of prey resources that support the entire suite of predators that we tagged. Variability in freshwater discharge is believed to af- fect estuai’ine fishes primarily by changing estuarine hydrodynamics that control prey resource availability. Interactions between freshwater discharge and tides control gravitational circulation in estuaries and the advection and concentration of the essential building blocks of estuarine food webs. As a result, estuaries Manderson et al.: Residence time and habitat duration for predators in a small mid-Atlantic estuary 155 are traps for autochthonous and allochthonous nutri- ents and organic matter from adjacent terrestrial and marine systems (MacCready and Geyer, 2010). Sta- ble isotope studies indicate that estuarine food webs are supported by inputs of freshwater and terrestri- al sources of nutrients and organic matter (Kostecki et ah, 2010). It is assumed that discharge effects on estuarine hydrodynamics and nutrient transport ulti- mately concentrate high secondary production of zoo- plankton in estuarine regions where fresher and saltier waters converge (North and Houde, 2006; Baptista et ah, 2010). These mechanisms are thought to produce a dome-shaped relationship between estuarine fish pro- duction and freshwater discharge (Dolbeth et ah, 2010; and references therein). In the Navesink River, tidal asymmetries produce a short-duration, high-velocity flood tide followed by a long, slow ebb (Chant and Stoner, 2001). During flood tides, particles are suspended and transported up- stream. When freshwater discharge from the Swimming River is sufficient, the water column in the Navesink River stratifies during the ebb and particles accumulate in the central and upper reaches of this river. Finer particles and flocculants can remain in suspension in the upper Navesink River (Fig. 1., between locations B and C), where a convergence zone is formed by the tur- bulent mixing of freshwater inflow from the Swimming River and tidal inflows of saltwater from Sandy Hook Bay and the Atlantic Ocean (Fugate and Chant, 2005). In this area, Shaheen et ah (2001) reported high con- centrations of the copepod Eurytemora affinis, an impor- tant constituent of estuarine food webs. We measured relatively sharp gradients in salinity and chlorophyll-a and, compared with levels observed in other areas in our study, higher abundances of small fish prey, includ- ing Atlantic Silverside and age-0 Atlantic Menhaden in combined hydrographic and gillnet surveys (L. Stehlik and senior author, unpubl. data). Additionally, most of the predators that we tagged established home ranges in this region for days to weeks when temperatures in the upper estuary remained below thresholds associated with egress (L. Stehlik and senior author, unpubl. data; see also Scharf et ah, 2004; Manderson et ah2). Estuaries change from stratified to well-mixed states when freshwater discharge decreases and salin- ity stratification weakens to the point that estuarine Richardson numbers reach a range of 0.08-0.8 (Fischer, 1979; MacCready and Geyer, 2010). This transition oc- curs in the Navesink River when freshwater discharge from the Swimming River falls to ~1 in3 s_1, (Chant3), 2 Manderson, J. P., J. Pessutti, J. E. Rosendale, and B. Phelan. 2007. Estuarine habitat dynamics and telemetered move- ments of three pelagic fishes: Scale, complexity, behavioral flexibility and the development of an ecophysiological frame- work. ICES Council Meeting (C.M.) Documents 2007/G:02, 36 p. 3 Chant, R. 2004. Personal commun. Institute of Coastal and Marine Science, Rutgers Univ., 71 Dudley Rd., New Brunswick, NJ 08901. a discharge rate similar to the value at which all the predators we tagged were likely to leave this small es- tuary in New Jersey. We speculate that the relation- ship between predator egress and freshwater discharge reflects a shift from a partially mixed to a fully mixed estuarine state and the relaxation of physical mecha- nisms that control and concentrate the high primary and secondary productivity that supports the 3 studied predators in the upper reaches of this estuary. Conclusions Our analyses of residence time and egress of individual Striped Bass, Bluefish, and Weakfish in the Navesink River, New Jersey, indicate that small estuarine tribu- taries contain the habitat resources required to sustain juvenile and small adult stages of these 3 predators for relatively long periods of time but that the resources that regulate habitat suitability are ephemeral. Re- quired resources include temperature, which regulates metabolic demand and predatory capacity in cold-blood- ed fishes (Magnuson et al., 1979; Neill et al., 1994). Summer temperatures in the Navesink River appeared to support smaller predators for longer durations than they did for larger fishes presumably because prey re- quirements increase with body size and temperature and because the small tributary is dominated by small rather than large prey during the warmest summer months. Freshwater discharge also appeared to be a critical habitat resource that controlled residence time for ani- mals in this estuary. We believe this relationship re- flects the essential role that freshwater discharge plays in regulation of physical processes that both drive and concentrate the secondary productivity required to meet the prey resource requirements of the predators. Other factors that we were not able to measure effec- tively, particularly dissolved oxygen concentrations and human predation pressure, also may have influenced habitat suitability and the residence time and timing of egress of predators in this small estuarine system (Brady et al., 2009). Because estuaries occur at the land-sea boundary, high-frequency variability in atmospheric temperature, precipitation, and wind is rapidly translated into vari- ability in water temperature, freshwater discharge, dissolved oxygen, and other biophysical processes that determine estuarine habitat suitability. Changes in seasonal rates of warming, cooling, and precipitation that alter and reduce the persistence of suitable es- tuarine habitats should require animals to undertake more frequent, long-distance movements that are ener- getically costly. Conversely, long durations of suitable habitat conditions require fewer shifts in local home range (Martinho et al., 2009) and allow the allocation of resources to the life-history processes of growth and reproduction instead of long-distance movements. 156 Fishery Bulletin 112(2-3) Increased movement costs should come at the ex- pense of strategies that reduce predation risk and increase growth and reproduction rates. Changes in atmospheric forcing with climate change are coherent over spatial scales of 1000s of kilometers (Hare and Able, 2007; Manderson, 2008; Shearman and Lentz, 2010). As a result, climate-driven changes in habitat and persistence should affect the energy budgets and survival of many individuals over broad areas. These effects should be translated across a level of ecologi- cal organization to affect the birth and death rates of regional fish populations. Literature cited Aarts, G., M. MacKenzie, B. McConnell, F. Mike, and J. Matthiopoulo. 2008. Estimating space-use and habitat preference from wildlife telemetry data. Ecography 31:140-160 Able, K. W. 2005. A re-examination of fish estuarine dependence: evidence for connectivity between estuarine and ocean habitats. Estuar. Coast. Shelf Sci. 64:5-17. Able. K. W., and M. P. Fahay. 2010. Ecology of estuarine fishes: temperate waters of the western North Atlantic, 566 p. Johns Hopkins Univ. Press, Baltimore. Anderson-Teixeira, K. J., V. M. Savage, A. P. Allen, and J. F. Gillooly. 2009. Allometry and metabolic scaling in ecology. In Encyclopedia of life sciences. John Wiley & Sons, Ltd, Chichester, UK. [Available from http://www.els.net/Wi- leyCDA/ElsArticle/refld-a002 1222.html.] Armsworth, P. R., and J. E. Roughgarden. 2005. The impact of directed versus random movement on population dynamics and biodiversity patterns. Am. Nat. 165:449-465. Baptista, J., F. Martinho, M. Dolbeth, I. Viegas, H. Cabral, and M. Pardal. 2010. Effects of freshwater flow on the fish assemblage of the Mondego estuary (Portugal): comparison between drought and non-drought years. Mar. Freshw. Res. 61:490-501. Belisle, M. 2005. Measuring landscape connectivity: the challenge of behavioral landscape ecology. Ecology 88:1988-1995. Bennetts, R. E., J. D. Nichols, J.-D. Lebreton, R. Pradel, J. E. Hines, and W. M. Kitchens. 2001. Methods for estimating dispersal probabilities and related parameters using marked animals. In Disper- sal (J. Clobert, E. Danchin, A. A. Dhondt and J. D. Nich- ols, eds.), p. 3-17. Oxford Univ. Press, Oxford, UK. Brady, D. C., D. M. Tuzzolino, and T. E. Targett. 2009. Behavioral responses of juvenile weakfish ( Cy - noscion regalis) to diel-cycling hypoxia: swimming speed, angular correlation, expected displacement, and effects of hypoxia acclimation. Can. J. Fish. Aquat. Sci. 66:415-424. Brown, J. H. 2004. Toward a metabolic theory of ecology. Ecology 85:1771-1789. Chant, R. J., and A. W. Stoner. 2001. Particle trapping in a stratified flood-dominated estuary. J. Mar. Res. 59:29-51. Charnov, E. L. 1976. Optimal foraging: the marginal value theo- rem. Theor. Popul. Ecol. 9:129-136. Chiarella, L. A., and D. O. Conover. 1990. Spawning season and first-year growth of adult bluefish from the New York Bight. Trans. Am. Fish. Soc. 119:455-462. Dolbeth, M., F. Martinho, V. Freitas, S. Costa-Dias, J. Campos, and M. Pardal. 2010. Multi-year comparisons of fish recruitment, growth and production in two drought-affected Iberian estuar- ies. Mar. Freshw. Res. 61:1399-1415. Fischer, H. B., E. J. List, R. C. Y. Koh, J. Imberger, and N. H. Brooks. 1979. Mixing in inland and coastal waters, 483 p. Aca- demic Press, San Diego, CA. Freitas, V., K. Cardoso, K. Lika, M. A. Peck, J. Campos, S. Kooijman, and H. W. van der Veer. 2010. Temperature tolerance and energetics: a dynamic energy budget-based comparison of North Atlantic ma- rine species. Philos. Trans. R. Soc. Lond., B: Biol. Sci. 365:3553-3565. Fugate, D. C., and R. J. Chant. 2005. Near-bottom shear stresses in a small, highly stratified estuary. J. Geophys. Res. 110:C03022. Gillooly, J. F., J. H. Brown, G. B. West, V. M. Savage, and E. L. Charnov. 2001. Effects of size and temperature on metabolic rate. Science 293:2248-2251. Grothues, T. M., and K. W. Able. 2007. Scaling acoustic telemetry of bluefish in an es- tuarine observatory: detection and habitat use pat- terns. Trans. Am. Fish. Soc. 136:1511-1519. Hagy, J. D. 2002. Eutrophication, hypoxia and trophic transfer ef- ficiency in Chesapeake Bay. Ph.D. diss., 330 p. Univ. Maryland, College Park, MD. Hare, J. A., and K. W. Able. 2007. Mechanistic links between climate and fisheries along the east coast of the United States: explaining population outbursts of Atlantic croaker (Micropogonias undulatus). Fish. Oceanogr. 16:31-45. Harrington, D. P, and T. R. Fleming. 1982. A class of rank test procedures for censored sur- vival data. Biometrika 69:553-566. Harris, L., C. Duarte, and S. Nixon. 2006. Allometric laws and prediction in estuarine and coastal ecology. Estuar. Coasts 29:340-344. Hartman, K. J., and S. B. Brandt. 1995. Comparative energetics and the development of bioenergetics models for sympatric estuarine pisci- vores. Can. J. Fish. Aquat. Sci. 52: 1647-1666 Johnson, J. C., J. J. Luczkovich, S. P. Borgatti, and T. A. B. Snijders. 2009. Using social network analysis tools in ecology: Markov process transition models applied to the sea- sonal trophic network dynamics of the Chesapeake Bay. Ecol. Model. 220:3133-3140. Kostecki, C., F. Le Loc’h, J. M. Roussel, N. Desroy, D. Huteau, P. Riera, H. Le Bris, and O. Le Pape. 2010. Dynamics of an estuarine nursery ground: the spatio-temporal relationship between the river flow and Manderson et al.: Residence time and habitat duration for predators in a small mid-Atlantic estuary 157 the food web of the juvenile common sole ( Solea solea, L.) as revealed by stable isotopes analysis. J. Sea Res. 64:54-60. Krause, A. E., K. A. Frank, D. M. Mason, R. E. Ulanowicz, and W. W. Taylor. 2003. Compartments revealed in food-web struc- ture. Nature 426:282-285. Livingston, R. J. 1987. Field sampling in estuaries: the relationship of scale to variability. Estuaries 10:194-207. MacCready, P., and W. R. Geyer. 2010. Advances in estuarine physics. Annu. Rev. Mar. Sci. 3:35-58. Magnuson, J. J., L. B. Crowder, and P. A. Medvick. 1979. Temperature as an ecological resource. Am. Zool. 19:331-343. Manderson, J. P. 2008. The spatial scale of phase synchrony in winter flounder (Pseudopleuronectes americanus) production increased among southern New England nurseries in the 1990s. Can. J. Fish. Aquat. Sci. 65:340-351. Manderson, J. P., J. Pessutti, C. Meise, D. Johnson, and P. Shaheen. 2003. Winter flounder settlement dynamics and the modification of settlement patterns by post-settlement processes in a NW Atlantic estuary. Mar. Ecol. Prog. Ser. 253:253-267. Manderson, J. P., J. Pessutti, P. Shaheen, and F. Juanes. 2006. Dynamics of early juvenile winter flounder preda- tion risk on a North West Atlantic estuarine nursery ground. Mar. Ecol. Prog. Ser. 329:249-265. Manderson, J. P, B. A. Phelan, C. Meise, L. L. Stehlik, A. J. Bejda, J. Pessutti, L. Arlen, A. Draxler, and A. W. Stoner. 2002. Spatial dynamics of habitat suitability for the growth of newly settled winter flounder Pseudopleuro- nectes americanus in an estuarine nursery. Mar. Ecol. Prog. Ser. 228:227-239. Mann, K. H 2000. Ecology of coastal waters: with implications for management, 2nd ed., 432 p. Blackwell Science, Inc., Malden, MA. Martinho, F., M. Dolbeth, I. Viegas, C. M. Teixeira, H. N. Ca- bral, and M. A. Pardal. 2009. Environmental effects on the recruitment vari- ability of nursery species. Estuar. Coast. Shelf Sci. 83:460-468. Mather, M. E. E, K. H. Ferry, L. A. Deegan, and G. A. Nelson. 2009. Use of non-natal estuaries by migratory striped bass (Morone saxatilis) in summer. Fish. Bull. 107:329. Munch, S. B., and D. O. Conover. 2000. Recruitment dynamics of bluefish ( Pomatomus saltatrix) from Cape Hatteras to Cape Cod, 1973- 1995. ICES J. Mar. Sci. 57:393-402. Neill, W. H., J. M. Miller, H. W. Van Der Veer, and K. O. Winemiller. 1994. Ecophysiology of marine fish recruitment: a con- ceptual framework for understanding interannual vari- ability. J. Sea Res. 32:135-152. Ng, C., K. W. Able, and T. M. Grothues. 2007. Habitat use, site fidelity, and movement of adult striped bass in a southern New Jersey estuary based on mobile acoustic telemetry. Trans. Am. Fish. Soc. 136:1344-1355. North, E. W., and E. D. Houde. 2006. Retention mechanisms of white perch (Morone americana ) and striped bass (Morone saxatilis ) early- life stages in an estuarine turbidity maximum: an in- tegrative fixed-location and mapping approach. Fish, Oceanogr. 15:429-450. Pautzke, S. M., M. E. Mather, J. T. Finn, L. A. Deegan, and R. M. Muth. 2010. Seasonal use of a New England estuary by forag- ing contingents of migratory striped bass. Trans. Am. Fish. Soc. 139:257-269. Peterson, M. S., M. R. Weber, M. L. Partyka, and S. T. Ross. 2007. Integrating in situ quantitative geographic infor- mation tools and size-specific laboratory-based growth zones in a dynamic river-mouth estuary. Mar. Freshw. Ecosys. 17:602-618. R Core Team 2013. R: a language and environment for statistical com- puting. R Foundation for Statistical Computing, Vien- na, Austria. [Available from http://www.R-project.org.] Scharf, F. S., J. A. Buckel, K. A. Rose, and F. Juanes 2006. Effects of variable prey and cohort dynamics on growth of young-of-the-year estuarine bluefish: evidence for interactions between spring- and summer-spawned cohorts. Trans. Am. Fish. Soc. 135:1266-1289. Scharf, F. S., J. P. Manderson, M. C. Fabrizio, J. P. Pessutti, J. E. Rosendale, R. J. Chant, and A. J. Bejda. 2004. Seasonal and interannual patterns of distribution and diet of bluefish within a middle Atlantic Bight estu- ary in relation to abiotic and biotic factors. Estuaries 27: 426-436. Shaheen, P. A., L. L. Stehlik, C. J. Meise, A. W. Stoner, J. P. Manderson, and D. L. Adams. 2001. Feeding behavior of newly settled winter floun- der ( Pseudopleuronectes americanus) on calanoid cope- pods. J. Exp. Mar. Biol. Ecol. 257:37-51. Shearman, R. K., and S. J. Lentz. 2010. Long-term sea surface temperature variabil- ity along the U.S. East Coast. J. Phys. Oceanogr. 40:1004-1017. Sousa, T., T. Domingos, J. C. Poggiale, and S. A. L. M. Kooijman. 2010. Dynamic energy budget theory restores coherence in biology. Philos. Trans. R. Soc. Lond., B: Biol. Sci. 365:3413-3428. Steinberg, N. S. 1994. Young-of-the-year bluefish ( Pomatomus salta- trix) consumption in the Hudson River estuary: a bio- energetic modeling approach. M.S. thesis, 100 p. State Univ. New York at Stony Brook, Stony Brook, NY. Stoner, A. W., J. P. Manderson, and J. Pessutti. 2001. Spatially explicit analysis of estuarine habitat for juvenile winter flounder (Pseudopleuronectes america- nus): combining generalized additive models and geo- graphic information systems. Mar. Ecol. Prog. Ser. 213:253-271. Taylor, D. L., R. Nichols, and K. W. Able. 2007. Habitat selection and quality for multiple cohorts of young-of-the-year bluefish (Pomatomus saltatrix): comparisons between estuarine and ocean beaches in southern New Jersey. Estuar. Coast. Shelf Sci. 73:667-679. 158 Fishery Bulletin 112(2-3) Turnure, J. T. 2010. Estuarine habitat ecology of adult weakfish (Cy- noscion regalis): a multi-scale approach. M.S. thesis, 139 p. Rutgers, State Univ. New Jersey, New Bruns- wick, NJ. Wingate, R. L., and D. H. Secor. 2007. Intercept telemetry of Hudson River striped bass resident contingent: migration and homing pat- terns. Trans. Am. Fish. Soc. 136:95-104. Winkler, K., J. H. Rappole, and M. A. Ramos. 1995. The use of movement data as an assay of habitat quality. Oecologia 101:211-216. Wood, S. 2006. Generalized additive models: an introduction with R. Chapman and Hall, Boca Raton, FL. 2012. gamm4: Generalized additive mixed models using mgcv and lme4. R package, vers. 0.1-6. [Available from http://CRAN.R-project.org/package=gamm4.] 159 Abstract— The link between ocean temperature and spatial and tempo- ral distribution patterns of 8 species of small cetaceans off Southern Cali- fornia was examined during the period 1979-2009. Averages and anomalies of sea-surface temperatures (SSTs) were used as proxies for SST fluctua- tions on 3 temporal scales: seasonal, El Nino-Southern Oscillations (ENSO), and Pacific Decadal Oscillations (PDO). The hypothesis that cetacean species assemblages and habitat associations in southern California waters co-vary with these periodic changes in SST was tested by using generalized additive models. Seasonal SST averages were included as a predictor in the models for Dali’s porpoise ( Phocoenoides dalli), and common dolphins ( Delphinus spp.), northern right whale dolphin ( Lisso - delphis borealis), and Risso’s dolphin ( Grampus griseus). The ENSO index was included as a predictor for north- ern right whale, long-beaked common (Delphinus capensis ), and Risso’s dol- phins. The PDO index was selected as a predictor for Dali’s porpoise and Pacific white-sided (Lagenorhynchus obliquidens), common, and bottlenose ( Tursiops truncatus) dolphins. A metric of bathymetric depth was included in every model, and seafloor slope was included in 5 of the 9 models, an indi- cation of a distinctive spatial distribu- tion for each species that may repre- sent niche or resource partitioning in a region where multiple species have overlapping ranges. Temporal changes in distribution are likely a response to changes in prey abundance or dis- persion, and these patterns associated with SST variation may foreshadow future, more permanent shifts in dis- tribution range that are due to global climate change. Manuscript submitted 1 April 2013. Manuscript accepted 28 March 2014. Fish. Bull. 112:159-177 (2014). doi:10.7755/FB.112.2-3.5 The views and opinions expressed or implied in this article are those of the author (or authors) and do not necessarily reflect the position of the National Marine Fisheries Service, NOAA. Effects of fluctuations in sea-surface temperature on the occurrence of small cetaceans off Southern California E. Elizabeth Henderson1 Karin A. Forney2 Jay P. Barlow3 John A. Hildebrand1 Annie B. Douglas4 John Calambokidis4 William J. Sydeman5 Email address for contact author: elizabeth.henderson@nmmpfoundation.org 1 Scripps Institution of Oceanography University of California, San Diego 9500 Gilman Drive Mailcode 0905 La Jolla, California 92093 Present address for contact author: National Marine Mammal Foundation 2240 Shelter Island Drive, Suite 200 San Diego, California 92106 2 Marine Mammal and Turtle Division Southwest Fisheries Science Center National Marine Fisheries Service, NOAA 1 10 Shaffer Road Santa Cruz, California 95060 3 Marine Mammal and Turtle Division Southwest Fisheries Science Center National Marine Fisheries Service, NOAA 8901 La Jolla Shores Drive La Jolla, California 92037 4 Cascadia Research Collective 2181/2 W. 4 th Ave. Olympia, Washington 98501 5 Farallon Institute for Advanced Ecosystem Research 101 H Street, Suite Q Petaluma, California 94952 Highlands, New Jersey 07732 Cetaceans are higher-trophic-level marine predators whose movement patterns and habitat preferences are typically related to the distribution of their prey (Wishner et ah, 1995; Gowans et al. 2007). Unlike baleen whales, small cetaceans (porpoises, dolphins, and small-toothed whales) generally do not undertake ocean- scale annual migrations to track prey or to move between breeding and feeding grounds. Rather, small cetaceans may display a high degree of site fidelity, or they may move sea- sonally inshore and offshore or along regional-scale coastlines (Leather- wood et ah, 1984; Dohl et ah, 1986; Shane et ah, 1986; Forney and Bar- low, 1998). Although many small cetacean species may overlap in any one re- gion of their total range, they often differ in their occurrence or habitat- use patterns, perhaps reflecting com- petitive exclusion or niche partition- ing. This separation of habitat and resources often occurs along depth, slope, and sea-surface temperature (SST) gradients (Reilly, 1990; Forney, 2000; Ballance et ah, 2006; MacLeod et al., 2008). Habitat preferences likely reflect differences in preferred prey. Dolphins may follow prey habi- tats as they shift not only season- ally but through large-scale climate- driven changes such as the El Nino- Southern Oscillation (ENSO) or the Pacific Decadal Oscillation (PDO) (Shane, 1995; Defran, 1999; Benson et ah, 2002; Ballance et ah, 2006). We examined the distribution and relative abundance of multiple spe- cies of small cetaceans across shift- ing temperature regimes off South- 160 Fishery Bulletin 112(2-3) ern California by using a unique coupled cetacean- oceanographic long-term data set. This data set enables a rare opportunity to assess interdecadal changes in cetacean distribution over a broad spatial extent. The co-occurrence of cold- and warm-water cetacean species makes this location an ideal one at which to examine potential effects of climate variation on regional dis- tribution patterns at different temporal scales (intra- annual, annual, and decadal). The Southern California region represents the con- vergence of warm- and cold-water masses and supports populations of both warm- and cold-water, small ceta- cean species (Forney and Barlow, 1998). During the summer, the cold, equatorward flowing California Cur- rent system has a seasonal maximum (7.8 Sverdrups [Sv], -7.8 million m3 s_1). The California Current turns shoreward (poleward) at approximately 32°N and be- comes the California Countercurrent. The California Countercurrent and California Undercurrent also have a seasonal maximum in late summer and into the fall and, therefore, dominate the Southern California Bight, with a combined maximum transport in October of 1.8 Sv. The California Undercurrent reaches its minimum (0.8 Sv) and turns equatorward in the spring. The Cali- fornia Countercurrent turns equatorward then as well; therefore, all flow through the Southern California re- gion becomes equatorward in the spring, allowing the California Current to dominate and transport cooler water farther south (Hickey, 1993; Hickey et ah, 2003). In the California Current system, strong El Nino years in the positive ENSO phase have been linked to increased downwelling, warmer SSTs, and a deepening of the thermocline observed off Southern California (Sette and Isaacs, 1960; McGowan, 1985; Caldeira et ah, 2005). During the warm, positive phase of the PDO, the California Current is weakened and the Counter- current is strengthened. This intensified current brings warmer waters farther north and west into and beyond the Southern California region, creating warm SST anomalies along the California coast. In contrast, dur- ing the cool, negative PDO phase, the California Cur- rent is stronger, bringing cool water farther south and east into the region (Mantua and Hare, 2002). A PDO regime shift from cool to warm occurred around 1977, before our study, and a shift back to a cool PDO may have occurred during the last decade starting in 1998- 99 (Peterson and Schwing, 2003; Zhang and McPhaden, 2006; Wang et al., 2010). Two long-term sets of ship-based surveys have been conducted in Southern California waters, making it an ideal region for this investigation. The California Coop- erative Oceanic Fisheries Investigations (CalCOFI) has been conducting quarterly cruises that have sampled a breadth of oceanographic and lower-trophic-level bio- logical data since 1949. Marine bird and mammal ob- servations were added in 1987 (Hyrenbach and Veit, 2003; Sydeman, et al., in press). The NOAA Southwest Fisheries Science Center (SWFSC) also regularly has conducted marine mammal abundance surveys that have included this region since 1979. Changes in SST have been linked to changes in all levels of the food web, from immediate phyto- and zoo- plankton responses to lagged alterations in numbers, diet, and even reproductive success of higher-level or- ganisms, such as Ashes, seabirds, and marine mammals (Tibby, 1937; Hubbs, 1948; McGowan, 1985; McGowan et ah, 2003; Sydeman et al., in press). It follows that small cetacean populations would respond to such vari- ations in SST, likely as a response to changes in prey populations, as has been shown for seabirds (Hyren- bach and Veit, 2003). In addition, population-level re- sponses to these fluctuations in temperature may pre- dict their reaction to future ocean conditions as global ocean temperatures rise. We investigated such responses by 8 species of small cetaceans across 30 years, using SST averages and anomaly indices as a proxy for environmental varia- tion on 3 temporal scales: seasonal (yearly), ENSO (2-7 years) and PDO (-30 years). We predicted that patterns in small cetacean occurrence and distribution within Southern California waters would follow simi- lar trends reported for seabirds (Hyrenbach and Veit, 2003; Yen et al., 2006; Sydeman et al., in press) and other cetaceans (Forney and Barlow, 1998; Becker et al., 2012). For small cetaceans off Southern Califor- nia, the following trends were predicted: 1) species as- semblages will differ depending on the dominant SST regime, 2) cold-water-associated species will be more abundant and broadly distributed when cold-water con- ditions prevail, 3) warm-water-associated species will dominate during warm-water conditions, and 4) the lat- ter 2 patterns will be compounded when SST fluctua- tions co-occur on multiple scales. Materials and methods Study area and survey methods Our study area was situated between 117°W and 125°W longitude and from 30°N to 35°N latitude (Fig. 1) and includes the Southern California Bight as well as deeper offshore waters. The Southern California Bight is a region of complex bathymetric features, including the Channel Islands and a series of deep basins and shallow ridges (Dailey et al., 1993). Beyond the steep 2000-m slope lies the ocean basin, with a mean depth of >3500 m. Three regions, associated with depth, were defined in the analyses for this study (Fig. 1): 1) the inshore and island region (with a mean depth <1100 m and a maximum depth <2000 m; 2) the slope region (with a mean depth of 1000-3200 m and a depth range of 500-3500 m); and the offshore region (with a mean depth >3500 m and a maximum depth >4000 m). The terms for these three regions will be used throughout the study. Henderson et al.: Effects of sea-surface temperature on the occurrence of small cetaceans off Southern California 161 Figure 1 Map of the study area located off Southern California in the eastern North Pacific Ocean, south of Point Conception and incor- porating the Channel Islands, in which small cetacean distributions were recorded during 1979-2009. Colored areas indicate 500-m depth contours. The blue area (mean depth <1 100 in, maximum depth <2000 m) was considered the inshore and island region, the green area (mean depth of 1000-3200 m, within a depth range of 500-3500 m) was considered the slope region, and the yellow area (mean depth >3500 m, maximum depth >4000 m) was considered the offshore region. We analyzed data from visual sightings of marine mammals from 105 separate survey cruises conducted by both CalCOFI and SWFSC from 1979 to 2009 (Fig. 2). During CalCOFI cruises from May 1987 to April 2004 (CalCOFIa) marine mammals were recorded as part of standardized CalCOFI top predator surveys that were focused primarily on marine birds. The strip-transect methods of Tasker et al. (1984) were fol- lowed. Observations were made with the naked eye by a single observer stationed on one side of the flying bridge or outside the main bridge. Marine mammals were recorded only if they occurred within the 300-m strip transect used for birds or within 1000 m of the vessel for large cetaceans. Each CalCOFI transect line extended from directly in front of the ship to 90° on the observation side. Group sightings of marine birds and mammals were summarized into 3-km bins, with the latitude and longitude determined for the centroid of each bin. Additional details of field methods are pro- vided in Veit et al. (1996; 1997), Hyrenbach and Veit (2003), and Yen et al. (2006). In July 2004, 2 dedicated marine mammal visual observers were added to the CalCOFI cruises (CalCO- Flb), and a standard line-transect protocol replaced the strip-transect protocol (Burnham et al., 1980; Buckland et al., 2001). A complete description of survey methods can be found in Soldevilla et al. (2006). Each observer monitored a 90° field of view from bow to abeam, one on each side of the ship, and alternated between scan- ning with Fujinon1 7x50 binoculars (Fujifilm Corp., To- kyo) and with the naked eye. Survey effort was calcu- lated on the basis of latitude and longitude at the start and end of each trackline. For all CalCOFI surveys, observations were made on daytime tracklines between stations, and no visual observation effort was conducted while the vessel was stationary. All visual effort was conducted in sea state conditions rated 5 or less on the Beaufort scale. Data used for analyses were generally from 4 surveys per year from 1987 to 2009, 1 survey per season (typically in the same month but with some variation). In 5 of these years, only 3 surveys were conducted. In 1998, surveys were carried out monthly to capture a time series of oceanographic measures in a strong El Nino year. However, to be consistent across all years for pur- poses of analysis, these cruise data were combined into 1 Mention of trade names or commercial companies is for iden- tification purposes only and does not imply endorsement by the National Marine Fisheries Service, NOAA. 162 Fishery Bulletin 112(2-3) Transect lines surveyed for all studies of small cetaceans off Southern California in 1979-2009. Orange lines indicate surveys conducted during California Cooperative Oceanic Fisheries Investigations (CC) cruises from 1987 to 2004, green lines indicate CC surveys conducted from 2004 to 2009, and purple lines indicate surveys conducted by the NOAA Southwest Fisheries Science Center (SWFSC) from 1979 to 2009. Black lines indicate latitude and longitude in 1° increments, which were used to create the grid sections for analyses in the generalized additive models. this analysis. The 3 warm-temperate and tropical spe- cies were short-beaked common dolphin ( Delphinus delphis), long-beaked common dolphin ( D . capensis), and striped dolphin ( Stenella coeruleoalba). The 3 cold- temperate species were Pacific white-sided dolphin ( Lagenorhynchus obliquidens), northern right whale dolphin (Lissodelphis borealis), and Dali’s porpoise ( Phocoenoides dalli). The remaining 2 species were considered cosmopolitan, distributed globally in tropi- cal and temperate waters: Risso’s dolphin (Grampus griseus) and bottlenose dolphin (Tursiops truncatus ) (Reeves et ah, 2002). All bottlenose dolphin sightings in this study were presumed to be offshore animals because most coast- al animals remain within about 1 km from the shore (Hanson and Defran, 1993) and surveys were conduct- ed at least 5-10 km from the coast. In addition to their individual species’ models, short- and long-beaked com- mon dolphins were combined into an additional Del- phinus species category because the 2 species were not recognized formally as distinct until 1994 (Heyning and Perrin, 1994). Furthermore, they were not distin- guished on SWFSC cruises before 1991 or on CalCOFI cruises before August 2004. Therefore, the data sets for long-beaked and short-beaked common dolphin are smaller than the data sets for all other species, and the data set for Delphinus spp. consists of all combined common dolphin sightings from all cruises. 4 quarters (winter, spring, summer, and fall; see the next section. Environmental data, for details). A full summary of surveys, along with total effort (in kilome- ters) and sightings per year for all species is provided in Appendix I. Data for analyses also came from 10 different SWF- SC cruises, conducted primarily in the summer and fall (from July through November) from 1979 through 2005 and covering an area that included Southern Califor- nia waters (Appendix I). For SWFSC cruises, standard line-transect protocols were followed, as described in Barlow and Forney (2007) and Kinzey et al.2 The latter cruises had 3 observers on the flying bridge, 2 of whom used “big eye” 25x150 binoculars to scan 90° from bow to abeam on either side of the flying bridge. The third observer monitored the entire forward 180° with 7x50 binoculars and the naked eye. Survey effort (in kilome- ters) was calculated either from the latitude and lon- gitude positions at the start and end of each trackline (1979-84 surveys) or from latitude and longitude posi- tions recorded approximately every 10 min along the track (1991-2005 surveys). Eight species of small cetaceans were examined in 2 Kinzey, D., P. Olson, and T. Gerrodette. 2000. Marine mam- mal data collection procedures on research ship line-transect surveys by the Southwest Fisheries Science Center. NOAA Southwest Fisheries Science Center, Admin. Rep. LJ-00-08, 32 p. Henderson et al . : Effects of sea-surface temperature on the occurrence of small cetaceans off Southern California 163 Environmental data Three variables were used to represent variations in SST on different temporal scales: quarterly SST aver- ages, ENSO indices, and PDO indices. Monthly aver- aged SST data from 1985 to 2009 were from NOAA Advanced Very High Resolution Radiometer (AVHRR) Pathfinder satellite data, which have a spatial resolu- tion of ~4.1-km (http://www.nodc.noaa.gov/Satellite- Data/pathfinder4km). For 1981-84, NOAA AVHRR data (multichannel averaged SST with a 5.7-km reso- lution) were also used. No satellite data were avail- able before 1981; therefore, a missing data filter and a single imputation method were used to create val- ues for 1979 and 1980 with the mean of the SSTs for the other years (Hastie, 1991; Nakagawa et al., 2001). With Windows Image Manager, vers. 6 (WimSoft, San Diego, CA), seasonal SST averages were calculated from the monthly SST data. These SST averages were estimated for each quarter and each grid cell (see the following paragraph) for the period 1979-2009 (spring: February-April; summer: May-July; fall: August- October; winter: November-January ). NOAA ENSO anomaly data, derived from the Oceanic Nino Index as a 3-month running mean of SST anomalies from 1971 to 2009 in the Nino 3.4 region (http://www.cpc.ncep. noaa.gov/products/analysis_monitoring/ensostuff/ ensoyears.shtml), were used as a proxy for ENSO for 1979-2009. The Nino 3.4 region is centered on the equator; therefore, the index indicates the relative strength of the ENSO event rather than SST anomaly values for Southern California waters. PDO anomaly data, averaged for the period from 1900 to 2009, from the University of Washington (http://jisao. Washing- ton. edu/pdo) were used as a proxy for the PDO regime from 1979 to 2009. The PDO index is derived from a monthly averaged SST for North Pacific waters pole- ward of 20°N. Depth data were taken from the NOAA Nation- al Geophysical Data Center’s ET0P02 2-min global relief database (http://www.ngdc.noaa.gov/mgg/ fliers/06mgg01.html). The study area was divided into 52 grid cells of 1° (111 km or 60 nmi) latitude by 1° longitude, leading to grid cell areas that ranged from 2940 to 3120 km2. The gridded depth data were then assigned to each of the grid cells, and minimum, maximum, and mean depth values were calculated for each grid cell, along with the maximum seafloor slope per cell. These large grid cells correspond to approxi- mately one day of effort for each of the surveys and were designed to be large enough to smooth out the mesoscale features that occur on shorter temporal and spatial scales than were of interest here. Although mesoscale features, such as fronts or eddies, are of- ten observed to be hotspots for marine mammals, the multidecadal data set used in our study allowed for a synoptic examination of changing distribution patterns throughout the study area. Modeling cetacean sighting rates Generalized additive models (GAMs) of species sight- ing rates as a function of the temperature data and depth values were created with the mixed GAM com- putational vehicle (mgcv) package in R software, vers. 2.14.2 (R Core Team, 2012) (Hastie and Tibshirani, 1990; Wood, 2006). GAMs use a link function to relate the predictor variables to the mean of the response variable. GAMs also allow nonparametric functions to be fitted to the predictor variables through the use of a smoothing function to describe the relationship be- tween the predictor and the response variables (Has- tie and Tibshirani, 1990). For model development, the grid cells described previously were used as data units, and all effort, sighting, and seasonal SST data were calculated for each cell. This approach allowed for the normalization of spatial and temporal differences in survey data. The type of survey was included as a categorical vari- able to account for differences in sighting rates due to survey method and platform. For example, because many vessels of different heights were used and heights for some vessels were not reported, standard- ization of observations for platform heights was not possible. Survey types included SWFSC (1979-2005), CalCOFIa (1987-2004), and CalCOFIb (2004-09). For each survey type, the number of group sightings of each species within each 1° cell, standardized by the log of the amount of effort per cruise (in kilometers), was modeled by assuming a Poisson distribution with a log link function. Potential predictor variables in the model were the following: seasonal SST averages of each grid section (SeasAv); ENSO index (ENSO); PDO index (PDO); the mean (DepthMean), minimum (DepthMin), and maxi- mum (DepthMax) depth (in meters) for each grid sec- tion; the maximum slope for each grid section (Slope); and the quarter (Quarter) as a categorical variable for identification of interannual patterns. Although sea state has been shown to be an important predic- tor of sighting rates in other cetacean habitat and trend models (Becker, 2007), the condition of the sea surface was not recorded in early CalCOFI observa- tions and, therefore, sea state was not included in this analysis. Instead, only data recorded when the sea state was rated 0-3 on the Beaufort scale during SWFSC cruises and later CalCOFI cruises were used to standardize for differences in survey effort and, thus, make the different platforms as comparable as possible. We used the number of group sightings, rather than the number of individuals, as our measure of relative encounter rate, essentially creating encounter rate models of group sightings per unit (kilometer) of survey effort (SPUE) (Bordino et al., 1999; Stockin, 2008). A correlation analysis of annual rates of group sightings in relation to mean group size per year 164 Fishery Bulletin 112(2-3) Results Sea-surface temperature for the study area over the period 1979-2009 ranged from 12.7°C to 19.4°C, with a mean of 16.2°C. Overall averaged seasonal anoma- lies ranged from -1.5°C to 1.1°C around the mean (Fig. 3). In comparison, seasonal anomalies by grid section ranged from -3.8°C to 3.4°C. Years with a strong posi- tive PDO (index>l) were 1983, 1987, 1993, 1997, and 2003, and a strong negative PDO (index<-l) occurred in 1999 and 2008 (Fig. 3). Strong positive ENSO years were 1982-83, 1987-88, 1991-92, 1997-98, and 2002- 03, and strong negative ENSO years were 1988-89 and 1999-2000 (Fig. 3). No long-term trends in SST were apparent in our data given the levels of seasonal and ENSO variation observed. However, a linear regression of PDO anomaly data shows an overall negative trend in the last 30 years: coefficient of multiple determina- tion (i?2)=0.215, P=0.009. This pattern is likely a result of the PDO regime switch in the last decade (Overland et al., 2008; Hodgkins, 2009). The correlation analysis revealed that annual sight- ing rates and mean group size were not correlated for any of the species examined. This finding indicates that, although species may be encountered with vary- ing frequency across years, the number of individual animals per group does not change in a correlated way. For example, if more groups of a given species were also was conducted to determine whether group size correlated with the number of groups encountered. To select predictor variables for inclusion in each model, a likelihood-based smoothness selection meth- od, instead of a traditional stepwise method, was ap- plied with the restricted maximum likelihood (REML) criterion (Patterson and Thompson, 1971; Wood, 2006). Each predictor variable was tested for inclusion in the model with a tensor product approach coupled with a smoothing function defined by a cubic regression spline with shrinkage. The best model was selected on the basis of a combination of the information-the- oretic descriptor Akaike’s information criterion (AIC; Akaike, 1976) and REML. Next, an interactive term selection method was applied to sequentially drop the single term with the highest nonsignificant P-value and then refit the model until all terms were signifi- cant. The best-fit model was therefore one that mini- mized AIC and maximized REML and the explained deviance and that included only significant predictor variables. In addition, the ENSO, PDO, and seasonal SST averages, as well as each of the depth metrics, were tested for correlation if more than one of them was included in a model as a significant predictor. These variables were then included together only if they were not correlated. If the variables were cor- related, then only the most significant variable re- mained in the final model. Henderson et al.: Effects of sea-surface temperature on the occurrence of small cetaceans off Southern California 165 Table 1 The final best-fit generalized additive models are presented here for each of the 8 species of small cetaceans investigated for this study in Southern California waters in 1979-2009. Also included are the restricted maximum likelihood (REML) score, explained deviance (Expl. dev.), and residual degrees of freedom (df) for each model. See Appendix 2 for the P-values of each variable in these models. The 8 species were the short-beaked common dolphin ( Delphinus delphis), long-beaked common dolphin ( D . capensis ), Risso’s dolphin ( Grampus griseus), northern right whale dolphin ( Lissodelphis borealis), Pacific white-sided dolphin (Lagenorhynchus obliquidens ), Dali’s porpoise ( Phocoenoides dalli), striped dolphin ( Stenella coeruleoalba), and bottlenose dolphin (Tursiops truncatus); a third model for common dolphins incorporated data for both the short- and long-beaked common dolphins. Variable abbreviations: DepthMin=minimum depth (m), DepthMean=mean depth (m), MaxDepth=maximum depth (m), SeasAv=seasonal averaged sea-surface temperature, ENSO=El Nino-Southern Oscillation, and PDO=Pacific Decadal Oscillation. Species Final model REML Expl. dev. Residual df Short-beaked common dolphin Quarter + Slope + DepthMean + SeasAv 754.0 23.9% 642 Long-beaked common dolphin Quarter + DepthMax + ENSO + SeasAv 211.7 57.4% 652 Both common dolphins Quarter + SurveyType + Slope + DepthMax + PDO + SeasAv 2751.8 32.5% 2415 Risso’s dolphin Quarter + SurveyType + Slope + DepthMean + ENSO + SeasAv 644.7 36.6% 2421 Northern right whale dolphin SurveyType + DepthMax + ENSO + SeasAv 270.3 26.1% 2428 Pacific white-sided dolphin Quarter + SurveyType + Slope + DepthMean + PDO 706.2 21.2% 2419 Dali’s porpoise Quarter + SurveyType + Slope + DepthMean + PDO + SeasAv 726.8 27.5% 2423 Striped dolphin SurveyType + DepthMin 88.3 41.8% 2437 Bottlenose dolphin SurveyType + Slope + DepthMean + PDO 376.0 46.4% 2429 encountered in a year, the group size would not neces- sarily also increase or decrease. The best-fit models are shown in Table 1. Values for explained deviance ranged between 21.2% and 57.4% across species. A summary of group sighting rates is given in Table 2. Six of the 9 models included the Quarter variable, indicating intra-annual variation in the SPUE for each species. Of the 9 models, 7 models included the SurveyType variable; the 1987-2004 Cal- COFI cruises ranked lowest and the SWFSC cruises ranked highest in sighting numbers for most species. Of the 9 models, 6 models included the seasonal SST average variable, and 7 models also included either the PDO or ENSO index. The latter results indicate the importance of those temperature fluctuations on small cetacean distribution. All models also included at least one depth metric, previously shown to be an important predictor variable for Southern California cetaceans (e.g., Becker, 2007). Finally, 5 of the 9 models included slope as a predictor. Common dolphins Three different models were used for common dol- phins: both species of common dolphin in a single combined category, short-beaked common dolphin, and long-beaked common dolphin. The similarities in the model results for both common dolphins and the short-beaked common dolphin indicate that the data for the combined category likely are dominated by sightings of short-beaked common dolphins. Com- mon dolphins were associated with seasonal SSTs of about 14-18°C in all 3 models, indicating possible avoidance of extremely warm or cold temperatures (Fig. 4). For all common dolphin groups, most sight- ings occurred in the summer and fall, and generally the fewest sightings occurred in the spring. Depth was an important predictor of common dolphin dis- tribution in all 3 models, and slope was included in the models for the combined category and the short- beaked common dolphin. Long-beaked common dol- phins were found almost exclusively inshore, and sightings of short-beaked common dolphins and dol- phins in the combined group were recorded both in- shore and offshore in areas with shallow slopes. The model for both common dolphins combined showed a very slight increase in sightings with negative PDO anomalies, although the overall response was fairly flat (Fig. 4). 166 Fishery Bulletin 112(2-3) Table 2 Summary of sightings, including the number of cruises con- ducted by the California Cooperative Oceanic Fisheries Inves- tigations and the NOAA Southwest Fisheries Science Center in which each species was encountered and the total number of groups sighted in 1979-2009 in Southern California wa- ters for each studied species of small cetacean: short-beaked common dolphin (Delphinus delphis), long-beaked common dolphin ( D . capensis), Risso’s dolphin (Grampus griseus), northern right whale dolphin (LissocLelphis borealis ), Pacific white-sided dolphin ( Lagenorhynchus obliquidens), Dali’s por- poise ( Phocoenoides dalli), striped dolphin ( Stenella coeruleo- alba), and bottlenose dolphin ( Tursiops truncatus ). Because the short- and long-beaked common dolphins were not rec- ognized formally as distinct until 1994, data for both species were used in a combined category in analyses. Number of Number of Species cruises groups Short-beaked common dolphin 29 387 Long-beaked common dolphin 22 93 Both common dolphins 105 1537 Risso’s dolphin 74 227 Northern right whale dolphin 32 71 Pacific white-sided dolphin 62 217 Dali’s porpoise 64 240 Striped dolphin 22 28 Bottlenose dolphin 50 180 ed during later SWFSC (1991-2005) cruises, mak- ing Survey Type an important predictor variable. Bottlenose dolphin Bottlenose dolphin groups tended to display a strong inshore and island association. They gen- erally were sighted over the continental shelf, although they were occasionally observed farther offshore, as shown in the depth residuals plots (Fig. 5). The PDO variable was significant, indi- cating that a slight increase in sightings occurred with negative PDO anomalies. Northern right whale dolphin Northern right whale dolphin is 1 of 3 cold-tem- perate species strongly associated with the Cali- fornia Current system. Therefore, the extent of this species into the Southern California study area was expected to correlate with cold-water intrusions. Sightings were associated with cool SSTs as expected. However, sightings were asso- ciated also with both positive and negative ENSO anomalies. Groups of northern right whale dol- phins showed a strong association with the slope region, with most sightings located at depths be- tween 2000 and 4000 m, as shown in the depth residuals plot (Fig. 6). Risso's dolphin Risso’s dolphins were largely observed inshore, al- though they were occasionally observed offshore and in areas of shallow depths and steep slope, as shown in the partial residuals plots for depth and slope (Fig. 5). Sightings peaked slightly during warmer seasonal SSTs, around 18°C, but occurred least frequently in the summer. ENSO also was included in the model and in- dicated slightly more sightings during positive ENSO phases. Striped dolphin Striped dolphins are a tropical and warm-temperate species associated with warm water masses, and dol- phins of this species were predominantly observed off- shore of the 2000-m depth contour with a deep min- imum depth (Fig. 5). Because of this strong offshore distribution, only 28 groups were sighted during 22 cruises. This low number of sightings is in part due to the limitation of including only sightings made in sea states rated 3 or less on the Beaufort scale; be- cause most striped dolphin sightings occurred offshore, many were made in higher-rated sea states and were, therefore, not included. Because of that exclusion, most sightings included for analyses came from data collect- Dall's porpoise Sightings of Dali’s porpoise, another cold-temperate species, peaked during the spring, fall, and winter (Fig. 6), and groups of Dali’s porpoises were associated with cool SSTs. However, they were associated with slightly positive PDO phases, as well. They were distributed inshore and offshore, in areas of slightly shallower slopes, as shown in the depth and slope residuals plots. Pacific white-sided dolphin Results were unexpected for Pacific white-sided dol- phin, the final cold-temperate species the sighting rates of which were anticipated to increase in cooler temperatures. Sightings peaked slightly during the spring quarter when the water temperature was cool- er. However, they also exhibited an association with slightly positive PDO indices (Fig. 6). This species was distributed largely inshore, as shown in the depth re- siduals plot. Discussion Patterns of seasonal sea-surface temperatures Patterns of encounter rate related to seasonal SSTs were largely consistent with past studies within this region Henderson et al.: Effects of sea-surface temperature on the occurrence of small cetaceans off Southern California 167 A B C Slope m :> £ T ~ Q. ' v “ 1 1 1 1 1 “‘t-" -0 5 0 0 0.5 ENSO 1 0 1000 2000 3000 4000 Depth Mean 1000 2000 3000 DepthMean 4000 . CO - CM ™ “ X ^ CO — ^ " “ co _ Q Q_ 1 “ co _ -1 « »■ ■< ■» 1 in ■ 1 n 1 i . 11 » 1 Q CL ■ “ crT — co _ 12 16 SeasAv 20 -1 0 -0 5 “T 00 05 PDO 10 15 00 05 PDO 10 15 t 1 r CalCOFIa CalCOFIb SWFSC SurveyType SeasAv CalCOFIa CalCOFIb SWFSC SurveyType Figure 6 | Generalized additive model func- J tions of sightings per unit of effort 5 from 1979 to 2009 of (A) northern 76 right whale dolphin (Lissodelphis <5 borealis), (B) Dali’s porpoise ( Pho - coenoides dalli), and (C) Pacific white-sided dolphin (Lagenorhyn- chus obliquidens ) in relation to sea-surface temperature (SST) in- £ dices and depth variables in South- g- ern California waters. Estimated 2 degrees of freedom of the smooth ^ function are given (in parenthe- £ ses on the y-axis) for all included e 03 variables. Solid lines represent °- the marginal effect of the given variable after controlling for the other variables in the model. The shaded band represents 2 standard errors. SurveyType is the variable for the cruises: SWFSC, CalCOFIa ( 1979— 2004) and CalCOFIb (2004-09). Variable abbreviations: DepthMin=minimum depth (m), DepthMean=mean depth (m), MaxDepth=maximum depth (m), SeasAv=seasonal averaged SST (°C), ENSO=El Nino-Southern Oscillation, and PDO=Pacific Decadal Oscillation. CalCOFIa CalCOFb SWFSC SurveyType 170 Fishery Bulletin 112(2-3) (e.g., Dohl et al., 1986; Barlow, 1995; Forney and Barlow, 1998; Forney, 2000; Barlow and Forney, 2007; Becker, 2007). Risso’s and common dolphins preferred waters of intermediate and warmer temperature (14-20°C) (For- ney, 2000; Reeves et al., 2002; Becker, 2007). In con- trast, sightings of Dali’s porpoises, Pacific white-sided dolphins, and northern right whale dolphins peaked in the cool spring season or with cool SSTs. In addition, long-beaked common, bottlenose, and Risso’s dolphins and Dali’s porpoises showed a preference for inshore or island-associated waters. Short-beaked common and Pacific white-sided dolphins were observed both inshore and slightly offshore. Northern right whale dolphins were associated with the slope region, and striped dol- phins were observed only in deep offshore waters. The relationship between SST and depth is complex and dif- ficult to separate, and these models likely oversimplify the observed trends. However, these results do seem to indicate some habitat or resource partitioning is occur- ring because these small cetacean species presumably follow preferred water conditions and prey. Although the seasonal distribution patterns here are generally consistent with those found by Forney and Barlow (1998) for temperate species, an increase in common dolphin sightings was observed in that study in winter rather than in summer for 1991-92. In contrast, a summer peak in sightings for common dolphins was found by Dohl et al. (1986). Our results, however, support the findings of both of these studies. ENSO was included as a predictor in the model for the long-beaked common dolphin. The strong El Nino that occurred in 1991-92 may explain the increase in winter sightings for common dolphins in the surveys conducted by Forney and Barlow (1998) over that time period. If the winter of 1991-92 was uncharacteristi- cally warm, then there may have been more common dolphins present than usual at that time of year. In contrast, the 1975-78 surveys conducted by Dohl et al. (1986) overlapped with the 1976-77 PDO regime shift from cool to warm; this shift could account for the in- crease in common dolphins during the warmer summer months of 1975-78. Patterns of temperature oscillation Temperature fluctuation patterns like ENSO, PDO, and the North Atlantic Oscillation have been documented to affect the prey of marine animals. An example of this effect is the strong relationship between the North Atlantic Oscillation, the life cycle of the copepod Calanus finmarchicus, and the recruitment of larval Atlantic Cod (Gadus morhua) that prey on copepods (Stenseth et al., 2002). Atlantic Cod in turn are a ma- jor food source for the gray seal (Halichoerus grypus), and Calanus spp. are an important prey for the North Atlantic right whale (Eubalaena glacialis) (Wishner et al., 1995; Mohn and Bowen, 1996). Calanoid copepods in the California Current system also have exhibited pop- ulation-level step changes in abundance in response to strong ENSO events and PDO shifts (Rebstock, 2002). For example, during the PDO phase switch in the late 1970s, 28% of the copepod species sampled increased in abundance. In contrast, around 1990 a biological step change occurred in copepod populations, when 28% of the species declined in abundance. Population fluctuations of small pelagic fishes, such as anchovies ( Engraulis spp.) and Pacific Sardine ( Sardinops sagax ), are also correlated strongly with both ENSO and PDO indices in the California Current system and in the Peru-Chile Current (Tibby, 1937; Hubbs, 1948; Niquen and Bouchon, 2004; Lehodey et al., 2006). These fish species are prey for many species of cetaceans in the California Current, including the short- and long-beaked common dolphins, bottlenose dolphin, Pacific white-sided dolphin, and Dali’s por- poise (e.g., Stroud et ah, 1981; Walker and Jones, 1993; Heise, 1997; Amano et al., 1998; Osnes-Erie, 1999). Isolated instances of cetaceans changing their dis- tribution patterns have been noted during and after strong climatic events. One example is the permanent expansion of the northern extent of the range of coastal bottlenose dolphins along the California coast during the 1982-83 El Nino (Defran et al., 1999). Another ex- ample is the increased abundance of humpback whales (Megaptera novaeatigliae) in Monterey Bay during the 1997-98 El Nino (Benson et al., 2002). SST fluctuations have been shown to affect the distribution and com- munity composition of seabirds in the California Cur- rent system as well (Hyrenbach and Veit, 2003; Yen et al., 2006). A decline of 2% per year in overall seabird density was recorded for the last 25 years — a drop that was attributed to declines in nearshore abundance of forage fishes (Sydeman et ah, in press). The models for most species included the PDO and ENSO indices as significant variables, although they were not strong predictors in most cases. During posi- tive PDO and ENSO phases, upwelling and productiv- ity decrease while water temperature increases, partic- ularly as warm equatorial waters are pushed poleward and the California Current system is found closer in- shore (Sette and Isaacs, 1960; McGowan, 1985). These conditions may explain the apparent association of the Dali’s porpoise and Pacific white-sided dolphin with positive PDO indices. These species may be pushed closer to shore by the contraction of the California Current, or they could be concentrating in the remain- ing areas of productivity, as has been hypothesized for the increase in rorquals in Monterey Bay during the 1997-98 El Nino (Benson et ah, 2002). Alternately, the patterns observed here may re- flect changes that occur in other parts of these spe- cies’ ranges. For example, during negative, cool PDO phases, the overall range of warm-temperate species may contract southward; therefore, a slight increase in the number of common dolphins and even bottlenose dolphins may occur during this phase. Likewise, if the cold-temperate species range as far south as Baja dur- ing negative PDO and ENSO periods, then their ranges Henderson et al.: Effects of sea-surface temperature on the occurrence of small cetaceans off Southern California 171 may contract northward during positive PDO phases, leading to an increase in sightings of Dali’s porpoises and Pacific white-sided dolphins in Southern California waters. Implications in regard to climate change We have demonstrated changes in distributions of small cetaceans on scales of months to decades. De- spite a limited understanding of the mechanisms be- hind those changes, the model results may help cre- ate a basis for understanding the potential effect of climate change upon these species. Studies of climate change in the California Current system indicate that, in addition to increasing temperatures, a rise in atmo- spheric carbon dioxide levels is predicted to lead to more intense upwelling (Bakun, 1990; Snyder et al., 2003), stronger thermal stratification, and a deepening of the thermocline (Roemmich and McGowan, 1995). These changes may alter large-scale circulation pat- terns (Harley et al., 2006). Fluctuations in these physi- cal mechanisms will lead to changes in ecosystem dy- namics and biodiversity from primary producers to top predators (Sydeman et al., 2001; Harley et al., 2006; Hooff and Peterson, 2006). Globally, species associated with sea ice or with highly limited ranges are the most obvious species to be affected by changing ocean temperatures and sea levels (Moore and Huntington, 2008). However, even pelagic species, such as the ones discussed here, are likely to be affected (Learmonth et al., 2006; Simmonds and Eliott, 2009). For example, as water temperatures off Scotland increased, the abundance of common dol- phins increased, whereas the number of white-beaked dolphins ( Lagenorhynchus albirostris), which are asso- ciated with cold water, decreased. Such trends could indicate a poleward shift in range for both species (Ma- cLeod et al., 2005; Simmonds and Isaac, 2007). In addi- tion, an influx of cold freshwater in the northern Gulf of Mexico in 2011 may have contributed to an unusu- ally high mortality rate in bottlenose dolphins (Char- michael, et al. 2012). We predicted that the ranges of the common dol- phins, Risso’s dolphin, and bottlenose dolphin would expand northward as ocean temperatures warmed, es- pecially as seasonal, ENSO, and PDO events were com- pounded (e.g., a positive PDO with a positive ENSO). Conversely, we predicted that the ranges of the Pacific white-sided dolphin, northern right whale dolphin, and Dali’s porpoise would contract poleward and inshore. These patterns have held true for observations made during previous shorter-term studies. For example, Dali’s porpoises and Pacific white-sided dolphins domi- nated the odontocete species assemblage off central California in the decade before the strong El Nino of 1997-98 (Benson et al., 2002; Keiper et al., 2005). Keiper et al. (2005) noted that during the strong El Nino of 1997-98 there was a deepened thermocline, a narrow, inshore distribution of Pacific Sardine eggs, and an overall decrease in abundance of macrozooplankton. During that El Nino, sightings of Dali’s porpoises were greatly reduced, whereas common and Risso’s dolphin sightings increased. Furthermore, Pacific white-sided dolphin sightings decreased after this period, while sightings of common (particularly the long-beaked spe- cies) and bottlenose dolphins increased (Keiper et al., 2005). However, over the longer-term, our study showed an association of the Pacific white-sided dolphin and Dali’s porpoise with positive PDO indices, of common and bottlenose dolphins with negative PDO indices, and of the northern right whale dolphin with positive ENSO indices. These results indicate a more complicat- ed relationship between distribution patterns and SST than we allowed for in our initial predictions or that has been observed on shorter temporal scales. Contin- ued monitoring efforts should be made to ensure that future changes in distribution or reproductive success are documented. Model considerations The results presented here provide insight into long- term distribution trends of small cetaceans over sev- eral decades. The results are both supported by and build upon the current knowledge base for these spe- cies. Nonetheless, we recognize some caveats to this study that warrant discussion. The PDO and ENSO indices were developed with the use of broad regions of the Pacific. Therefore, they may not reflect precisely the specific dynamics of the South- ern California study area. The seasonal SSTs, although averaged for each grid section and quarter, were also still quite broad, as was the selected size of grid cells. However, this scope was used intentionally to capture the large temporal- and spatial-scale dynamics of these changing SST patterns, rather than to examine meso- scale dynamics on shorter temporal scales. In addition, the SST, ENSO, and PDO variables have the potential to be correlated, as the indices are similar over time. A correlation analysis was conducted, and correlations between ENSO and PDO and between seasonal SSTs and PDO were detected for some species. In those cas- es, they were not included together, and only the most significant predictors were included. Only one cruise occurred per year before 1987. To account for potential differences between Survey Types, we repeatedly reran each model while randomly drop- ping out data from different years. The results indicat- ed that the models were robust against missing years of data, and the variation in the number of surveys per year did not affect the results. The survey methods from each Survey Type were quite different, making it a challenge to combine these data sets. However, by using only the group SPUE and by limiting our sight- ings to the ones made in sea states of 3 or less on the Beaufort scale, we tried to make the data as compa- rable as possible. 172 Fishery Bulletin 112(2-3) The inclusion of the Survey Type variable in most models reflects some of those differences in survey ef- fort. The CalCOFI cruises in 1987-2004 consistently ranked lowest in sighting numbers for all species al- though those surveys had the most effort. This rank- ing was likely due to a single observer who covered both birds and mammals with a smaller effective strip width rather than to the multiple observers dedicated to monitoring marine mammals for the other 2 types of surveys. The SWFSC cruises had the highest number of ob- servations for 6 of the 7 models in which they were in- cluded, although those cruises had less effort than the CalCOFI cruises. The high number of observations may have been due to optimal sighting conditions during the SWFSC cruises, which were largely conducted in summer and fall. In addition, big eye binoculars were used on SWFSC cruises but were not used regularly on CalCOFI cruises. In another difference in Survey Type, CalCOFI surveys always were conducted in pass- ing mode in which the survey vessel does not leave the transect line when animals are sighted, but SWFSC ships operated in closing mode and could deviate from the transect line to confirm species. Finally, we used the number of groups sighted rath- er than the number of individuals observed as our met- ric for encounter rate. The correlation analysis did not indicate a strong relationship between the number of groups encountered and the size of the group for any of the modeled species. Therefore, our models may have misidentified trends if a change in group size as a re- sponse to any of these variables had better explanatory power than the overall encounter rates. Conclusions The models presented in this study indicated that fluc- tuations in SST regimes influenced the distribution of small cetaceans. However, the relationships were not as straightforward as predicted. The observed com- plexities likely are related to effects of SSTs on prey and subsequent responses by cetaceans. Dolphins have been shown previously to be sensitive to changes in SST and to shift their distributions in response to re- gime oscillations like ENSO. However, this study is the first one to model responses to multiple temperature shifts over a long time period for a variety of cetacean species in this region of the California Current system. The resulting models were unique to each of the 8 species studies. This finding indicates that each spe- cies is characterized by a distinct pattern in habitat occurrence related to SST dynamics in this study area, despite the overlap in the overall distributions of the examined species in the Southern California study area. Results herein can be used to begin to predict the future distribution of these small cetaceans through- out the waters off Southern California. Results also provide a tool to understand, as global climate change intensifies, potential responses of these species to ris- ing ocean temperatures and the ecological mechanisms responsible for those responses. Acknowledgments We are grateful to SWFSC and CalCOFI for the use of their data sets, without which this analysis could not have been conducted, and to all of the visual observers over the years who gathered that data. We thank N. Mantua and S. Hare for permission to use their PDO index and NOAA scientists for permission to use their ENSO index. We also thank M. Ferguson, E. Archer, J. Moore, and E. Becker for assistance with the modeling and M. Kahru for help with the satellite data of sea- surface temperatures and WimSoft software. Literature cited Akaike, H. 1976. An information criterion (AIC). Math. Sci. 14:5-9. Amano, M., M. Yoshioka, T. Kuramochi, and K. Mori. 1998. Diurnal feeding by Dali’s porpoise, Phocoenoides dalli. Mar. Mamm. Sci. 14:130-135. Bakun, A. 1990. Global climate change and intensification of coast- al ocean upwelling. Science 247:198-201. Ballance, L. T., R. L. Pitman, and P. C. Fiedler. 2006. Oceanographic influences on seabirds and ceta- ceans of the eastern tropical Pacific: a review. Prog. Oceanogr. 69:360-390. Barlow, J. 1995. The abundance of cetaceans in California waters. Part 1. Ship surveys in summer and all of 1991. Fish. Bull. 93:1-14. Barlow, J., and K. A. Forney. 2007. Abundance and population density of cetaceans in the California Current ecosystem. Fish. Bull. 105:509-526. Becker, E. A. 2007. Predicting seasonal patterns of California ceta- cean density based on remotely sensed environmental data. Ph.D. diss., 253 p. Univ. California Santa Bar- bara, Santa Barbara, CA. Becker, E. A., D. G. Foley, K. A. Forney, J. Barlow, J. V. Redfern, and C. L. Gentemann. 2012. Forecasting cetacean abundance patterns to en- hance management. Endang. Species Res. 16:97-112. Benson, S. R., D. A. Croll, B. B. Marinovic, F. P. Chavez, and J. T. Harvey. 2002. Changes in the cetacean assemblage of a coastal upwelling ecosystem during El Nino 1997-98 and La Nina 1999. Prog. Oceanogr. 54:279-291. Bordino, P., G. Thompson, and M. Iniguez. 1999. Ecology and behaviour of the franciscana (Ponto- poria blainvillei) in Bahia Anegada, Argentina. J. Ce- tacean Res. Manage. 1:213-222. Buckland, S. T., D. R. Anderson, K. P. Burnham, J. L. Laake, D. L. Borchers, and L. Thomas. 2001. Introduction to distance sampling estimating abundance of biological populations, 448 p. Oxford Univ. Press, Oxford, UK. Henderson et al.: Effects of sea-surface temperature on the occurrence of small cetaceans off Southern California 173 Burnham, K. P., D. R. Anderson, and J. L. Laake. 1980. Estimation of density from line transect sampling of biological populations. Wildl. Monogr. 72:1-202. Caldeira, R. M. A., P. Marchesiello, N. P. Nezlin, P. M. DiGia- como, and J. C. McWilliams. 2005. Island wakes in the Southern California Bight. J. Geophys. Res. 110:011012. Charmichael, R. H., W. M. Graham, A. Aven, G. Worthy, and S. Howden. 2012. Were multiple stressors a ‘perfect storm’ for north- ern Gulf of Mexico bottlenose dolphins ( Tursiops trunca- tus) in 2011? PLoS ONE 7(7):e41155. Dailey, M. D., J. W. Anderson, D. J. Reish, and D. S. Gorsline. 1993. The Southern California Bight: background and setting. In Ecology of the Southern California Bight: a synthesis and interpretation (M. D. Dailey, J. W. An- derson, and D. J. Reish, eds.), p. 1-18. Univ. California Press, Berkeley, CA. Defran, R. H., D. W. Weller, D. L. Kelly, and M. A. Espinosa. 1999. Range characteristics of Pacific coast bottlenose dolphins ( Tursiops truncatus) in the Southern Califor- nia Bight. Mar. Mamm. Sci. 15:381-393. Dohl, T. P, M. L. Bonnell, and R. G. Ford. 1986. Distribution and abundance of common dolphin, Delphinus delphus, in the Southern California Bight: a quantitative assessment based upon aerial transect data. Fish. Bull. 84:333-343. Forney, K. A. 2000. Environmental models of cetacean abundance: re- ducing uncertainty in population trends. Conserv. Biol. 14:1271-1286. Forney, K. A., and J. Barlow. 1998. Seasonal patterns in the abundance and distribu- tion of California cetaceans, 1991-1992. Mar. Mamm. Sci. 14:460-489. Gowans, S., B. Wursig, and L. Karczmarski. 2007. The social structure and strategies of delphinids: Predictions based on an ecological framework. Adv. Mar. Biol. 53:195-294. Hanson, M. T., and R. H. Defran. 1993. The behavior and feeding ecology of the Pacific coast bottlenose dolphin, Tursiops truncatus. Aquat. Mamm. 19:127-142. Harley, C. D. G., A. R. Hughes, K. M. Hultgren, B. G. Miner, J. B. Sorte, C. S. Thornber, L. F. Rodriguez, L. Tomanek, and S. L. Williams. 2006. The impacts of climate change in coastal marine systems. Ecol. Lett. 9:228-241. Hastie, T. J. 1991. Generalized additive models. In Statistical mod- els in S (J. M. Chambers and T. J. Hastie, eds.), p. 249- 307. Chapman and Hall/CRC, Boca Raton, FL. Hastie, T. J., and R. J. Tibshirani. 1990. Generalized additive models, 335 p. Chapman and Hall/CRC, Boca Raton, FL. Heise, K. 1997. Diet and feeding behavior of Pacific white-sided dolphins (Lagenorhynchus obliquidens) as revealed through the collection of prey fragments and stomach content analyses. Rep. Int. Whal. Commn. 47:807-815. Heyning, J. E., and W. F. Perrin. 1994. Evidence for two species of common dolphin (Ge- nus Delphinus ) from the eastern North Pacific. Nat. Hist. Mus. Los Ang. Cty., Contrib. Sci. 442:1-35. Hickey, B. M. 1993. Physical Oceanography. In Ecology of the South- ern California Bight: a synthesis and interpretation (M. D. Dailey, D. J. Reish, and J. W. Anderson, eds.), p. 19-70. Univ. California Press, Berkeley, CA. Hickey, B. M., E. L. Dobbins, and S. E. Allen. 2003. Local and remote forcing of currents and tempera- ture in the central Southern California Bight. J. Geo- phys. Res. 108:1-26. Hodgkins, G. A. 2009. Streamflow changes in Alaska between the cool phase (1947-1976) and the warm phase (1977-2006) of the Pacific Decadal Oscillation: the influence of gla- ciers. Water Resour. Res. 45:6502-6507. Hooff, R. C., and W. T. Peterson. 2006. Copepod biodiversity as an indicator of changes in ocean and climate conditions of the northern California current ecosystem. Limnol. Oceanogr. 51:2607-2620. Hubbs, C. L. 1948. Changes in the fish fauna of western North Amer- ica correlated with changes in ocean temperature. J. Mar. Res. 7:459-382. Hyrenbach, K. D., and R. R. Veit. 2003. Ocean warming and seabird communities of the southern California Current System (1987-98): re- sponse at multiple temporal scales. Deep Sea Res. (II Top. Stud. Oceanogr.) 50:2537-2565. Keiper, C. A., D. G. Ainley, S. G. Allen, and J. T. Harvey. 2005. Marine mammal occurrence and ocean climate off central California, 1986 to 1994 and 1997 to 1999. Mar. Ecol. Prog. Ser. 289:285-306. Learmonth, J. A., C. D. MacLeod, M. B. Santos, G. J. Pierce, H. Q. P. Crick, and R. A. Robinson. 2006. Potential effects of climate change on marine mammals. Oceanogr. Mar. Biol. 44:431-464. Leatherwood, S., R. R. Reeves, A. E. Bowles, B. S. Stewart, and K. R. Goodrich. 1984. Distribution, seasonal movements and abundance of Pacific white-sided dolphins in the Eastern North Pa- cific. Sci. Rep. Whales Res. Inst. 35:129-157. Lehodey, P, J. Alheit, M. Barange, T. Baumgartner, G. Beau- grand, K. Drinkwater, J.-M. Fromentin, S. R. Hare, G. Ot- tersen, R. I. Perry, C. Roy, C. D. van der Lingen, and F. Werner. 2006. Climate variability, fish, and fisheries. J. Clim. 19:5009-5030. MacLeod, C. D., S. M. Bannon, G. J. Pierce, C. Schweder, J. A. Learmonth, J. S. Herman, and R. J. Reid. 2005. Climate change and the cetacean community of north-west Scotland. Biol. Conserv. 124:477-483. MacLeod, C. D., C. R. Weir, M. Begona Santos, and T. E. Dunn. 2008. Temperature-based summer habitat partitioning between white-beaked and common dolphins around the United Kingdom and Republic of Ireland. J. Mar. Biol. Assoc. U.K. 88:1193-1198. Mantua, N. J., and S. R. Hare. 2002. The Pacific Decadal Oscillation. J. Oceanogr. 58:35-44. McGowan, J. A. 1985. El Nino 1983 in the Southern California Bight. In El Nino North: Nino effects in the eastern subarctic Pa- cific Ocean (W. S. Wooster and D. L. Fluharty, eds.), p. 166-184. Washington Sea Grant Program, Univ. Wash- ington, Seattle, WA. 174 Fishery Bulletin 112(2-3) McGowan, J. A., S. J. Bograd, R. J. Lynn, and A. J. Miller. 2003. The biological response to the 1977 regime shift in the California Current. Deep Sea Res. (II Top. Stud. Oceanogr. ) 50:2567-2582. Mohn, R., and W. D. Bowen. 1996. Grey seal predation on the eastern Scotian Shelf: modelling the impact on Atlantic cod. Can. J. Fish. Aquat. Sci. 53:2722-2738. Moore, S. E., and H. P. Huntington. 2008. Arctic marine mammals and climate change: im- pacts and resilience. Ecol. Appl. 18:157-165. Nakagawa, S., J. R. Waas, and M. Miyazaki. 2001. Heart rate changes reveal that little blue penguin chicks ( Eudyptula minor) can use vocal signatures to discriminate familiar from unfamiliar chicks. Behav. Ecol. Sociobiol. 50:180-188. Niquen, M., and M. Bouchon. 2004. Impact of El Nino events on pelagic fisheries in Peruvian waters. Deep Sea Res. (II Top. Stud. Ocean- ogr.) 51:563-574. Osnes-Erie, L. D. 1999. Food habits of common dolphin ( Delphinus delphis and D. capensis) off California. M.S. thesis, 56 p. San Jose State Univ., Moss Landing, CA. Overland, J., S. Rodionov, S. Minobe, and N. Bond. 2008. North Pacific regime shifts: definitions, issues and recent transitions. Prog. Oceanogr. 77:92-102. Patterson, H. D., and R. Thompson. 1971. Recovery of interblock information when block sizes are unequal. Biometrika 58:545-554. Peterson, W. T., and F. B. Schwing. 2003. A new climate regime in northeast Pacific eco- systems. Geophys. Res. Lett. 30(17), 1896, doi: 10. 1029/2003GL0 17528 R Core Team. 2012. R: a language and environment for statistical com- puting. R Foundation for Statistical Computing, Vienna, Austria. [Available from http://www.R-project.org, ac- cessed September 2013.] Rebstock, G. A. 2002. Climatic regime shifts and decadal-scale variabil- ity in calanoid copepod populations off southern Califor- nia. Global Change Biol. 8:71-89. Reeves, R. R., B. S. Stewart, P. J. Clapham, and J. A. Powell. 2002. Guide to marine mammals of the world, National Audibon Field Guide Series, 527 p. Alfred A Knopf, New York. Reilly, S. B. 1990. Seasonal changes in distribution and habitat dif- ferences among dolphins in the eastern tropical Pacif- ic. Mar. Ecol. Prog. Ser. 66:1-11. Roemmich, D., and J. A. McGowan. 1995. Climatic warming and the decline of zooplankton in the California Current. Science 267:1324-1326. Sette, O. E., and J. D. Isaacs. 1960. The changing Pacific Ocean in 1957 and 1958. CalCOFI Rep. 7:13-217. Shane, S. H. 1995. Relationship between pilot whales and Risso’s dol- phins at Santa Catalina Island, California, USA. Mar. Ecol. Prog. Ser. 123:5-11. Shane, S. H., R. S. Wells, and B. Wiirsig. 1986. Ecology, behavior and social organization of the bottlenose dolphin: a review. Mar. Mamm. Sci. 2:34-63. Simmonds, M. P., and W. J. Eliott. 2009. Climate change and cetaceans: concerns and recent developments. J. Mar. Biol. Assoc. U. K. 89:203-210. Simmonds, M. P., and S. J. Isaac. 2007. The impacts of climate change on marine mam- mals: early signs of significant problems. Oryx 41:19-26. Snyder, M. A., L. C. Sloan, N. S. Diffenbaugh, and J. L. Bell. 2003. Future climate change and upwelling in the Cali- fornia Current. Geophys. Res. Lett. 30:1823-1827. Soldevilla, M. S., S. M. Wiggins, J. Calambokidis, A. Douglas, E. M. Oleson, and J. A. Hildebrand. 2006. Marine mammal monitoring and habitat inves- tigations during CalCOFI surveys. CalCOFI Rep. 47:79-91. Stenseth, N. C., A. Mysterud, G. Ottersen, J. W. Hurrell, K.-S. Chan, and M. Lima. 2002. Ecological effects of climate fluctuations. Science 297:1292-1296. Stockin, K. A. 2008. Factors affecting the occurrence and demograph- ics of common dolphins (Delphinus sp.) in the Hauraki Gulf, New Zealand. Aquat. Mamm. 34:200-211. Stroud, R. N., C. H. Fiscus, and H. Kajimura. 1981. Food of the Pacific white-sided dolphin, Lageno- rhynchus obliquidens, Dali’s porpoise, Phocoenoides dalli, and northern fur seal, Callorhinus ursinus, off California and Washington. Fish. Bull. 78:951-959. Sydeman, W. J., M. M. Hester, J. A. Thayer, F. Gress, P. Martin, and J. Buffa. 2001. Climate change, reproductive performance and diet composition of marine birds in the southern Cali- fornia Current system, 1969-1997. Prog. Oceanogr. 49:309-329. Sydeman, W. J., S. A. Thompson, J. A. Santora, J. A. Koslow, R. Goericke, and M. D. Ohman. in press. Climate-ecosystem change off southern Cali- fornia: seabird numerical responses and regime-specific predator-prey interactions. Deep Sea Res. (II Top. Stud. Oceanogr.). Tasker, M. L., P. H. Jones, T. Dixon, and B. F. Blake. 1984. Counting seabirds at sea from ships: a review of methods employed and a suggestion for a standardized approach. The Auk 101:567-577. Tibby, R. B. 1937. The relation between surface water temperature and the distribution of spawn of the Calfornia sar- dine. Calif. Fish Game 23:132-137. Veit, R. R., P. Pyle, and J. A. McGowan. 1996. Ocean warming and long-term change in pelagic bird abundance within the California current sys- tem. Mar. Ecol. Prog. Ser. 139:11-18. Veit, R. R., J. A. McGowan, D. G. Ainley, T. R. Wahls, and P. Pyle. 1997. Apex marine predator declines ninety percent in association with changing oceanic climate. Global Change Biol. 3:23-28. Walker, W. A., and L. L. Jones. 1993. Food habits of northern right whale dolphin, Pa- cific white-sided dolphin, and northern fur seal caught in the high seas driftnet fisheries of the North Pacific Ocean, 1990. Int. N. Pac. Fish. Comm. Bull. 53:285-295. Wang, X. J., R. Murtugudde, and R. Le Borgne. 2010. Climate driven decadal variations of biological Henderson et al.: Effects of sea-surface temperature on the occurrence of small cetaceans off Southern California 175 production and plankton biomass in the equatorial Pacific Ocean: Is this a regime shift? Biogeosci. Disc. 7:2169-2193. Wishner, K. F., J. R. Schoenherr, R. Beardsley, and C. Chen. 1995. Abundance, distribution and population structure of the copepod Calanus finmarchicus in a springtime right whale feeding area in the southwestern Gulf of Maine. Cont. Shelf Res. 15:475-507. Wood, S. N. 2006. Generalized additive models: an introduction with R, 392 p. Chapman and Hall/CRC, Boca Raton, FL. Appendix I Search effort (in linear kilometers) and number of groups seen for each species on each of the surveys conducted by the NOAA Southwest Fisheries Science Center (SWF- SC) and the California Cooperative Oceanic Fisheries Investigations (CalCOFI) and included in the analyses for this study. El Nino cruises were combined into sea- sons for analyses. Seasons are defined as follows: spring was February-April, summer was May-July, fall was August-October, and winter was November-January. Species abbreviations are as follows: Dsp -Delphinus Yen, P. P. W., W. J. Sydeman, S. J. Bograd, and K. D. Hyrenbach. 2006. Spring-time distributions of migratory marine birds in the southern California Current: oceanic eddy associations and coastal habitat hotspots over 17 years. Deep Sea Res. (II Top. Stud. Oceanogr.) 53:399-418. Zhang, D., and M. J. McPhaden. 2006. Decadal variability of the shallow Pacific meridio- nal overturning circulation: relation to tropical sea sur- face temperatures in observations and climate change models. Ocean Model. 15:250-273. sp.; Dd -Delphinus delphis; Dc -Delphinus capen- sis\ Gg-Grampus griseus; Lb =Lissodelphis borealis ; Lo =Lagenorhynchus obliquidens; Pd =Phocoenoides dalli; Sc =Stenella coeruleoalba\ Tt =Tursiops truncatus. NA=not available. The SWFSC cruises are as follows: CAMMS=The California Marine Mammal Survey; PODS =Population of Delphinus Stocks; ORCAWALE=Oregon, California, Washington Line-Transect and Ecosystem cruise; CSCAPE=The Collaborative Survey of Cetacean Abundance and the Pelagic Ecosystem. Effort Cruise Year Quarter Dsp Gg Lb Lo Pd Sc Tt Dd Dc (km) CalCOFI CC198705 1987 Spring 5 1 0 0 4 0 3 NA NA 1559 CC198709 1987 Summer 9 4 0 5 0 0 3 NA NA 1704 CC198711 1987 Fall 3 1 0 1 2 0 3 NA NA 1468 CC198801 1988 Winter 3 2 0 2 3 0 3 NA NA 1501 CC198804 1988 Spring 3 1 4 0 0 0 3 NA NA 1346 CC 198808 1988 Summer 15 0 0 0 3 0 3 NA NA 1810 CC198810 1988 Fall 3 7 0 0 9 0 3 NA NA 1420 CC198901 1989 Winter 2 0 1 0 2 0 3 NA NA 1338 CC198904 1989 Spring 1 0 0 3 7 0 3 NA NA 1596 CC198907 1989 Summer 25 3 1 1 1 0 3 NA NA 1932 CC198911 1989 Fall 11 4 1 1 4 0 3 NA NA 1496 CC199003 1990 Winter 1 0 1 0 1 0 3 NA NA 407 CC199004 1990 Spring 13 0 0 5 3 1 3 NA NA 1509 CC199007 1990 Summer 18 2 0 0 1 0 3 NA NA 1887 CC 1990 11 1990 Fall 5 3 0 1 4 0 3 NA NA 1349 CC199101 1991 Winter 3 1 0 1 3 0 3 NA NA 1332 CC199103 1991 Spring 7 2 0 1 5 0 3 NA NA 1162 CC199107 1991 Summer 28 0 0 2 0 0 3 NA NA 1668 CC199109 1991 Fall 12 2 0 0 0 0 3 NA NA 1635 CC199201 1992 Winter 5 1 2 3 1 0 3 NA NA 1265 CC199204 1992 Spring 5 2 0 1 2 0 3 NA NA 2427 CC199207 1992 Summer 18 1 1 2 1 0 3 NA NA 1437 CC199209 1992 Fall 5 0 0 0 2 0 3 NA NA 1625 CC199301 1993 Winter 10 1 0 1 1 0 3 NA NA 1249 CC199303 1993 Spring 9 2 0 0 0 0 0 NA NA 1630 CC199308 1993 Summer 9 0 0 0 0 0 0 NA NA 1843 CC199310 1993 Fall 2 1 0 0 2 0 0 NA NA 1549 CC199401 1994 Winter 12 2 2 0 4 0 1 NA NA 1369 CC199403 1994 Spring 2 0 0 0 0 0 0 NA NA 1552 CC199410 1994 fall 15 6 0 2 0 0 2 NA NA 1590 CC199501 1995 Winter 15 2 0 0 1 0 0 NA NA 1331 176 Fishery Bulletin 112(2-3) Effort Cruise Year Quarter Dsp Gg Lb Lo Pd Sc Tt Dd Dc (km) CC199504 1995 Spring 12 2 0 0 2 0 0 NA NA 1629 CC199507 1995 Summer 26 1 0 2 1 0 0 NA NA 1900 CC199510 1995 Fall 9 1 0 1 0 0 0 NA NA 1589 CC199604 1996 Spring 6 1 0 0 1 0 1 NA NA 1214 CC199608 1996 Summer 8 0 0 3 0 0 0 NA NA 1729 CC199610 1996 Fall 4 0 0 0 0 0 0 NA NA 1434 CC199701 1997 Winter 7 8 1 5 3 0 0 NA NA 1442 CC199707 1997 Summer 25 3 0 0 1 0 2 NA NA 1724 CC199709 1997 Fall 9 1 0 2 0 0 1 NA NA 1511 CC199712 1997 El Nino 1 2 3 0 0 0 0 0 NA NA 361 CC199801 1998 Winter 10 1 0 0 0 0 0 NA NA 696 CC199803 1998 El Nino 2 6 2 0 4 1 0 1 NA NA 701 CC199804 1998 Spring 13 1 0 2 0 0 1 NA NA 1491 CC199805 1998 El Nino 3 14 0 0 1 1 0 2 NA NA 818 CC199806 1998 El Nino 4 14 1 1 2 0 0 1 NA NA 812 CC199807 1998 Summer 25 0 0 1 0 0 1 NA NA 1652 CC199809 1998 Fall 9 1 0 0 0 0 0 NA NA 1499 CC199810 1998 El Nino 5 21 0 0 1 0 0 0 NA NA 1308 CC 199904 1999 Spring 9 1 3 3 1 0 0 NA NA 1633 CC 199908 1999 Summer 33 2 0 0 0 0 1 NA NA 1457 CC199910 1999 Fall 17 1 0 0 1 0 0 NA NA 1212 CC200004 2000 Spring 9 0 0 3 5 0 2 NA NA 1667 CC200007 2000 Summer 9 0 0 10 1 0 3 NA NA 1754 CC200010 2000 Fall 9 0 2 4 1 0 2 NA NA 1425 CC200101 2001 Winter 7 2 0 1 1 0 0 NA NA 1434 CC200104 2001 Spring 5 2 1 1 2 0 0 NA NA 1428 CC200107 2001 Summer 16 0 0 0 0 0 0 NA NA 1547 CC200110 2001 Fall 25 3 0 5 0 0 0 NA NA 1322 CC200201 2002 Winter 7 4 3 0 3 0 1 NA NA 1172 CC200203 2002 Spring 6 0 0 5 4 0 0 NA NA 1454 CC200207 2002 Summer 23 1 0 2 2 0 4 NA NA 1741 CC200211 2002 Fall 7 0 0 0 0 0 0 NA NA 1443 CC200301 2003 Winter 14 2 0 1 2 0 0 NA NA 1712 CC200304 2003 Spring 6 3 1 7 12 0 0 NA NA 3503 CC200307 2003 Summer 16 4 1 0 1 0 0 NA NA 1680 CC200310 2003 Fall 12 0 0 0 0 0 0 NA NA 1542 CC200401 2004 Winter 16 1 0 0 4 0 5 NA NA 1380 CC200403 2004 Spring 13 6 3 7 14 0 0 NA NA 2301 CC200407 2004 Summer 21 2 0 6 1 2 0 16 0 2003 CC200411 2004 Fall 19 2 0 6 0 0 2 8 8 1552 CC200501 2005 Winter 16 2 1 4 3 0 0 11 1 1376 CC200504 2005 Spring 7 4 6 13 4 0 2 0 4 2024 CC200507 2005 Summer 64 0 0 3 0 0 0 16 18 2264 CC200511 2005 Fall 32 5 1 1 1 0 1 10 7 1357 CC200602 2006 Winter 4 0 0 4 0 0 7 6 4 1292 CC200604 2006 Spring 6 3 2 3 8 0 3 1 1 2070 CC200607 2006 Summer 53 0 0 0 0 0 0 41 3 1964 CC200610 2006 Fall 17 0 1 3 1 1 4 11 2 1731 CC200707 2007 Winter 42 7 0 1 0 0 0 14 10 2180 CC200711 2007 Spring 22 0 2 2 1 0 1 12 0 1630 CC200701 2007 Summer 20 1 0 1 7 0 0 14 0 1454 CC200704 2007 Fall 9 4 1 2 9 0 2 2 1 900 CC200801 2008 Winter 15 4 1 5 4 0 0 8 0 1264 CC200803 2008 Spring 19 2 2 6 22 0 2 13 2 1182 CC200808 2008 Summer 31 1 0 1 0 0 6 10 3 1224 CC200810 2008 Fall 30 2 0 2 0 0 1 21 1 1505 CC200901 2009 Winter 30 1 0 0 13 0 4 18 3 1273 CC200903 2009 Spring 14 0 0 0 2 0 1 5 4 707 CC200907 2009 Summer 34 7 0 0 0 1 7 9 9 931 CC200911 SWFSC 2009 Fall 12 2 1 1 1 0 0 6 1 713 564 1979 Sept-Oct 17 8 1 1 2 1 3 NA NA 1662 Henderson et al.: Effects of sea-surface temperature on the occurrence of small cetaceans off Southern California 177 Effort Cruise Year Quarter Dsp Gg Lb 646 1980 June-July 8 0 0 798 1982 April 16 15 11 674 1983 Dec 19 7 0 905 1984 Dec 42 13 1 CAMMS 1991 July-Oct 50 8 10 PODS 1993 July-Oct 23 3 0 ORCAWALE 1996 Aug-Nov 30 6 1 ORCAWALE 2001 July-Dee 20 7 0 CSCAPE 2005 Aug-Dec 42 2 0 Appendix 2 The effective degrees of freedom (EDF) and P-values for each of the parameters included in the generalized additive model of sightings per unit effort in Southern California waters in 1979-2009 for each studied spe- cies of small cetacean: short-beaked common dolphin (Delphinus delphis ), long-beaked common dolphin ( D . capensis), Risso’s dolphin (Grampus griseus ), striped dolphin ( Stenella coeruleoalba ), bottlenose dolphin (Tursiops truncatus), northern right whale dolphin ( Lissodelphis borealis), Pacific white-sided dolphin (Lagenorhynchus obliquidens ), and Dali’s porpoise (Phocoenoides dalli). Because the short- or long-beaked Lo Pd Sc Tt Dd Dc (km) 0 2 0 3 NA NA 2045 8 16 0 3 NA NA 1842 18 4 0 3 NA NA 562 10 3 0 3 NA NA 1179 0 0 6 3 45 2 4210 0 0 4 1 21 0 2610 9 0 6 4 24 2 3936 2 0 1 7 16 1 2540 0 6 5 4 29 6 2951 common dolphins were not recognized formally as dis- tinct until 1994, data for both species were used in a combined category in analyses. Note that no EDF was available for the 2 parametric variables (Quarter and SurveyType). SurveyType is the variable for the cruises, which were conducted by the NOAA South- west Fisheries Science Center and the the California Cooperative Oceanic Fisheries Investigations. Vari- able abbreviations: DepthMin=minimum depth (m), DepthMean=mean depth (m), MaxDepth=maximum depth (m), SeasAv=seasonal averaged sea-surface temperature (°C), ENSO=El Niho-Southern Oscilla- tion, and PDO=Pacific Decadal Oscillation. NA=not available. Species Parameter EDF P-value Species Parameter EDF P-value Both common Quarter NA <0.01 Striped dolphin SurveyType NA <0.01 dolphins SurveyType NA <0.01 DepthMin 1.73 <0.01 Slope 6.97 <0.01 Bottlenose dolphin SurveyType NA <0.01 DepthMax 2.09 <0.01 DepthMean 4.97 <0.01 SeasAv 5.87 <0.01 PDO 1.4 0.03 PDO 1.97 <0.01 Northern right SurveyType NA <0.01 Short-beaked Quarter NA <0.01 whale dolphin DepthMax 2.51 <0.01 common dolphin Slope 3.12 <0.01 ENSO 3.32 <0.01 DepthMean 5.32 <0.01 SeasAv 2.19 <0.01 SeasAv 7.1 <0.01 Pacific white-sided Quarter NA <0.01 Long-beaked Quarter NA 0.02 dolphin SurveyType NA <0.01 common dolphin DepthMax 1.98 <0.01 Slope 1.57 <0.01 SeasAv 2.46 0.03 DepthMean 4.06 <0.01 ENSO 3.82 0.03 PDO 1.84 0.01 Risso’s dolphin Quarter NA <0.01 Dali’s porpoise Quarter NA <0.01 SurveyType NA <0.01 SurveyType NA <0.01 Slope 5.33 <0.01 Slope 1.61 <0.01 DepthMean 5.09 <0.01 DepthMean 3.81 <0.01 ENSO 2.43 <0.01 PDO 1.74 0.01 SeasAv 3.39 0.01 SeasAv 2.83 <0.01 178 Age, growth, and reproduction of Southern Kingfish (Menticirrhus americanus ): a multivariate comparison with life history patterns in other sciaenids Samuel D. Clardy1 Nancy J. Brown-Peterson2 Mark S. Peterson2 Robert T. Leaf2 Email address for contact author: samuel.clardy@usm.edu 1 Marine Education Center Gulf Coast Research Laboratory University of Southern Mississippi 703 East Beach Drive Ocean Springs, Mississippi 39564 2 Department of Coastal Sciences College of Science and Technology University of Southern Mississippi 703 East Beach Drive Ocean Springs, Mississippi 39564 Abstract— Southern Kingfish (Men- ticirrhus americanus) is an abundant sciaenid in the northcentral Gulf of Mexico, but little is known of its life history. Our objectives were to describe demographic traits and compare the characteristics of this population with those of other recreationally and com- mercially important populations in the U.S. Exclusive Economic Zone (U.S. EEZ). We report significant differences in sex-specific weight at length. Otolith annulus formation occurred in April and May and maximum age was dr- years for both sexes. Length-at-age analysis indicated that mean asymp- totic total length (TL; TL^ maie=244 mm, femaie=303 mm) and mean in- stantaneous growth rates (&maie=1.12 y_1 and &female=0-95 y-1) were signifi- cantly different between sexes. The mean length at 50% maturity (TL50) for females was 171 mm TL, corre- sponding to an age at maturity of 1 year. Gonadosomatic indices and his- tological examination of ovarian ma- turity phases indicated rapid gonadal development in February and March with females actively spawning from April to September. The interval be- tween spawning averaged 6.9 days, and the most frequent spawning occurred in June and July. Mean relative batch fecundity was 231.1 number of eggs g-1 of ovary-free body weight ^stan- dard error 35.7). Principal component analysis (PCA) of 5 variables from 17 sciaenid populations in the U.S. EEZ identified 2 principal components that explained 68.1% of variation among populations; these components rep- resent a size-related gradient and a gradient of spawning season dynamics. Five distinct groups were identified on the basis of fish size, age at maturity, spawning-season duration, and batch fecundity. Manuscript submitted 12 April 2013. Manuscript accepted 25 April 2014. Fish. Bull. 112:178-197 (2014). doi:10.7755/FB.112.2-3.6 The views and opinions expressed or implied in this article are those of the author (or authors) and do not necessarily reflect the position of the National Marine Fisheries Service, NOAA. The family Sciaenidae include many commercially and recreationally im- portant species and occur worldwide in temperate and subtropical marine, estuarine, and fresh waters (Chao, 1995, 2002). There are 79 sciaenids that occur in North America (Page et ah, 2013); 36 of these species com- monly occur in waters of the U.S. Exclusive Economic Zone (U.S. EEZ), and 29 of those 36 species support either commercial or recreational fisheries or both (FishBase, http:// www.fishbase.org). However, only 11 of these fishes are reported individu- ally in the NOAA National Marine Fisheries Service commercial (http:// www.st.nmfs.noaa.gov/commercial- fisheries/commercial-landings/annu- al-landings/index, accessed Septem- ber 2013) and recreational (http:// www.st.nmfs.noaa.gov/stl/recre- ational/queries/catch, accessed Sep- tember 2013) statistics databases. Despite the conspicuousness of this family, complete life-history infor- mation is not available for many of the commonly occurring species, in- cluding some economically important members. Understanding life-history strate- gies and the interaction among spe- cies traits, environmental factors, and population dynamics is not only important for fisheries management but also vital to predict population responses to a variety of natural and anthropogenic disturbances (Win- emiller and Rose, 1992; Winemiller, 2005). A comparison of life-history traits of species within the single, di- verse family Sciaenidae can further the understanding of the dynamics of this valuable group of fishes. For example, analysis of the reproductive life-history traits of Sciaenidae in the Gulf of Mexico (GOM) and Caribbean Sea resulted in the recognition of 3 major groups within this family that were defined by maximum length, fecundity, and duration of spawning season (Waggy et ah, 2006). Simil- iarly, Militelli et al. (2013) recently examined reproductive traits of 7 sciaenids from the Buenos Aires, Ar- gentina, coastal zone and found that composite groups could be described on the basis of the length of the re- productive period and size of the spawning area. Clardy et al.: Life history of Menticirrhus americanus and other sciaenids 179 Life-history patterns of many of the Sciaenidae in the northcentral GOM are not well understood, and limited life-history information is available on the Menticirrhus complex, which includes Gulf Kingfish, (Menticirrhus littoralis), Northern Kingfish (M. saxa- tilis), and Southern Kingfish (M. americanus). These species are targeted by both recreational and commer- cial fishermen in Mississippi, and all of them current- ly are unregulated in the northcentral GOM. Results from analysis of NOAA harvest data for 2000-11 indi- cate a decline in annual landings of Southern Kingfish from 117,967 to 36,332 kg in Mississippi waters dur- ing this time period (http://www.st.nmfs.noaa.gov/stl/ recreational/queries/catch, accessed September 2013). This decline likely was affected by a number of factors, and it may be cause for concern for the sustainability of the population. Southern Kingfish is a widely distributed estuarine sciaenid with a range in coastal waters from southern New England to the southern tip of Florida, the GOM and Caribbean Sea, and as far south as Argentina (Armstrong and Muller1; Chao, 2002; Haluch et ah, 2011; Militelli et ah, 2013); it is the most commonly oc- curring Menticirrhus species in northcentral GOM wa- ters. Although most frequently encountered over sandy bottoms (Chao, 2002), this species has been found in deep channels between barrier islands generally over sand, in shallow muddy bottoms, in seagrasses, or on shell hash (Reid, 1954; Bearden, 1963; Crowe, 1984). Seasonal movements of Southern Kingfish appear to occur from shallow waters, where it occurs in early spring through late fall, to deeper waters during the winter (Hildebrand and Cable, 1934; Bearden, 1963; Lagarde2; Fritzsche and Crowe3; Crowe, 1984). Updated information on age, growth, and reproduc- tion of Southern Kingfish in the middle South Atlantic Bight (SAB) was published last year (McDowell and Ro- billard, 2013) and includes the first age estimates from otoliths and fecundity estimates. However, similar cur- rent information, including histological assessment, for Southern Kingfish in the GOM had not been completed until we concluded the study presented here. McDow- ell and Robillard (2013) reported that the spawning 1 Armstrong, M. P., and R. G. Muller. 1996. A summary of biological information for southern kingfish (Menticirrhus americanus). Gulf kingfish (M. littoralis), and northern king- fish ( M . saxatilis) in Florida waters. Florida Marine Research Institute (FMRI) In-house Report IHR 1996-004, 25 p. [Available from http://research.myfwc.com/publications/pub- lication_info.asp?id=43619.] 2 Lagarde, C. C. 1981. Environmental requirements of se- lected coastal finfish and shellfish of the Mississippi Sound and vicinity: Southern kingfish Menticirrhus americanus (Linnaeus), 5 p. [Available from Mississippi State Univ. Research Center, NSTL Station, Mississippi, 105 Mill St.; Starkville, MS 39759.] 3 Fritzsche, R. A., and B. J. Crowe. 1981. Contributions of the life history of the southern kingfish, Menticirrhus ameri- canus (Linnaeus), in Mississippi, 84 p. BMR Project No. CO-ST-79-022. [Available from Mississippi Department of Marine Resources, 1141 Bayview Ave.; Biloxi, MS 39532.] season for this species in the SAB was March-August and that peak activity occurred in April. That find- ing is similar to previous results from studies in the GOM; for those studies gonads were examined macro- scopically and indicated that the spawning season in the GOM lasts from February or March to November (Irwin, 1970; Crowe, 1984; Harding and Chittenden, 1987), with a peak in April (Fritzsche and Crowe3). Size at sexual maturity appears similar between the GOM and the SAB; Harding and Chittenden (1987) re- ported on the basis of macroscopic assessment that fe- males in the northwestern GOM reach 100% maturity at an age of 12-14 months or a size of 250 mm total length (TL), and McDowell and Robillard (2013) sug- gested that females from the SAB reach 50% maturity at 1.1 years or 199 mm TL. Knowledge of the reproductive biology and somatic traits of fish populations is required to assess the re- siliency of populations to fishing (Nielsen and Johnson, 1983; Fulford and Hendon, 2010). To address gaps in knowledge, we describe a variety of reproductive and somatic traits, namely annual spawning season, spawn- ing frequency, batch fecundity, and size at 50% maturi- ty, for Southern Kingfish within the northcentral GOM that were determined by histological analysis and with standard techniques. We also quantified length at age and weight at length from analyses of nonlinear rela- tionships. Finally, we compare these traits with traits of other recreationally and commercially important sci- aenid populations in the U.S. EEZ to assess differences and similarities in life-history patterns of members of this family. Materials and methods Southern Kingfish were sampled from several loca- tions in the Mississippi Sound off the coast of Missis- sippi from April 2008 to May 2009 (Fig. 1). We targeted with hook and line a minimum of 50 fish each month from February to October and 15 fish from November to January. Haphazardly collected samples from crab pots and otter trawls during spring and summer were also obtained. Fish were identified according to Chao (2002). Upon collection, fish were immediately placed on ice and processed in the laboratory within 24 h. Each fish was measured for TL and standard length (SL) in millimeters and weighed in grams. The left sagittal otoliths of these fish were removed, cleaned, and dried; sex was determined macroscopically when gonads were removed and weighed (0.1 g). Sex-specific gonadosomatic indices (GSIs) were calculated: GSI = ( GW/GFBW ) x 100, (1) where GW = gonad weight; and GFBW = gonad-free body weight of the fish (Gree- ley et ah, 1986). A small cross section from the middle of the right gonad was removed and fixed in 10% neutral buffered 180 Fishery Bulletin 112(2-3) Figure 1 Map of locations where Southern Kingfish ( Menticirrhus americanus ) were sampled with hook and line and crab pots between April 2008 and May 2009 in the Mississippi Sound, Mississippi, to describe reproductive and somatic traits of this species and compare its life history with that of other sciaenids. formalin for a minimum of 1 week for histological analysis. Aging and aging validation of otoliths Age estimates were determined from Southern King- fish allocated in 10-mm-TL size classes. We randomly chose and aged up to 5 specimens per size bin; otoliths were processed as described in VanderKooy (2009). Otoliths were embedded in a resin block (Buehler Epoxicure4 resin and hardener, Buehler, an ITW Co., Lake Bluff, IL) and sectioned at the junction of the ostium and sulcus with a saw equipped with a Nor- ton diamond wheel (Saint-Gobain, Valley Forge, PA). These sections were sanded, mounted on labeled slides with clear Crystalbond 509 adhesive (Aremco Products, Inc., Valley Cottage, NY), and the slide was cooled and dried. The otolith sections were polished with a clear Flox-Texx mounting medium (Thermo Fisher Scientific, Inc., Waltham, MA). Annuli were counted with a Motic BA200 microscope (Motic North America, Richmond, Canada) under trans- mitted light by 2 independent readers. Fully formed annuli were counted to determine the age of the fish specimen, and the outer edge margin was coded on the basis of the percentage of translucent area beyond the final opaque ring (margin codes: 1=0%, 2=33%, 3=66%, and 4=99%; [VanderKooy, 2009]). A margin code of 1 was assumed to signify the month the annulus was formed. Numbers of annuli and margin codes were compared between readers, and any discrepancies were 4 Mention of trade names or commercial companies is for iden- tification purposes only and does not imply endorsement by the National Marine Fisheries Service, NOAA. reexamined by a third reader; if agreement could not be reached, the otolith was removed from analysis. Mc- Dowell and Robillard (2013) validated that Southern Kingfish off Georgia produced a single annulus once a year. Reproductive biology Preserved ovaries were dehydrated, embedded in paraf- fin, sectioned at 4 pm, and stained with Hematoxylin 2 and Eosin Y (Thermo Fisher Scientific) for histological examination. Slides were assessed microscopically to determine ovarian phases defined by Brown-Peterson et al. (2011); the 5 reproductive phases of immature, developing, spawning capable, regressing, and regen- erating, as well as the associated subphases of early developing and actively spawning, were recognized. A Southern Kingfish was defined as sexually mature once it entered the developing phase (DE) and cortical al- veoli (CA) oocytes were observed. Two quantitative approaches were used to estimate the spawning frequency of Southern Kingfish: 1) the percentage of fish in the spawning-capable phase with or without a 24-h postovulatory follicle (POF) complex (hereafter termed the POF method) and 2) the percent- age of fish undergoing oocyte maturation (OM) (Hunter and Macewicz, 1985) (hereafter termed the OM meth- od). Spawning frequency is expressed as the number of days between spawning events (see Brown-Peterson and Warren [2001] for details). Batch fecundity (BF) was estimated for female Southern Kingfish in the actively spawning subphase. A subsample from the middle of the gonad was re- moved, weighed (to the nearest 0.1 g), and preserved Clardy et al. : Life history of Menticirrhus americanus and other sciaemds 181 in modified Gilson’s fluid (Bagenal, 1966) for a mini- mum of 3 months. Once oocytes were separated from ovarian tissue, they were suspended in water and all hydrated or OM oocytes occurring in six 1-mL aliquots were counted. Graphs of oocyte size-frequency distribu- tions of spawning-capable and actively spawning fish were developed to identify and determine the size of OM oocytes used for counts. Batch fecundity (number of eggs) and relative batch fecundity (RBF; number of eggs per gram of ovary-free body weight leggs g_1 OFBW]) were calculated by the volumetric method (Ba- genal and Braum, 1971). All values for BF and RBF provided here are reported as a mean ± standard error of the mean ( SE). Comparison of sciaenid life histories Life-history data were summarized for recreationally and commercially important sciaenid species com- monly occurring in the U.S. EEZ. Species that occur infrequently in U.S. catches, have their centers of dis- tribution outside U.S. waters, or that are important only in the aquarium trade, were excluded from the comparison. In addition, commercially or recreationally important species for which there is little life-history information were excluded from the data summary. To treat geographically separated populations of the same species individually, we recognized GOM, Pacific, and Atlantic populations within species. Data for each of these species were obtained from the literature, with the exception of Southern Kingfish, which is described in this study. We summarized 8 somatic and reproduc- tive traits for each species: 1) maximum age (years), 2) maximum TL (millimeters), 3) spawning-season dura- tion (months), 4) spawning frequency (days), 5) RBF (eggs g_1 OFBW), 6) age at maturity (years), 7) TL at maturity (millimeters), and 8) the parameter b (mor- phology index: exponent of weight at length from the power function) in the 2-parameter power function used to describe weight at length (W=aTLb). Data analyses Somatic patterns of Southern Kingfish were quantified with linear and nonlinear relationships. The mean TL at 50% maturity (TL50) was estimated with a 2-param- eter logistic model: M'rL = l + e-r(,TL-TL50) ’ (2) where female maturity was coded binomially as imma- ture (0) or mature (1). The relationship between sex-specific SL and TL was estimated with a linear model: SL = /30+PlxTL. (3) We used a 2-parameter von Bertalanffy growth func- tion (VBGF) to describe length at age: TLt = TLJl-ekt), (4) where TLt = TL of a fish with age t: The VBGF model parameters were TL„ (the mean hypothetical maximum TL achieved by an individual, in millimeters) and the growth coefficient k (the rate of growth, per year). We estimated longevity as the age (in years) taken to reach 95% and 99% of predicted TL„ (Fabens, 1965; Ricker, 1979) and used a power function to determine mean weight at length. The 95% confidence intervals (Cl) of the mean model parameters for the VBGF, power, and logistic functions were de- termined by using the confint algorithm distributed in the stats package in the base version of the program R (vers. 2.15.1; R Core Team, 2012). The GSI values were arcsine square-root trans- formed before analysis and were tested for normality (Kolmogorov-Smirnov one-sample test) and homogene- ity of variance (Levene’s test). A one-way analysis of variance (ANOVA) tested for GSI differences among months by sex. If a significant F-value was observed, monthly values were separated with a Sidak pairwise test; if the data were heterogeneous, a Games-Howell (GH) test was used (Field, 2005). Differences in spawn- ing frequency among seasons (early, mid, and late) were tested with a chi-square test. Linear regressions of logiQ-transformed data were used to determine if there were relationships between either BF or RBF as the dependent variable and TL, OFBW, and age as the independent variables. The SPSS program (vers. 20.0; IBM Corp., Armonk, NY) was used, and the significance level was P<0.05. Principal components analysis (PCA) was used to compare and describe the multivariate somatic and reproductive traits of 17 geographic populations (14 species) in the family Sciaenidae. We limited our PCA to those populations that had available data for 5 so- matic and reproductive traits. Somatic traits included the maximum TL and the parameter b\ reproductive traits included duration of spawning season, RBF, and age at maturity. In cases where more than one value was reported for a trait, the median value was used for the PCA. The PCA was performed by eigenvalue decomposition of the correlation matrix with varimax rotation to maximally resolve loadings, and only eigen- values >1.0 were used. We considered any variable that loaded on a component at | >0.60 | to make a sig- nificant contribution to interpretation of that compo- nent (Hair et ah, 1984). The SPSS program was used, and the significance level was P<0.05. Results Somatic characteristics From April 2008 to May 2009, 519 Southern Kingfish (434 females, 85 males) were sampled; no fish were col- lected in January 2009. Of these fish, 508 were collect- 182 Fishery Bulletin 112(2-3) Table 1 Sample size (n), mean total length (TL, in millimeters) and minimum and maximum TL values by month and sex for Southern Kingfish (Menticirrhus americanus ) collected in Mississippi Sound, Mississippi, from April 2008 to May 2009. Female Male Date n Mean TL Min TL Max TL n Mean TL Min TL Max 1 April 2008 45 232 192 267 3 234 227 244 May 2008 3 270 231 303 - - - - June 2008 45 217 164 312 29 193 173 224 July 2008 42 243 163 348 10 226 180 256 August 2008 33 254 219 331 7 231 212 267 September 2008 51 252 191 328 3 233 226 242 October 2008 49 238 182 341 6 222 215 240 November 2008 16 258 205 307 2 212 203 220 December 2008 3 266 246 278 2 233 221 244 February 2009 45 239 181 283 10 217 198 263 March 2009 49 229 135 279 6 210 171 244 April 2009 11 246 148 280 2 199 198 200 May 2009 42 232 184 300 5 207 191 224 Overall 434 - 135 348 85 - 171 267 ed with hook and line, 7 were collected in otter trawls, and 4 were collected from crab pots. All haphazardly sampled specimens were within the normal size dis- tribution (trawls, 187-300 mm TL; crab pots, 263-280 mm TL). Sizes of females were recorded in ranges of 135-348 mm TL and 24.8-530.2 g; males fell in size ranges of 171-267 mm TL and 49.4-213.4 g (Table 1; Fig. 2, A and B). There was a significant linear rela- tionship between TL and SL for both sexes (Table 2; males: P<0.0001, coefficient of multiple determination [i?2]=0.98; females: P<0.0001, R2= 0.96). Comparison of the mean values and confidence inter- vals derived in the VBGF analysis indicated that males reach a significantly smaller asymptotic TL (a=0.05) and grow faster than females (Fig. 2, A and B; Table 2). Females at all ages had a greater mean length than that of males (Fig. 20. The estimated longevity, cal- culated with the mean VBGF parameters, was 3. 1-4.3 years for males and 3.6-5. 1 years for females (Table 2). Significant differences in both a and b parameters were observed by comparing mean and 95% Cl values in the weight-at-length relationship (parameter a was the coefficient of the power function). Females had a significantly greater mean value of the power function exponent (parameter b) than had males: 3.26 versus 2.88 (a=0.05; Table 2). Age determined from otoliths Alternating opaque and translucent annuli were dis- tinctive in the prepared cross sections of Southern Kingfish otoliths (Fig. 3A). Marginal increment analy- sis showed that Southern Kingfish in the GOM form opaque annuli during April and May, on the basis of the presence of fish with a margin code of 1 (0% translucence above the opaque annulus; Fig. 3B). A single fish had a margin code of 1 in June, although 2 was the dominant margin code in June and July (78% and 73% of fish, respectively). Fish with a mar- gin code of 2 showed active accretion of the otolith beyond the formation of the opaque annulus. Mar- gin codes 3 and 4 were most common from August to March; margin code 4 was dominant in February and March, just before the deposition of the opaque annu- lus (Fig. 3B). There were 5 age classes (age 0-4) observed in the sampled mature population (n=123) of Southern King- fish in the northcentral GOM. Fish of age 1 and 2 were the largest groups, accounting for 39% and 33% of the total, respectively. Because of the sampling technique used, young-of-the-year (age-0) fish composed only 12%, and fish of age 3 and 4 represented 13% and 3% of the sample, respectively. Reproductive characteristics Of the 519 Southern Kingfish collected, ovary samples from 397 females were processed for histological analy- sis. The smallest female observed to reach sexual ma- turity was 163 mm TL and age 1. Only 3.5% of the 397 females histologically examined were found to be immature. Estimated size at L50 was 171 mm TL (Fig. 4; i?2=0.71, az=396 ). All females >211 mm TL were sexu- ally mature and were age 1 or older. Clardy et al Life history of Menticirrhus americanus and other sciaenids 183 Female GSI increased from February to April, re- mained relatively high from May to August, and de- creased thereafter (Fig. 5A). Female GSI among months was significantly different (ANOVA: F\ o 423=32.49; PcO.OOOl) with values for March-September signifi- cantly higher than those for all other months (GH, Pc0.05), indicating Southern Kingfish may spawn in the northcentral GOM from March to September. Male GSI mirrored that of females (Fig. 5A) and was signifi- cantly different among months (ANOVA: F 10, 74=7-65; PcO.OOOl), peaking in April and reaching the lowest values in November (Sidak; P<0.05). Histological analysis confirmed the Southern King- fish spawning season that had been indicated by GSI value in our analysis. The majority of the females were in the early developing subphase in February; this percentage decreased in March as fish moved into the developing phase (Table 3). Females were first observed in the spawning capable phase (Fig. 5B) in March, although no POFs were seen in March. From April to August, >78% of captured females were in the spawning-capable phase; this percentage decreased to 59% by September (Table 3). Females in the actively spawning subphase were found from April to Septem- ber (Table 3). Although some females were found in the spawning-capable phase in October (13%), the ma- jority of fish were in the regressing and regenerating phases, signaling the end of the reproductive season. By November, all females were reproductively inactive (Table 3). Therefore, on the basis of histological obser- vations, the spawning season for Southern Kingfish in the northcentral GOM appears to be April-September. Furthermore, the asynchronous oocyte development ob- served in females in the spawning-capable phase (Figs. 5B and 6) indicates that Southern Kingfish is a batch- spawning species in the northcentral GOM. Seasonal spawning frequencies were estimated with both the POF and OM calculation methods, and they were calculated only for months when actively spawn- ing females were observed (Table 4). Results were similar between methods for the early and mid-season spawning frequencies but differed during the late sea- son. There was a significant difference seasonally for frequencies calculated with the POF method 0.05). Both methods revealed a spawning frequency of 11.3 days between spawns for the early season (April-May). During the mid-season (June-July), spawning frequency ranged from 3.5 to 5.8 days between spawnings depending on the method. Spawning frequency in the late season (Au- gust-September) was much greater for the OM method (5.4 days) compared with the POF method (19.7 days). The mean annual spawning frequency was calculated for both methods at 6.9 days between spawnings (Table 4). Therefore, over the course of the entire spawning season (April-September), Southern Kingfish spawn on average once every 7 days and an individual has the potential to spawn 26 times. Batch fecundity estimates were compared for 11 fe- males that were in the actively spawning subphase. A distinct group of oocytes >350 pm was observed in ac- tively spawning fish; therefore, oocytes >350 pm were considered to be undergoing OM (Fig. 6) and were counted for fecundity analysis. Overall BF ranged from 17,338 to 80,495 eggs and had a mean of 35,571 eggs (SE 6405) (Table 5). There were no significant re- lationships between log10 BF and logio TL (P2=0.199, Pi g=2.23, P=0.169) or logio BF and logio OFBW 184 Fishery Bulletin 1 12(2-3) Table 2 Summary of equations and related statistics used in this study for analyses of Southern Kingfish ( Menticirrhus americanus ) collected in Mississippi Sound, Mississippi, between April 2008 and May 2009. W=somatic wet weight (in grams), TL=total length (in millimeters), SL=standard length (in millimeters), coefficient of multiple determination (R2), coefficient of cor- relation (r), CI=confidence intervals, y=year, (3i=slope of simple linear regression, Po=y- intercept of simple linear regression, k-\s a growth rate constant (y-1), TLM= is the mean maximum TL (mm), TLt=TL at time t (mm), t=time, r=instantaneous rate of increase (mm-1), a= coefficient of the power function, b= exponent of the power function, TL5 0= total length at 50% maturity, pgs% and P99%= the age (y) taken to reach 95% and 99% of predicted TL Estimated mean para- Model name Para- meter value (minimum Analysis (model formulation) Sex n R2 P meter to maximum 95% Cl) SL vs. TL Linear Female 434 0.96 <0.001 Po -6.26 (-10.43 to -2.08) SL = p0 + Pi x TL Pi 0.85 (0.83 to 0.87) Male 85 0.98 <0.001 Po -8.26 (-13.45 to -3.08) Pi 0.86 (0.84 to 0.89) Length-at-age 2-parameter von Bertalanffy Female 89 0.57 — k 0.95 y-1 (0.76 to 1.16 y-1) growth function TLt = TLJ1 - e~kt) TL„ 303.2 mm (284.7 to 328.6 mm) Male 43 0.75 - k 1.12 y-1 (0.99 to 1.26 y-1) TL «, 243.7 mm (235.6 to 252.7 mm) Weight at Power function Female 80 0.98 _ a 2.73 x 10-6 (1.42 x lO-6 to 4.78 x 10-6) length W = aTLh b 3.26 (3.16 to 3.37) Male 43 0.96 - a 2.16 x 10-5 (7.14 x 10-6 to 5.48 x 10-5) b 2.88 (2.71 to 3.09) Longevity 95% and 99% of predicted L„ Female — — — p95% 3.6 y P99% 5.1 y Male - - - P95% 3.1 y p99% 4.3 y Maturity 2-parameter logistic function Female 396 0.37 — r 0.71 y-1 (0.06 to 0.09 y-1) MTL = (1 + e-'-(TL-rL5o))-1 tl50 170.6 mm (166.8 to 176.6 mm) Table 3 Percentages of spawning female Southern Kingfish (Menticirrhus americanus ) by month and spawning phase on the basis of histological analysis. Samples were collected in Mississippi Sound, Mississippi, between April 2008 and May 2009. Actively spawning is a subphase of the spawning-capable phase. Spawning Month Feb. phases n 44 Mar. 49 Apr. 44 May 45 Jun. 40 July 41 Aug. 33 Sep. 49 Oct. 37 Nov. 13 Dec. 2 Immature 2 6 2 13 2 5 0 0 0 0 0 Developing 2 31 0 0 2 0 0 0 3 0 0 Early developing 64 24 5 2 8 0 0 0 0 0 0 Spawning-capable 0 39 86 76 63 80 76 47 13 0 0 Actively spawning 0 0 7 9 15 15 15 12 0 0 0 Regressing 0 0 0 0 10 0 0 8 35 8 0 Regenerating 32 0 0 0 0 0 9 33 49 92 100 Clardy et al.: Life history of Menticirrhus americanus and other sciaenids 185 A B CD "O O O C TO ro ■ Margin code 4 □ Margin code 3 □ Margin code 2 B Margin code 1 Month Figure 3 Image of a sagittal otolith and a plot of marginal increments (shown by code) from Southern Kingfish ( Men- ticirrhus americanus) collected in Mississippi Sound, Mississippi, between April 2008 and May 2009. (A) Transverse section of a sagittal otolith on a Southern Kingfish that was 4+ years old, with the numbered ar- rows indicating deposited annuli and MC 3 indicating the margin code. (B) Percentage of fish in each marginal increment group, pooled by gender and age, from February to November. There were no data for December and January. Margin codes are based on the percentage of translucent area beyond the final opaque ring of the otolith (margin code 1=0%, 2=33%, 3=66%, and 4=99%). The numbers above the bars represent, for each month, the number of otoliths analyzed for marginal increments. 186 Fishery Bulletin 112(2-3) 100 150 200 250 300 350 TL (mm) Figure 4 Plot of percentage of female Southern Kingfish (Menticirrhus americanus) collected from the Mississippi Sound, Mississippi, be- tween April 2008 and May 2009 that reached maturity (entered the reproductive cycle) in relation to total length (TL) by 10-mm- TL size bins. The vertical line at 171 mm TL indicates the TL where 50% of individuals were mature (TL50). (i?2=0.164, Fi g=1.77, P=0.216). The 11 specimens used in the fecundity calculations were collected in May, July, August, and September, and the highest mean BF occurred in August at 38,722 eggs (SE 20,288) and the lowest mean BF, in September at 33,730 eggs (SE 12,544). There were only 2 age classes represented in the fecundity analysis: age-1 (n- 4) and age-2 (n= 7) fish (Table 5). Mean BF was slightly higher in age-2 fish at 36,622 eggs (SE 6122) than in age-1 fish at 33,730 eggs (SE 15,591); however, there was no significant re- lationship between logqo BF and logio age (P2=0.015, Fh 9=0.136, P=0. 721). Relative batch fecundity had a range of 94.4-509.5 eggs g_1 OFBW and a mean of 213.1 eggs g_1 OFBW (SE 35.7) (Table 5). Like the BF results, results for RBF had its highest mean in August at 259.6 eggs g_1 OFBW (SE 124.9) and lowest mean in September at 168.1 eggs g'1 OFBW (SE 70.3). Unlike the BF results, results for mean RBF were lower in age-2 fish, with 200.2 eggs g_1 OFBW (SE 29.4), than in age-1 fish, with 235.7 eggs g_1 OFBW (SE 91.5). Potential annual fe- cundity for Southern Kingfish in the northcentral GOM indicates that a female with a somatic wet weight of 176 g could potentially spawn 924,846 eggs over the course of the spawning season. Sciaenidae life-history traits: a comparative analysis Wide variation was observed in the somatic and repro- ductive traits examined for the 21 sciaenid populations that were analyzed in this study (Appendix table). For instance, Silver Seatrout ( Cynoscion nothus), a GOM species, has the lowest reported maximum age (1.5 years) of the 21 species, but Black Drum ( Pogonias cromis), another GOM species, has the oldest reported maximum age (43 years). Duration of spawning season ranged from 2 months (Black Drum) to 7-10 months for White Croaker ( Genyonemus lineatus), a Pacific species. Size at maturity varied from 110-120 mm TL in Silver Perch ( Bairdiella chrysoura ), a GOM species, to 900 mm TL for Atlantic populations of the Red Drum ( Sci - aenops ocellatus). Life-history traits least frequently reported for sciaenid species in- clude RBF and spawning frequency (Appen- dix table). Of the 11 species or populations with available spawning-frequency data, Queenfish ( Seriphus politus ) has the longest interspawning interval of 7.4 days and Sil- ver Perch has the shortest interval of 1.3-1. 6 days. The PCA of the 5 somatic and reproductive traits produced 2 meaningful components that accounted for 68.1% of the total varia- tion of the original data set. The first prin- cipal component (PC 1) accounted for 43.1%, whereas the second principal component (PC 2) accounted for 25.0%. Maximum TL and age at maturity both positively loaded on PC 1, but RBF loaded negatively on PC 1 (Table 6). In contrast, dura- tion of spawning season and parameter b loaded posi- tively on PC 2 (Table 6). We interpret PC 1 as a size- related component of life history, and PC 2 represents spawning season dynamics. Overall, the PCA indicated 3 general trends in sciaenid life history: 1) fishes that reach a larger maximum TL and older age at maturity tend to have lower RBF, 2) a higher RBF is associated with a shorter spawning-season duration, and 3) a lon- ger spawning-season duration is associated with fishes that have higher b values. Five groups of sciaenid populations can be discerned in the PCA biplot (Fig. 7). Group A, consisting of fishes with high maximum TL, greater age at maturity, and low RBF, is composed of GOM species Red Drum and Black Drum. Group B consists of fishes with longer spawning seasons, greater b values, low RBF, smaller maximum TL, and lower age at maturity and is repre- sented by Sand Seatrout (Cynoscion arenarius ), South- ern Kingfish, White Croaker, and Queenfish, a mixture of species from all 3 regions, the GOM as well as the Atlantic and Pacific. Group C, made up of fishes with higher RBF, shorter spawning season, young age at maturity, small maximum TL, and low b values, is rep- resented by 2 GOM species: Spot (Leiostomus xanthur- us) and Silver Perch. Group D consists of fishes with more intermediate traits, such as small maximum TL, low age at maturity, and moderate spawning season, b values, and RBF. A species each from the Atlantic and GOM represent group D: Atlantic Croaker ( Micropogo - nias undulatus) and Silver Seatrout. Finally, group E also is an intermediate group with characteristics of Clardy et al.: Life history of Menticirrhus americanus and other sciaenids 187 6 Figure 5 Gonadal development of Southern Kingfish (Menticirrhus americanus) sampled from the Mississippi Sound, Mississippi, from April 2008 to May 2009. (A) Plot of mean values in a gonadosomatic index for male and female Southern Kingfish (for sample sizes of each sex, see Table 1). Error bars indicate ±1 standard error; in several months the error bars for males, are hidden by the data point symbol. Data points are significantly different (P<0.05) if letter labels are not the same. (B) Photomicrograph of a female Southern Kingfish in the spawning-capable phase. Labels indicate the following oocyte stages: cortical alveoli (CA), primary vitellogenic (Vtgl), secondary vitellogenic (Vtg2), tertiary vitellogenic (Vtg3), and postovulatory follicle complex (POF). moderate maximum TL, age, RBF, duration of spawning season, and b values. Group E comprises fishes from all 3 regions; Spotted Seatrout (Cynoscion nebulosus) from Atlantic and GOM populations, as well as Weakfish ( Cynoscion regalis), Spotfin Croaker (Roncador stearn- sii), and Yellowfin Croaker ( Umbrina roncador). Discussion In this article, we identify and address gaps in the understanding of the life history of the recreationally and commercially important Southern Kingfish in the northcentral GOM. We report that Southern Kingfish 188 Fishery Bulletin 112(2-3) TabSe 4 Seasonal spawning frequency for Southern Kingfish (Menticirrhus ameri- canus) in the northcentral Gulf of Mexico determined with 2 histological methods. The postovulatory follicle (POF) method is based on the pres- ence of POF <24 h and the oocyte maturation (OM) method is based on the presence of oocytes in OM. Only females in the spawning capable phase (including the actively spawning subphase) were used in this anal- ysis. Samples were collected from Mississippi Sound, Mississippi, between April 2008 and May 2009. Season n POF spawning frequency (days) OM spawning frequency (days) Early (April-May) 79 11.29 11.29 Mid (June-July) 70 3.50 5.83 Late (Aug.-Sept.) 59 19.67 5.36 Total (April-Sept.) 208 6.93 6.93 has a 6-month spawning period that occurs from spring to summer and that an individual female spawns, on average, 213.1 eggs g_1 OFBW once per week. Females reach sexual maturity at a relatively small size (171 mm TL) and young age (1+ years). We report significant differences in sex-specific length at age and weight at length. Finally, we document how the life-history traits of this species are part of the multivariate continuum of sciaenid stocks in the coastal waters of the continen- tal United States. Our analysis of the reproductive traits of the South- ern Kingfish provided the timing of spawning through the use of both GSI and histological analysis, and we identified a single, extended spawning period. The spawning season for Southern Kingfish in the north- central GOM begins in late March and ends in Septem- ber, with peaks in male ( 1.56%) and female (4.75%) GSI values in April. Other investigators have previously re- ported similar spawning seasons for Southern Kingfish in both the Atlantic and GOM on the basis of macro- scopic inspection of the ovary and GSI values (Hildeb- rand and Cable, 1934; Bearden, 1963; Lagarde2; Smith and Wenner, 1985; McDowell and Robillard, 2013). An extended spawning season (>3 months) is characteris- tic of most sciaenid species (Appendix). Our analysis of this reproductive trait places Southern Kingfish in group B (see Fig. 7), a group partially characterized by having the longest spawning season. Our finding of the existence of a single spawning season is in contrast to results from Harding and Chit- tenden (1987), who found on the basis of male and fe- male GSI and macroscopic classification of maturity that the spawning period in the northwestern GOM occurred from February or March to November and comprised 2 primary, discrete spawning periods (spring and fall). However, the Harding and Chittenden (1987) collections were from the deeper part of the bathymet- ric range of Southern Kingfish; more thorough collec- tions in estuaries, the surf zone, or the shallow inshore waters could resolve whether there are 2 discrete spawn- ing periods, as they have suggested, or 1 spawning period with 2 periods of recruitment. Although GSI values are typically good indicators of spawning prepared- ness, histological analysis can re- fine and more precisely delineate the spawning season and has not been previously performed for Southern Kingfish from the GOM. The spawning season estimated through our histolog- ical analysis matched our GSI results; spawning is initiated in April and ends in September. Additionally, there were fish in the spawning-capable phase in both March and October, although no females were in the actively spawning subphase during these months. Our estimate of the Southern Kingfish spawning period is supported by Anderson et al. (2012), who studied daily growth rings in otoliths from juveniles in the north- central GOM and estimated birth dates with back cal- culation. Although there were fish in that study with 250 A 707 7S7 2o7 3Sj 30j 3Sj ?o7 0.60 | (shown in bold) are considered useful for naming components (Hair et al., 1984). Maximum total length is presented in millimeters, duration of spawn- ing season is shown in months, relative batch fecundity is reported in number of eggs per gram of ovary-free body weight and age at maturity is presented in years; 6=slope of length-weight power function. Variable PC 1 PC 2 Maximum total length 0.853 -0.210 Spawning-season duration -0.297 0.830 Relative batch fecundity -0.626 -0.396 Age at maturity 0.861 -0.252 b 0.023 0.708 the determination of the spawning potential of South- ern Kingfish at different lengths and ages; however, fecundity measurements are poorly understood for the Southern Kingfish across its range. Previously report- ed estimates for 20 females in the northcentral GOM (Fritzsche and Crowe3) were a mean BF of 105,359 eggs (range: 46,024-332,229 eggs) and a mean RBF of 527 eggs g_1 OFBW, but these 2 estimates are based on all oocytes >300 pm. Our data show that vitello- genic oocytes between 300 and 350 pm do not undergo OM and, therefore, represent several potential batch- es. Batch fecundity estimates in our study were lower than those found by Fritzsche and Crowe3 because only hydrated oocytes or those oocytes undergoing OM were used. Militelli et al. (2013) reported RBF estimates for 8 Southern Kingfish from Argentina as 217 eggs g_1 OFBW (SE 70) — a finding that is similar to our results of 231 eggs g_1 OFBW (SE 36) for fish in the north- central GOM. McDowell and Robillard (2013) did not directly report RBF data for Southern Kingfish from Georgia, but we calculated RBF to be 308.6 eggs g_1 total weight (SE 27.6) on the basis of information in their manuscript. The lack of a significant relationship in our data be- tween BF or RBF and TL, OFBW, or age is unusual, but it may be linked to our small sample size and the potential for significant variation in fecundity among individuals within a protracted spawning season (Low- erre-Barbieri et al., 2009). McDowell and Robillard (2013) had a more robust sample size of 36 fish and found a significant relationship between batch fecundi- ty and both length and weight. Therefore, our fecundity estimates should be viewed with caution, despite their similarity to fecundity estimates of the Argentinean stock. Overall, Southern Kingfish exhibit some of the lowest reported RBF values for sciaenids (Appendix), a trait shared with the other members of group B in the PCA (see Fig. 7): Sand Seatrout, White Croaker, and Queenfish. 190 Fishery Bulletin 112(2-3) 1.5 -r -1.5 -1 0 -0 5 0 0 0 5 1 0 1.5 Principal component one Figure 7 Biplot of first (43.3%) and second (25.0%) principal components from principal component analysis (PCA) on the basis of the relative measures — each denoted with an arrow — of reproductive traits (spawning-season duration [SPAWN_DUR], relative batch fecundity [REL_FEC], and age at maturity [AGE_MAT]) and somatic traits (maximum total length [MAX_TL] and the weight-at-length parame- ter [6]) for 17 populations of sciaenids in 3 regions: Gulf of Mexico (GOM), Pacific Ocean, and Atlantic Ocean. Numbered points indicate sciaenid populations: (1) GOM Silver Perch ( Bairdiella chrysoura ), (2) GOM Sand Seatrout (Cynoscion arenarius), (3) Atlantic Spotted Seatrout (C. nebulosus), (4) GOM Spotted Seatrout, (5) GOM Silver Seatrout (C. nothus), (6) Atlantic Weakfish (C. regalis), (7) Pacific White Croaker ( Genyonemus lineatus), (8) GOM Spot ( Leiostomus xanthurus), (9) Atlantic Southern Kingfish (. Menticirrhus americanus), 10. GOM Southern Kingfish, (11) Atlantic Atlantic Croaker (Mi- cropogonias undulatus), (12) GOM Atlantic Croaker, (13) GOM Black Drum ( Pogonias cromis ), (14) Pacific Spotfin Croaker (. Roncador stearnsii), (15) GOM Red Drum ( Sciaenops ocellatus), (16) Pacific Queenfish (, Seriphus politus), and (17) Pacific Yellowfin Croaker (Umbrina roncador). The letters A-E indicate the 5 groups of sciaenids identified through PCA. The large circle represents a maximum cor- relation of 1 between original variables in the PCA analysis and the factor loadings for each variable that were generated by the PCA. Spawning frequency for batch-spawning species like Southern Kingfish, when combined with batch fecun- dity, provides an estimate of the total annual repro- ductive output of a species. Histological confirmation of multiple spawnings per season for populations of Southern Kingfish in Brazil and the SAB have been reported recently on the basis of the appearance of POFs and surrounding oocytes in multiple stages of development (Haluch et al., 2011; McDowell and Rob- illard, 2013). In our study, the seasonal spawning fre- quency was found to be about 7 days between spawn- ings with a peak of 3-6 days (depending on method, POF or OM) during June and July in the northcen- tral GOM. This interspawning interval is one of the longest ones reported among the Sciaenidae (see Ap- pendix) and contrasts with that of 2-4 days between spawnings for the SAB population of Southern King- fish (McDowell and Robillard, 2013). However, spawn- ing-frequency data are based on a 4-month spawn- ing season in Georgia and a 6-month season in the GOM — a difference that may account for some of this variation. Clardy et ai Life history of Menticirrhus americanus and other sciaenids 191 Unfortunately, spawning-frequency data are avail- able for only 11 of the 24 economically important sci- aenid species or populations and, for that reason, were not included in the PCA of reproductive and somatic traits. Therefore, the importance of spawning frequency to the reproductive and somatic relationships between Southern Kingfish and other sciaenids is unknown. The OM method appeared to be a reliable indica- tor of spawning frequency of Southern Kingfish in the GOM. Similarly, McDowell and Robillard (2013) used the OM method (specifically, presence of hydrated oo- cytes) to determine spawning frequency of Southern Kingfish from the SAB, although they did not pres- ent seasonal differences in spawning frequency. In our study, the OM method indicated that no significant difference existed between seasons as was observed with the POF method. The interspawning interval at the end of the season was longer when calculated by the POF method than when estimated with the OM method — a result that was likely due to a sampling bias. Fish that had finished spawning for the year may have moved out of the sample areas, whereas females that were still preparing to spawn remained in those areas. Therefore, fish that contained oocytes in the OM stage may have been more vulnerable to capture than fish with POFs. Similar differences in spawning fre- quency between methods were noted in the late sea- son of Silver Perch (Grammer et ah, 2009), where the OM method revealed an estimated 1.6 days between spawnings and the POF method indicated an estimated 16 days between spawnings. Although Grammer et al. (2009) stated their results may have been a function of low sample size (n=16), our study had a larger sample size (n- 59) that should not have been a contributing factor to the large difference observed. Interestingly, on the basis of percentage of spawning fish captured (OM method), spawning frequency of Southern King- fish from Georgia also was highest at the beginning of the reproductive season (March and April; McDowell and Robillard, 2013). Female Southern Kingfish in the northcentral GOM reached sexual maturity as small as 163 mm TL, cor- responding to 1 year of age. Females reached 50% ma- turity by 171 mm TL and 100% maturity by 211 mm TL, both at age 1. These results are consistent with reports from Smith and Wenner (1985), who estimated that females from the SAB reached TL§q at 192 mm TL at age 1 and TL\qq at 230 mm TL, also at age 1. Re- cently, McDowell and Robillard (2013) have confirmed these estimates of TL50 (199 111m TL) and 50% maturi- ty (1.1 years), indicating little change in the population of Southern Kingfish from the SAB over the past 25 years. Similarly, Haluch et al. (2011) reported female TL50 at 167 mm TL from the area of Santa Catarina, Brazil, with TL\qq at 228 mm TL. In contrast, Militelli et al. (2013) found TL50 for females in the coastal zone of Buenos Aires, Argentina, to be 223 mm TL, although this result was based on a relatively small sample size (n=54). Harding and Chittenden (1987) noted TL100 at 250 mm TL (with few maturing, virgin fish past 220 mm TL) for fish in the northwestern GOM, providing further evidence that Southern Kingfish in the GOM reach sexual maturity by age 1. Many sciaenids have developed a strategy to mature in the first year of life (Waggy et ah, 2006; see Appendix); 10 of the 17 species or populations analyzed in the PCA were in the low age-at-maturity quadrats of the plot (groups B, C, and D in Fig. 7). Population-level characteristics, such as mortal- ity rates and longevity, have been shown to correlate with individual growth parameters (Beverton and Holt, 1959; Lorenzen, 2005). Therefore, the somatic growth characteristics of Southern Kingfish discussed here provide a proxy for the determination of these char- acteristics that can aid in the management of the spe- cies. Our results indicate geographic differences in the maximum length and age of Southern Kingfish, as well as sex-specific differences in growth and condition. The maximum length from our study (348 mm TL) is similar to the lengths reported by Bearden (1963) on the East Coast of the United States (338 mm TL) and by Harding and Chittenden (1987) for the northwest- ern GOM (345 mm TL), but this result is smaller than the maximum size of 404 mm TL (Smith and Wenner, 1985) and 419 mm TL (McDowell and Robillard, 2013) reported for fish off the Atlantic coast of the southeast- ern United States. In our study, males from the GOM were found to have a smaller mean size (211.2 mm TL) than that of females (238.5 mm TL). Sex-specific differ- ences in maximum length also were reported by Hard- ing and Chittenden (1987) and McDowell and Robillard (2013) for Southern Kingfish, and such differences are common for species in this family (Chao, 1995, 2002). On the basis of annuli counts in sagittal otoliths, maximum age for both males and females in the north- central GOM was 4+ years. The oldest reported South- ern Kingfish was an individual that reached age 6 from the Atlantic coast of the southeastern United States (Smith and Wenner, 1985), although this age determi- nation was made with scale annuli, which is less accu- rate than age determination from otoliths (VanderKooy, 2009). Recent aging of Southern Kingfish from Georgia with the use of otoliths revealed a maximum age of 5 years, with the majority of fish 2000.5 m) were determined for the 11 most com- monly encountered species. The fol- lowing are values of uncorrected density (individuals/1000 km2, coef- ficients of variation in parentheses) for the seasonal period and depth with greatest density for a selec- tion of the species in this study: blue whale ( Bcilaenoptera musculus ), summer-fall, shallow, 3.2 (0.26); fin whale (B. physalus), summer-fall, shallow, 3.7 (0.30); humpback whale ( Megaptera novaeangliae), summer- fall, shallow, 3.1 (0.36); short-beaked common dolphin ( Delphinus del- phis), summer-fall, shallow, 1319.7 (0.24); long-beaked common dolphin ( D . capensis), summer-fall, shallow, 687.9 (0.52); and Dali’s porpoise ( Phocoenoides dalli ), winter-spring, deep, 48.65 (0.28). Seasonally, den- sity varied significantly by depth for humpback whales, fin whales, and Pacific white-sided dolphins. Manuscript submitted 24 January 2013. Manuscript accepted 23 May 2014. Fish. Bull. 112:198-220 (2014). doi:10.7755/FB.112.2-3.7 The views and opinions expressed or implied in this article are those of the author (or authors) and do not necessarily reflect the position of the National Marine Fisheries Service, NOAA. Seasonal distribution and abundance of cetaceans off Southern California estimated from CalCOFI cruise data from 2004 to 2008 Annie B. Douglas1 (contact author) Andrea M. Havron15 John Calambokidis1 Dominique L. Camacho1'5 Lisa M. Munger2 Greg S. Campbell2 Melissa S. Soldevilla3 2 John A. Hildebrand2 Megan C. Ferguson4 Email address for contact author: abdouglas@cascadiaresearch.org 1 Cascadia Research Collective 21814 West Fourth Avenue Olympia, Washington 98501 2 Scripps Institution of Oceanography University of California, San Diego 8635 Discovery Way La Jolla, California 92093-0210 3 Southeast Fisheries Science Center National Marine Fisheries Service, NOAA 75 Virginia Beach Drive Miami, Florida 33149 4 National Marine Mammal Laboratory Alaska Fisheries Science Center National Marine Fisheries Service, NOAA 7600 Sand Point Way NE Seattle, Washington 98115-6349 5 Spatial Ecosystems P.O. Box 2774 Olympia, Washington 98507 At least 30 species of cetaceans are found in the California Current (Leatherwood et ah, 1982), includ- ing 5 species of large whales listed as endangered under the U.S. En- dangered Species Act. The abun- dance and diversity of species along the West Coast of the United States and the continental slope are closely linked to the high level of biological production that is caused by upwell- ing and mixing of 4 different water masses along the California coast on a seasonal and interannual basis (Reid et ah, 1958; Smith et ah, 1986; Munger et ah, 2009). Although these waters are important to marine fau- na, they are also increasingly impor- tant to humans who use them for commercial shipping and fishing; oil and gas exploration, development, and production; naval exercises; and recreation. The combined use of these highly productive waters by cetaceans and humans can lead to ships striking large whales (Jen- sen and Silber, 2003; Berman-Kow- alewski et ah, 2010), entanglements of cetaceans in fishing gear (Julian and Beeson, 1998; Laist et ah, 2001; Carretta et ah, 2011b), and disrup- tion of normal behaviors by under- water sound (McDonald et ah, 2006; Weilgart, 2007). To assess long-term impacts of fisheries, industry, and ecosystem variability on marine mammals, it is necessary to estimate abundance, understand stock struc- ture, and determine seasonal habitat use by the species that inhabit these waters. Abundance for the summer and fall seasons has been estimated for many cetacean species in waters off California, Oregon, and Washington through the use of ship-based line- transect surveys or mark-recapture techniques of photographically iden- tified whales (Calambokidis and Bar- low, 2004; Barlow and Forney, 2007; Carretta et ah, 2011b). However, weather conditions make ship-based line-transect surveys difficult to con- duct year-round, and few studies have quantified habitat and distribution shifts of marine mammals during the Douglas et al. Seasonal distribution and abundance of cetaceans off Southern California 199 124°W l 122“W _l 120°W l 1 18°W l 116°W i Santa Barbara Los Angeles Study area zones 167.087 km2; Deep (>=2000 5 m) 71,407 km2 Shallow (<2000 5 m) Survey effort ■ CalCOFI stations Bathymetry contour (500 m) N I 1 1 1 1 0 25 50 100 km San Diego l). S. A._ MEXICO CANADA 124°W 122°W 120°W 118°W 116°W Figure 1 Map of the study area and 6 southern transect lines of 16 quarterly California Cooperative Oceanic Fisheries Investigation (CalCOFI) surveys conducted from July 2004 to April 2008 off Southern Cali- fornia. Shallow survey area (<2000.5 m) is lighter gray and deep survey area (>2000.5 m) is dark gray. Dark squares on CalCOFI transect lines indicate oceanographic sampling stations. The CalCOFI study area (238,494 km2) occurs completely inside the larger southern stratum (318,500 km2) of the South- west Fisheries Science Center shipboard surveys that extends farther to the northwest. winter and spring months (Dohl et al.1; Forney and Bar- low, 1998; Becker, 2007). In this study, new estimates of cetacean abundance were calculated off Southern California with marine mammal sighting data collected during 16 cruises undertaken as part of the California Cooperative Oceanic Fisheries Investigation (CalCOFI) (Bograd et al., 2003; Ohman and Venrick, 2003; Soldev- illa et al., 2006). Marine mammal surveys on CalCOFI cruises are conducted quarterly, along predetermined transect lines between oceanographic water sampling stations (Fig. 1). The depth of the study area is ex- tremely variable, with shallow waters inshore of the Channel Islands, a steep slope west of these islands, and an expansive deepwater plain offshore. Almost 1 Dohl, T. P., K. S. Norris, R. C. Guess, J. D. Bryant, and M. W. Honig. 1980. Summary of marine mammal and seabird surveys of the Southern California Bight area, 1975-1978. Part II. Cetacea of the Southern California Bight. Final report to the Bureau of Land Management, NTIS Rep. No. PB81248189, 414 p. 30% of CalCOFI transect lines occur in water depths of 20-2000 m, and 69% of them occur in water depths of 3001-4600 m, and there is minimal (1%) coverage with transect lines in water depths of 2001-3000 m owing to the steep slope and orientation of the transect lines. Therefore, this study provides information on seasonal and interannual presence of cetaceans in both coastal and deep offshore waters. Such information is important to understanding and potentially mitigating effects of human activities on cetacean populations off Southern California. Materials and methods Data collection Dedicated visual observers conducted line-transect surveys for marine mammals (Buckland et al., 1993; Kinzey and Gerrodette, 2003) during 16 quarterly Cal- 200 Fishery Bulletin 112(2-3) Table 1 Individual cruise identification, vessel, schedule (beginning, ending, season), line-transect-survey effort in kilome- ters, and marine mammal visual observers of 16 California Cooperative Oceanic Fisheries Investigation (CalCOFI) surveys conducted from July 2004 to April 2008. Height of observer platform on the 3 research vessels from which these surveys were completed: 13.2 m on RV Roger Reuelle [RR], 8.1 m on RV New Horizon [NH], and 11.0 m on NOAA Ship David Starr Jordan [DSJ]. CalCOFI cruise ID Vessel Begin End Season Survey effort (km) Visual observers1 0407JD DSJ 13-Jul-2004 28-Jul-2004 Summer 1543 RWB, ABD 0411RR RR 2-Nov-2004 19-Nov-2004 Fall 1295 ABD, AM, MS, SEY 0501NH NH 4-Jan-2005 20-Jan-2005 Winter 1006 DLC, EV 0504NH NH 15-Apr-2005 30-Apr-2005 Spring 1485 DLC, SMC 0507NH NH l-Jul-2005 16-Jul-2005 Summer 1571 DLC, ABD, VI 0511NH NH 4-Nov-2005 20-Nov-2005 Fall 1104 DLC, SMC 0602JD DSJ 4-Feb-2006 25-Feb-2006 Winter 1144 GSC, SMC 0604NH NH l-Apr-2006 17-Apr-2006 Spring 1624 DLC, ABD 0607NH NH 8-Jul-2006 24-Jul-2006 Summer 1595 DLC, AMH 0610RR RR 21-0ct-2006 5-Nov-2006 Fall 1208 ABD, AMH 0701JD DSJ 12-Jan-2007 2-Feb-2007 Winter 1080 GSC, ABD 0704JD DSJ 28-Mar-2007 18-Apr-2007 Spring 911 GSC, SMC 0707NH NH 28-Jun-2007 13-Jul-2007 Summer 1502 AMH, SEY 0711NH NH 02-Nov-2007 18-Nov-2007 Fall 1109 DLC, LJM 0801JD DSJ 07-Jan-2008 23-Jan-2008 Winter 935 DLC, GSC 0803JD DSJ 24-Mar-2008 09-Apr-2008 Spring 884 DLC, GSC Observers: R. W. Baird, D. L. Camacho, G. S. Campbell, S. M. Claussen, A. B. Douglas, A. M. Havron, V. Iriarte, A. Miller, L. J. Morse, M. Smith, E. Vazquez, and S. E. Yin. COFI cruises from July 2004 to April 2008 (Table 1). Covering an area of 238,494 km2, the study area con- sisted of coastal, shelf, and pelagic oceanic habitat from nearshore waters to waters 700 km offshore and up to 4600 m deep. Observers used unaided eye or handheld 7x50 reticle Fujinon2 binoculars (Fujifilm Corp., Tokyo) to sight, identify, and estimate group sizes of cetaceans and pinnipeds encountered along the transect lines be- tween CalCOFI hydrographic sampling stations (Fig. 1). The Southern California hydrographic sampling station sites are set along 6 parallel lines running southwest to northeast, with lines increasing in length from north to south (470-700 km). Stations occur ev- ery 37 km in coastal and continental shelf waters and every 74 km in offshore locations (Fig. 1). Occasionally, transect lines were interrupted by naval activity or adverse weather conditions; in these cases, the observ- ers discontinued effort until their vessel adjusted to a course that intersected with the interrupted transect line. Transit lines that ran along the CalCOFI tran- sect lines, as well as to and from the study area, were surveyed opportunistically in addition to the primary transect lines; however, these data were excluded from the analyses and results described here. Five northern 2 Mention of trade names or commercial companies is for iden- tification purposes only and does not imply endorsement by the National Marine Fisheries Service, NOAA. lines were surveyed partially during 2 winter cruises; however, only a few sections of these lines were sur- veyed with acceptable sea conditions, and therefore these data also have been excluded from the analyses and results in this study. Three vessels were used for the line-transect sur- veys: the RV Roger Revelle (2 surveys) and RV New Horizon (8 surveys) of the Scripps Institution of Ocean- ography, University of California, San Diego, and the NOAA Ship David Starr Jordan (6 surveys) (Table 1). Survey speeds ranged from 18.5 to 22.2 km/h. Height of the observer platform varied by vessel from 8.1 to 13.2 m, raising the possibility that there would be a vessel or a vessel-season bias. To test these biases, we ran single-factor analyses of variance (ANOVAs) to de- termine whether visual observers made initial sight- ings at significantly different distances for each vessel or vessel-season combination. Additionally, we ran tests to determine whether the number of transect line ki- lometers surveyed in good weather varied by season. Scanning from directly abeam to 10° past the bow on either side of the vessel, 2 observers recorded marine mammal sightings. During 2 survey cruises, an addi- tional person was available to record data and provide relief for observers at meal times (Table 1). Recorded sighting data included date, time, vessel latitude and longitude, vessel true heading, distance of animal from the vessel, sighting angle, 0, from the transect line, de- Douglas et al. Seasonal distribution and abundance of cetaceans off Southern California 201 termined with an angle board (zero at bow, negative to port, positive to starboard), sighting number, spe- cies, group-size estimate (best, high, low), presence of calves, general behavior of animals, photographs (if taken), and comments pertaining to the sighting. Be- cause of the surfacing behavior of cetaceans and lack of visibility while submerged, animal counts were only estimates, with the recorded “low” estimate being the minimum number of individuals observed during the sighting, the “high” estimate being the maximum, and the “best” estimate was always recognized as the value closest to the actual number of individuals. Group- size estimates and species confirmation were generally made by the lead observer but were agreed upon by all observers present. Transect lines were surveyed in “passing mode,” which does not allow for any alteration of course for closer examination of groups encountered. Barlow3 stated that surveys conducted in passing mode yield less biased estimates of encounter rates but result in a higher number of unidentified groups and more biased estimates of group size and species percentages com- pared with surveys conducted in closing mode. Clos- ing mode allows for all observers to go “off effort” and to adjust the course and speed of the vessel in order to approach animals sighted at a distance from the transect line. To assist with species identification and group-size estimation, 25x binoculars were available for all surveys on the Roger Revelle , for most surveys on the David Starr Jordan , and occasionally on the New Horizon. The 25x binoculars were used only as an aid once a group had been located with the naked eye or 7x50 binoculars to maintain a consistent search method between surveys. Observers recorded effort (on or off), weather (sea state, swell height, visibility [or estimated distance that observers can detect a small cetacean], and pre- cipitation), changes in course and speed, and sight- ing information onto data sheets. Observations were considered on effort if observers were actively search- ing with the unaided eye and 7x50 binoculars in a sea state 0-5, the vessel was traveling no less than 6 km/h, and there was a minimum visibility of 0.9 km (0.5 nmi) in front of the vessel. All sightings of all spe- cies of marine mammals were recorded with the excep- tion of California sea lions ( Zalophus californianus), which were sighted most often near the coast because estimation of group size and documentation of sight- ing details of high numbers of California sea lions in coastal waters would have compromised the ability of observers to sight and record other species that occur in the same area. Because these surveys were conducted in pass- ing mode and with limited use of 25x binoculars, 3 Barlow, J. 1997. Preliminary estimates of cetacean abun- dance off California, Oregon, and Washington based on a 1996 ship survey and comparisons of passing and closing modes. Southwest Fish. Sci. Cent. Admin. Rep. LJ-97-11, 25 p. some common dolphins ( Delphinus sp.) that were en- countered could be confirmed only to the genus level. Short-beaked common dolphins ( D . delphis) and long- beaked common dolphins ( D . capensis) have very simi- lar morphological features and pigmentation, and they are difficult to distinguish at a distance (Rosel et al., 1994); therefore, observers were encouraged to obtain photographs if there was doubt about the identification of these or other species. Photographs were reviewed onboard, compared with identification guides (Reeves et al., 2002), and occasionally shared with experienced colleagues for species confirmation. Additionally, there are 2 forms of Pacific white-sided dolphins ( Lagenorhyn - chus obliquidem) along the coast of California (Walker et al., 1986; Lux et al., 1997; Soldevilla et al., 2011); because these forms are indistinguishable from a dis- tance, density and abundance for this species likely in- cludes both forms. Analytical methods Sighting and effort data from the 16 CalCOFI cruises were split into 2 effort categories: 1) on-effort sight- ings on the 6 CalCOFI transect lines, the sightings that form the basis of all analyses and findings; and 2) opportunistic effort (sightings made off effort or when a vessel was not on a CalCOFI transect line), which is presented to show species diversity and presence or absence of species. We calculated encounter rates by season (number of on-effort sightings per 1000 km of transect line surveyed) for each species with 10 or more sightings. Distance r of the animal(s) sighted from the vessel and sighting position were determined from the reticle value (or estimated distance), height of observer platform, sighting angle (A) to animal from the bow of the vessel, and position of the vessel. To character- ize the depth distribution of effort, sample points were created at 1-km intervals along transect lines within ArcGIS (vers. 9.2; Esri, Redlands, CA). Sighting data also were plotted in ArcGIS, with sighting and effort data linked to a coastline shape- file and bathymetry data set from the ETOPOl global relief model (Amante and Eakins, 2009) with 1850-m resolution (National Geophysical Data Center [NGDC], http://www.ngdc.noaa.gov/mgg/global/global.html); where highest resolution was not available from ET- OPOl, the NGDC coastal relief model (NGDC, http:// www.ngdc.noaa.gov/mgg/coastal/crm.html ) provided 90-m resolution (73% of sighting depths and 49% of effort depths came from the 90-m resolution data set). Distance to the closest point of land, distance to the mainland, depth, and distance to the shelf break (200- m isobath) were also calculated for each sighting and effort location. Using the cold (winter-spring, defined as Janu- ary-April) and warm (summer-fall, defined as July- November) water distinctions in Forney and Barlow (1998), we tested for differences between the number of encounters of each positively identified species by 202 Fishery Bulletin 112(2-3) Table 2 Covariates and summary statistics for the best-fit detection function models by species group from analyses of data from line-transect surveys conducted off Southern California from 2004 to 2008. The distribution of perpendicular sighting dis- tances for pooled species was used to parameterize the detection function models summarized here, where f( 0) is the prob- ability density function evaluated at a perpendicular distance of zero, bin width (m) is the interval chosen to show the fraction of probability distribution, truncation value excludes the 5% of sightings by species group that were farthest from the transect line and therefore considered outliers, ESW is the effective strip (halDwidth for which the number of groups outside the ESW is equal to the number missed inside the ESW, and CV is the coefficient of variation. The hazard-rate key function was used in the best-fit model for all species groups. Species groups were selected on the basis of factors that influence sightability, such as common school sizes, body shape, and behavior. The species group of large whales consisted of the blue, fin, humpback, sperm, and killer whales. Delphinids included the short- and long-beaked common, northern right whale, Pacific white-sided, Risso’s, and bottlenose dolphins and the Dali’s porpoise. For detection function plots, see Figure 2. Species groups for estimating fl 0) Number of observations Covariates Bin width (m) Truncation (m) Average ESW (m) CV Large whales 127 Perpendicular distance 500 2348 1294 0.11 Delphinids 211 Perpendicular distance, sea state, species, log( group size) 100 1098 298 0.10 Dali’s porpoise 48 Perpendicular distance, sea state 100 1098 305 0.22 seasonal period (winter-spring and summer-fall) and between the distance from shore and depth by the season, with Systat (vers. 13; Systat Software, Chica- go, IL). Because our data had a bimodal distribution, Kruskal-Wallis one-way ANOVAs and Mann-Whitney [/-tests were run to examine variation in encounters by habitat, year, and season with Minitab software (vers. 15.1.30; Minitab, State College, PA). We used line-transect methods (Buckland et al., 2001) with multiple covariates (Marques and Buck- land, 2003) to estimate cetacean abundance for 2 sea- sonal periods and 2 depth categories (defined below) in Distance software (vers. 6.0; Research Unit for Wild- life Population Assessment, University of St. Andrews, UK; Thomas et al., 2010). Distance software uses the perpendicular distance of the encounter to the transect line rather than the straight line distance (observer to animal) made at the time of the sighting; therefore, we calculated perpendicular distance as r sin 0 before uploading these data into Distance. Small sample size by season precluded our ability to estimate abundance quarterly; therefore, we estimated abundance within 4 strata in the study area. The strata were defined by 2 seasonal periods, winter-spring (cold water) or sum- mer-fall (warm water), and 2 depth categories, shal- low (<2000.5 m and deep (>2000.5 m). On the basis of known ecological differences between shelf or slope and basin, with greater density of cetaceans near the coast, we chose to look at these data in 2 depth cat- egories. Additionally, a histogram of effort as a func- tion of depth showed that depth in our study area was strongly bimodal and that the depth of 2000.5 m was an appropriate cutoff point. Sample units were speci- fied by survey number, line number, season, and depth to ensure that each of the 6 transect lines from each survey would be divided into a shallow and a deep sample unit. To estimate density and abundance for a species, it is necessary to reliably estimate the detection func- tion (the probability of seeing an animal at x distance from the transect line), and that requires a relatively large sample size (Buckland et al., 2001). The neces- sary sample sizes were not available for all species in this study; therefore, we pooled multiple species with similar surfacing characteristics (Barlow et al., 2001) and pooled (binned) sightings across season-and-depth strata to estimate the detection function (Table 2, Fig. 2). Pooled species groups were defined as 1) large whales, which included blue (Balaenoptera musculus), fin ( Balaenoptera physalus ), humpback (Megaptera no- vaeangliae), sperm (Physeter macrocephalus), and killer whales ( Orcinus orca)\ 2) delphinids, which included short- and long-beaked common, northern right whale (Lissodelphis borealis ), Pacific white-sided, Risso’s ( Grampus griseus ), and common bottlenose ( Tursiops truncatus) dolphins; and 3) Dali’s porpoises ( Phocoe - noides dalli). Beaked whales and several other species of delphinids were encountered too infrequently to esti- mate abundance. Cetaceans that could not be identified to genus and species levels were not included in the pooled species groups for estimation of the detection function, and density and abundance levels were not estimated for them. Potential covariates for building the detection func- tion models included group size (a categorical variable that denotes whether sightings were greater or less than 20 individuals), cluster size (best estimate of group size), sea state (a numerical variable of 0-5), vessel, and spe- cies. Cut off points for group size were based on obvious breaks in histograms of group size for each species cat- Douglas et al Seasonal distribution and abundance of cetaceans off Southern California 203 r 0 500 — i 1 1 1000 1500 2000 Distance (m) B Distance (m) c Distance (m) Figure 2 Detection function plots by species group ([A] large whales, I B] delphinids, and |C] Dali’s porpoise [ Phocoenoides dalli ]) were created to visualize the correct detection functions to estimate density and abundance for the species most commonly encoun- tered in the study area for line-transect surveys conducted off Southern California in 2004-08 during 16 quarterly California Cooperative Oceanic Fisheries Investigation cruises. The points are the probability of detection for each encounter dependent on its perpendicular distance and chosen covariate(s) for the best fit. The sighting data showed some evidence of heaping ( i.e. , rounding to certain distances) because of the limitations of the use of reticles to estimate distance. Therefore, sightings were binned to facilitate data analysis (Buckland et al., 2001). For detection function covariates and summary statistics, see Table 2. egory. Selection of a detection function model was based primarily on the Akaike’s information criterion (AIC) value (generated with Distance) and then confirmed by visual examination of detection plots (Burnham and An- derson, 2002). Half-normal and hazard-rate key func- tions often provide a good fit to data used to model detection functions (Thomas et ah, 2010). Although both were considered in the models tested, the hazard-rate 204 Fishery Bulletin 112(2-3) model was chosen for all 3 groups on the basis of AIC values and visual inspection (Thomas et al., 2010). Line-transect theory assumes that the probability of detecting an animal on the transect line, g(0) equals 1.0 (perfect detectability), although this assumption is rarely true for marine mammals and can be relaxed if a correction factor is estimated. Estimation of a cor- rection factor was beyond the scope of this study, and g(0) is assumed to equal 1.0 and to be constant across sea states. The 5% of sightings made at the greatest distances to the vessel were assumed to be outliers and were truncated to improve the ability to fit the prob- ability density function f( 0) (Buckland et al., 2001). Truncation distance was 2348 m for large whales and 1098 m for delphinids and porpoises. There were 7 whale, 11 dolphin, and 2 porpoise sightings recorded beyond the truncation distance and these were exclud- ed from density and abundance analyses. We calculated density, Dp for a given species within the study area i as D = J_y-i f{0 1 2 Lj ^=1 Sj<0) ’ where L[ = A0|zj) = gj(0> - n = the length of on-effort transect lines within the study area /; the probability density function at zero with associated covariates z for group; the number of individuals of that species in group j; the transect line detection probability of group j; and the number of groups of that species encoun- tered in the study area i. Group abundance for each species in each stratum was estimated as Nr A x~^n Z-/i= 1 n 2 wL (2) where A L n w Pi = the area of the stratum; = the total search effort in the stratum; = the number of unique groups; = the truncation distance by species group; and = the estimated probability of detecting group i obtained from the fitted detection model. Results Survey effort and sightings Line-transect surveys were conducted during 16 cruis- es over 5 years, with 3 years of 4-season effort and 2 years of 2-season effort (Table 1). Including all survey effort from the southern CalCOFI transect lines, ob- servers collected visual data on marine mammals over 267 days and searched 25,079 km of transect lines. Of that total distance of transect lines covered, 19,996 km was surveyed on the 6 southern CalCOFI lines (hereafter referred to as “the study area”) in accept- able weather conditions (sea state of 0-5). Within the study area, on-effort transect line kilometers surveyed did not vary significantly by season (ANOVA, F= 0.078, P= 0.97). Sea states varied by season, with greatest sea states during spring and summer and lowest sea states during winter (Ivruskal-Wallis one-way ANOVA, P<0.001). The median sea state was 3 in all seasons, except for summer, when it was 4. We found no sig- nificant difference in perpendicular distance to tran- sect line for any species group by vessel (large whales, ANOVA, D=2.08, P=0.16; delphinids and porpoises ANOVA, F=1.01, P- 0.39) or by vessel-season (large whales, ANOVA, F= 2.76, P=0.15; delphinids and por- poises ANOVA, F= 0.36, P- 0.56). As stated in the Materials and methods section, acceptable survey conditions required >0.9 km of es- timated visibility on the transect line. Only 0.67% of effort was conducted with a visibility of <900 m, and that effort resulted in 4 sightings of common dolphins; because those dolphins were not identified to species level, their sightings were not used in the detection function. All encounters used in detection functions for i the 3 species groups were made with at least 2.77 km visibility, with 93% of large whales, 94% of delphinids, and 100% of Dali’s porpoises encountered with visibil- ity >7.4 km. In the study area during the 16 survey cruises, 29 marine mammal species were encountered, including 22 cetaceans, 6 pinnipeds, and a single mustelid spe- cies (Table 3). There were 931 on-effort sightings in the study area, with California sea lions ( 154 recorded sightings) the most commonly encountered species, fol- lowed by short-beaked common dolphins ( 122 sight- ings), northern fur seals ( Callorhinus ursinus, 59 sight- ings), and fin whales (53 sightings). The most common- ly encountered large cetaceans in the study area were fin, humpback (34 sightings), blue (25 sightings), and sperm (20 sightings) whales (Fig. 3), and the most com- monly encountered small cetacean species were short- beaked common dolphin, Pacific white-sided dolphin (46 sightings), and Dali’s porpoise (49 sightings) (Fig. 4). The ratio of on-effort sightings to opportunistic and off-effort sightings of cetaceans (1:0.76) in this study is higher than would be expected for other line-transect surveys where very little sighting effort is conducted off transect or in poor sea conditions. The large number of off-effort and opportunistic sightings in our study is mostly due to the large amount of opportunistic ef- fort between the primary transect lines or at the water sampling stations along the coast. Multispecies sightings were observed on 15 occa- sions for 7 dolphin species; the northern right whale dolphin and Pacific white-sided dolphin mixed most frequently (5 times), bottlenose dolphin and common dolphins mixed 3 times, Pacific white-sided dolphin and short-beaked common dolphin mixed 2 times, striped dolphin (Stenella coeruleoalba ) and unidentified com- mon dolphins mixed 2 times, and single occurrences Douglas et al. Seasonal distribution and abundance of cetaceans off Southern California 205 Table 3 Sighting data (number of encounters and estimated number of individuals) from 16 line-transect surveys conducted off Southern California from July 2004 to April 2008 on the 6 southern transect lines of quarterly California Cooperative Oceanic Fisheries Investigation (CalCOFI) cruises. Opportunistic effort is the category for sightings made off effort or when a vessel was not on a CalCOFI transect line. On effort Opportunistic effort Total Number of Number of Number of Number of Number of Number of Species encounters individuals enccounters individuals encounters individuals Humpback whale 34 60 6 13 40 73 Blue whale 25 34 20 27 45 61 Fin whale 53 100 9 10 62 110 Minke whale 10 10 8 9 18 19 Sei whale 0 0 1 1 1 1 Bryde’s/sei whale 1 1 1 1 2 2 Gray whale 8 16 16 33 24 49 Sperm whale 20 36 11 18 31 54 Baird’s beaked whale 1 20 0 0 1 20 Cuvier’s beaked whale 3 3 1 3 4 6 Killer whale 2 9 0 0 2 9 Short-finned pilot whale 1 33 1 30 2 63 False killer whale 1 10 0 0 1 10 Short-beaked common dolphin 122 11,067 96 8677 218 19,744 Long-beaked common dolphin Common dolphin 17 3259 39 7652 56 10,911 (unknown Short- or long-beaked) 72 6532 75 12,443 147 18,975 Pacific white-sided dolphin 46 573 34 467 80 1040 Risso’s dolphin 14 205 25 505 39 710 Bottlenose dolphin 12 220 18 221 30 441 Northern right whale dolphin 13 480 12 245 25 725 Rough-toothed dolphin 0 0 1 9 1 9 Striped dolphin 2 77 1 7 3 84 Harbor porpoise 0 0 1 2 1 2 Dali’s porpoise 49 267 24 115 73 382 Unidentified large cetacean 123 153 74 98 197 251 Unidentified small cetacean 63 5496 61 3703 124 9199 Unidentified Mesoplodon 1 1 0 0 1 1 Total for cetaceans 693 28,662 535 34,289 1228 62,951 California sea lion 154 1318 85 606 239 1924 Northern fur seal 59 78 14 20 73 98 Guadalupe fur seal 0 0 2 2 2 2 Steller sea lion 1 1 0 0 1 1 Northern elephant seal 11 11 2 2 13 13 Harbor seal 0 0 3 3 3 3 Sea otter 0 0 1 1 1 1 Unidentified fur seal 1 1 0 0 1 1 Unidentified pinniped 12 19 4 4 16 23 Total for pinnipeds or mustelids 238 1428 111 638 349 2066 Grand total Total transect line surveyed (km) 931 19,996 30,090 646 5083 34,927 1577 25,079 65,017 were recorded of Pacific white-sided dolphin with un- identified common dolphins, Risso’s dolphin with north- ern right whale dolphin, and Risso’s dolphin with bot- tlenose dolphin. Harbor porpoise ( Phocoena phocoena) were encoun- tered on a single survey north of Point Conception and were excluded from analyses because the study area in- cluded only the very southern tip of the regular habitat of this species off California (Barlow, 1988; Forney et al., 1991). Rough-toothed dolphins (Steno bredanensis) and false killer whales (Pseudorca crassidens ) were also encountered once each; both encounters were excluded 206 Fishery Bulletin 112(2-3) A 124°W 123°W 122°W 121°W 120°W 119'W 118°W 117°W B 124°W 123"W 122°W 121°W 12CTW 119”W 118“W 117°W c 124°W 123°W 122°W 121“W 120°W 119°W 118°W 117°W D 124°W 123°W 122°W 121°W 120°W 119°W 118”W 117°W PACIFIC OCEAN N A * i 1 0 100 km > Santa Barbara Los Angeles 0 % * V * San Diego n V-X \x Study area zones 167,087 km2 Deep (>=2000 5 m) 71,407 km2 Shallow (<2000.5 m) — 2000-m line Figure 3 Maps of on-effort sightings of the 4 most commonly encountered species of large whales, the (A) humpback whale (Mega- ptera novaeangliae ), (B) blue whale ( Balaenoptera musculus), (C) fin whale (B. physalus), and (D) sperm whale ( Physeter macrocephalus), recorded during the 16 shipboard line-transect surveys conducted quarterly during 2004-08 as part of the California Cooperative Oceanic Fisheries Investigation. from analyses because the encounter records likely rep- resent extralimital occurrences for both species. Abundance estimates from line-transect data Significant covariates for estimation of detection func- tions varied by species and species group (Table 2). On the basis of AIC values and visual examinations of test models, we selected sea state as a significant covariate for both the delphinid and Dali’s porpoise group. Addi- tionally, group size and dolphin species were chosen for the delphinid detection model. Average effective strip width (ESW) was 1294 m for large whales, 298 m for delphinids, and 305 m for porpoises. Baleen whales We encountered 6 or possibly 7 species of baleen whales (one encounter was undetermined; it was not possible to distinguish whether it was a Bryde’s or sei whale [ Balaenoptera edeni brydei, B. e. edeni, or B. borealis]), and sample sizes by species were sufficient to calculate seasonal abundance and density estimates for 3 of those species. Fin whales had the 124°W 123°W 122”W 121°W 120°W 119°W 118°W 117°W Douglas et al. Seasonal distribution and abundance of cetaceans off Southern California 207 n.se Nofr£ n.ee n.ze n.ie n.oe n.6z NoSE N.K: N„££ N.ZE N.LE N.OE N.6Z N.SE N.t?£ N.SE N.ZE N.IE N.OE N.6Z N.SE N.^S N.EE N.ZE N.IE N.OE N.6Z < N.SE N.t’E N.EE N.ZE N.tE N.OE N.6Z N.SE N.t’E N.EE N.ZE N.IE N.OE N.6Z "tf 01 u □ m il c ^ 00 S c =f o .2 1 u £ © Q* o CM T3 > CC bo & • j-t 4-i £ S- Q; CT *5 ^ ' u V o ^ G D — T3 W G <3 ►?> .G a G 5 G Gh Sh g V S g O TD _ G Sm "O •- 03 <1) 43 o 44 ajs 2 o .2- ^ ^ rG O CO 5 •S Q bp ^ G ’£ C/3 p ? 'G S 3T3 a ■~a S ^ c o 3 ^ o a o o o cj G — o a> <- .-So, 42 ~ c a a) a ^ a T3 o a; CJ 'TJ o .3 CD O G cn ^ bb — ' a o a 43 ~ tm to a « a ?> a a) o ^ .2 a G -e o ^ ~ a a >u 03 -73 ~C 5! ® o G Co part of the California Cooperative Oceanic Fisheries Investigation. 208 Fishery Bulletin 112(2-3) Table 4 Encounter rate, the number of encounters (Enc) per 1000 km, and number of sightings n of cetacean species by season on the 6 southern transect lines during 16 quarterly California Cooperative Oceanic Fisheries Investigation (CalCOFI) cruises conducted from 2004 to 2008 off Southern California, for species seen on 10 or more occasions. The value under each season represents the combined length of the transect lines surveyed. Winter Spring Summer Fall All seasons (4165 km) (4904 km) (6211 km) (4716 km) (19,996 km) Species Enc/1000 km, n Enc/1000 km, n Enc/1000 km, n Enc/1000 km, n Enc/1000 km, n Humpback whale 0,0 2.2, 11 1.6, 10 2,7, 13 1.7,34 Blue whale 0,0 0,0 3.1, 19 1.3, 6 1.3,25 Fin whale 0.7,3 1.0, 5 3.7,23 4.7, 22 2.7,53 Minke whale 0,0 1.4, 7 0.3, 2 0.2, 1 0.5, 10 Sperm whale 1.0, 4 0.4, 2 1.9, 12 0.4, 2 1.0,20 Short-beaked common dolphin 7.2,30 0.8,4 9.5,59 6.1, 29 6.1, 122 Long-beaked common dolphin Common dolphin (unknown short- 0.2, 1 0.2, 1 1.3,8 1.5,7 0.9, 17 or long-beaked) 2.6, 11 1.6,8 5.5,34 4.0, 19 3.6, 72 Pacific white-sided dolphin 2.6, 11 4.7, 23 1.1, 7 1.1, 5 2.3, 46 Risso’s dolphin 0.7,3 1.6,8 0.2, 1 0.4, 2 0.7, 14 Bottlenose dolphin 0.5,2 1.0, 5 0.2, 1 0.8,4 0.6, 12 Northern right whale dolphin 0.2, 1 1.8,9 0,0 0,6,3 0.7, 13 Dali’s porpoise 2.6, 11 7.1,35 0.2, 1 0.4, 2 2.5,49 highest encounter rate (Table 4) and were the most abundant of large whales in the study area (Table 5), with the greatest density estimate from summer-fall surveys in shallow water, 3.67 individuals/1000 km2 (CV=0.30) (Tables 6 and 7), and with the greatest abun- dance during the summer-fall surveys in deep water. Fin whales were the only whale species that showed a significant difference in depth, distance to land, and distance to shelf by seasonal period (Table 7; Figs. 5 and 6). Humpback whale density was highest during the summer-fall surveys in shallow water with 3.08 individuals/1000 km2 (CV=0.36) (Table 5). Least abun- dant of the large whales, blue whales were encountered only during the summer-fall surveys, with the greatest density and abundance in shallow water, 3.20 individu- als/1000 km2 and 228 individuals (CV=0.26) (Table 5). Odontocetes Although sperm whales were most abun- dant during the summer-fall surveys in deep water, 158 individuals (CV=0.36) (Table 5), density was simi- lar for both shallow areas (0.94 individuals/1000 km2 [CV=0.44]) and deep areas (0.95 individuals/1000 km2 [CV=0.36]) for that seasonal period. Short-beaked com- mon dolphins were the most abundant cetacean spe- cies, encountered in all seasons and at all depths; the highest encounter rate was observed in the summer months (Table 4) and the greatest density estimate was obtained from summer-fall surveys in shallow wa- ter, 1319.69 individuals/1000 km2 (CV=0.24) (Table 5). Long-beaked common dolphins were the second-most abundant cetacean species; however, this species was encountered only in shallow water and, seasonally, there was a dramatic shift in density with 22 times more long-beaked common dolphins observed during the summer-fall surveys than during the winter- spring surveys (Table 5). Because of the difficulty of distinguishing between short- and long-beaked common dolphins from a survey conducted in passing mode, 72 out of 211 on-effort common dolphin sightings were not identified to species (Table 3). Densities of Pacific white-sided, northern right whale, and Risso’s dolphins were greatest during the winter-spring period in shal- low water, and Dali’s porpoises were most abundant during the winter-spring seasonal period in deep wa- ter; these species were least abundant during the sum- mer-fall period. Abundance of Dali’s porpoises varied strongly by seasonal period but not by depth. Beaked whales were encountered on 6 occasions, with Cuvier’s beaked whale ( Ziphius cavirostris) the most commonly encountered (3 occasions). Discussion Monitoring and management of marine mammal spe- cies off Southern California has often relied heavily on abundance estimates generated from line-transect surveys conducted during the summer and fall, despite year-round anthropogenic activities and significant sea- sonal spatial movements of many species (Forney and Barlow, 1998). Our observations from the 16 CalCOFI surveys conducted between 2004 and 2008 provide the most current and consistent data set on seasonal shifts in movements and abundance for the most commonly Douglas et al. Seasonal distribution and abundance of cetaceans off Southern California 209 Table 5 Density and abundance of cetaceans by species and season-and-depth stratum. The total area of the study area was 238,494 km2, with 71,407 km2 in shallow depths (<2000.5 m) and 167,087 km2 in deep depths (>2000.5 m). Coefficients of variation (CV) apply to both density and abundance estimates, and “NA” indicates that no CV was available because of a sample size equal to zero. Asterisks indicate mean values derived from the separate strati- fied densities for shallow and deep waters. See Table 2 for covariates used to estimate density by species group. Species Number of Mean group Density (individuals/ Uncorrected Season, depth groups size 1000 km2> abundance CV Blue whale Winter— spring, shallow 0 0.0 0.00 0 NA Winter— spring, deep 0 0.0 0.00 0 NA Summer— fall, shallow 19 1.4 3.20 228 0.26 Summer-fall, deep 5 1.2 0.36 59 0.33 Summer-fall, all depths 24 1.4 1.21* 288 0.23 Fin whale Winter-spring, shallow 7 2.0 2.33 166 0.30 Winter-spring, deep 0 0.0 0.00 0 NA Summer-fall, shallow 22 1.4 3.67 262 0.30 Summer-fall, deep 21 2.0 2.49 417 0.42 Summer-fall, all depths 43 1.6 2.84* 679 0.25 Humpback whale Winter-spring, shallow 8 2 2.66 190 0.33 Winter-spring, deep 3 1.7 0.34 56 0.58 Winter-spring, all depths 11 1.9 1.03* 246 0.29 Summer-fall, shallow 19 1.4 3.08 220 0.36 Summer-fall, deep 2 4.5 0.53 89 0.58 Summer-fall, all depths 21 1.7 1.29* 309 0.32 Sperm whale Winter-spring, shallow 1 5 0.83 59 0.70 Winter-spring, deep 5 1.2 0.40 67 0.40 Summer— fall, shallow 3 2.7 0.94 67 0.44 Summer-fall, deep 10 1.6 0.95 158 0.36 Short-beaked common dolphin Winter-spring, shallow 11 57.6 307.83 21,981 0.33 Winter-spring, deep 22 127.8 609.86 101, 900 0.45 Winter-spring, all depths 33 104.3 519.43* 123,881 0.21 Summer-fall, shallow 33 104.8 1319.69 94,235 0.24 Summer-fall, deep 51 37.5 454.35 75,916 0.20 Summer-fall, all depths 84 63.9 713.44* 170,151 0.14 Long-beaked common dolphin Winter-spring, shallow 1 60 30.90 2207 0.78 Winter-spring, deep 0 0.0 0.00 0 NA Summer-fall, shallow 14 217.2 687.87 49,118 0.52 Summer-fall, deep 0 0.0 0.00 0 NA Pacific white-sided dolphin Winter-spring, shallow 19 13.2 110.57 7896 0.38 Winter-spring, deep 14 14 41.91 7002 0.35 Winter-spring, all depths 33 13.5 62.46* 14,898 0.21 Summer-fall, shallow 11 6.6 29.24 2088 0.34 Summer-fall, deep 1 3 0.6967 116 0.72 Summer-fall, all depths 12 6.3 9.24* 2204 0.35 Risso’s dolphin Winter-spring, shallow 9 19 35.65 2546 0.36 Winter-spring, deep 0 0.0 0.00 0 NA Summer— fall, shallow 3 8.6 3.90 279 0.55 Summer-fall, deep 0 0.0 0.00 0 NA table continued 210 Fishery Bulletin 112(2-3) Table S (continued) Density Species Season, depth Number of groups Mean group size (individuals/ 1000 km2' Uncorrected abundance CV Bottlenose dolphin Winter-spring, shallow 4 11.2 22.12 1580 0.54 Winter-spring, deep 1 2 0.97 161 0.79 Summer-fall, shallow 5 13.6 40.32 2879 0.69 Summer-fall, deep 0 0.0 0.00 0 NA Northern right whale dolphin Winter-spring, shallow 5 39 107.31 7662 0.50 Winter-spring, deep 4 13.5 20.30 3392 0.49 Summer-fall, shallow 1 6 6.72 480 0.78 Summer-fall, deep 2 25 11.10 1855 0.68 Dali’s porpoise Winter-spring, shallow 13 4.8 45.50 3249 0.32 Winter-spring, deep 32 5.3 48.65 8128 0.28 Winter-spring, all depths 45 5.1 47.71* 11,378 0.26 Summer-fall, shallow 2 3 2.11 151 0.58 Summer-fall, deep 1 17 2.73 456 0.78 encountered marine mammal species in the Southern California region. Although seasonal variation was seen in cetacean encounters, large numbers of whales and dolphins were observed year-round off Southern California, both on and between transect lines (Table 4; Figs. 3 and 4). Abundance and density of cetaceans: overall comparisons with previous surveys Although our analyses of relative density by seasonal period and depth are robust for the most commonly en- countered species, absolute densities and uncorrected abundances reported here may differ from values re- ported by the NOAA Southwest Fisheries Science Cen- ter (SWFSC) for its previous studies in Southern Cali- fornia; those differences primarily are due to 5 factors. First, our study relied on data collected by the naked eye or with 7x binoculars, hence ESWs by species group in our study were calculated as half (or less) of the ESWs used for the same species groups from 5 years of pooled SWFSC sighting data, which were collected with 25x binoculars as the primary search method (Barlow and Forney, 2007). Second, we assumed that detection on the transect line was certain, org(0)=l; this decision likely had the greatest negative impact on density of cryptic or long-diving species, like sperm whales. Third, we had a relatively high proportion of sightings that were not identified to species, and we did not prorate unidentified cetaceans, thinking that it would be bet- ter to compute a best estimate of cetaceans positively identified to species than to make assumptions about the detectability of unidentified and identified species. Fourth, we used uncalibrated group-size estimates — an approach different from the one for SWFSC cruises in which observers make group-size estimates inde- pendently and each observer is “calibrated” with the use of photogrammetry of select sightings (Gerrodette and Perrin4). Carretta et al. (2011b) stated that uncali- brated group-size estimates could result in estimated counts that were 50% lower than actual group sizes. Fifth, we did not correct for reactions to vessel ap- proach by small cetaceans — an issue that is primarily a concern with the Dali’s porpoise and vessel-attracted dolphin species, like the short-beaked common dolphin. Lastly, the SWFSC southern stratum, with an area of 318,500 km2, is larger than the study area of the Cal- COFI surveys by 25%. We compared density and abun- dance of species from these 2 studies because the 2 ar- eas overlap by 75% and the CalCOFI study area occurs completely inside the SWFSC southern stratum. That said, our study provides the most recent and best repli- cated shipboard assessment of seasonal densities for 11 species of cetaceans off Southern California, including 3 species of baleen whales, the sperm whale, 6 species of delphinids, and the Dali’s porpoise. Baleen whales The most commonly occurring large whales that used this area for feeding were fin, hump- back, and blue whales; because of their presence along the coast in greater numbers during the summer and fall, compared with other seasons, these species have been well represented in previous line-transect and 4 Gerrodette, T., and C. Perrin. 1991. Calibration of ship- board estimates of dolphin school size from aerial photo- graphs. Southwest Fish. Sci. Cent. Admin. Rep. W-91-36, 73 P- Douglas et al. Seasonal distribution and abundance of cetaceans off Southern California 211 212 Fishery Bulletin 112(2-3) Table 7 Summary of results from Kruskal-Wallis and Mann-Whitney tests of seasonality for depth, dis- tance from vessel to land, and distance from vessel to shelf break for cetaceans encountered on the transect lines of the California Cooperative Oceanic Fisheries Investigation (CalCOFI) surveys conducted from 2004 to 2008 across all 4 seasons. Probability values that indicate significant results are shown in bold type. Asterisks indicate species for which tests were run with a limited number of sightings for one or more seasons; see Table 4 for number of sightings. Species Depth (P-value) Distance to land (P-value) Distance to shelf break (P-value) Humpback whale 0.007 0.149 0.191 Blue whale 0.799 0.611 0.484 Fin whale* 0.003 <0.000 0.001 Minke whale 0.079 0.040 0.143 Sperm whale* 0.783 0.729 0.732 Short-beaked common dolphin* 0.122 0.005 0.008 Long-beaked common dolphin 0.247 0.037 0.015 Common dolphin (unknown short- or long-beaked) 0.773 0.422 0.587 Pacific white-sided dolphin 0.014 0.005 0.008 Risso’s dolphin* 0.581 0.136 0.087 Bottlenose dolphin* 0.256 0.223 0.252 Northern right whale dolphin* 0.518 0.518 0.518 Dali’s porpoise* 0.889 0.941 0.929 photoidentification studies. The large whales repre- sented here are highly visible from a distance and of- ten occur in small groups; therefore, confidence levels on group-size estimates are higher and abundance es- timates likely are more accurate for them than for the smaller cetaceans that occur in large and more variable size groups. Group sizes for fin, humpback, and blue whales (Table 5) were similar to group sizes reported by Barlow (2010) and Barlow and Forney (2007). The number of unidentified large cetacean encounters (123) is to be expected from a survey conducted in passing mode. Although we did not apportion encounters of un- identified species in our analyses, on the basis of the proportion of large whale species positively identified, it is likely that fin, humpback, and blue whales made up the majority of these sightings. Fin whales were the most commonly encountered and most abundant large whale in the study area. As has been documented by Forney and Barlow (1998), fin whales were encountered during all seasons, but the encounter rate for this species increased during the summer and fall seasons. Our abundance of 679 individuals (CV=0.25) in the study area during the summer-fall seasonal period is similar to Barlow’s (2010) estimate of 499 individuals (CV=0.27) from a 2008 survey for Southern California and higher than the abundance estimate of 359 individuals (CV=0.40) from surveys conducted in 1991-2005 (Barlow and For- ney, 2007). Given the differences in survey design, we would have expected our abundance estimate to have been lower than the values presented in Barlow (2010); however, annual variability, which we do not address in this study of multiyear CalCOFI surveys, may account for the difference in abundance estimates. Broadly, these patterns of increasing abundance are consistent with the recently documented trend of an increasing population for fin whales (Moore and Barlow, 2011). Although there were few encounters during winter, fin whales used nearshore waters in the winter and spring and shifted into offshore waters in the summer and fall (Tables 6 and 7; Figs. 5 and 6); this movement seems to coincide with the observed coldest temperatures in nearshore waters recorded in winter and spring and with a slight increase of zooplankton biomass that oc- curs in the spring (Munger et ah, 2009). Although humpback whales were the second-most frequently encountered large cetacean, none were sighted during the winter. Our abundance estimate of 309 individuals (CV=0.32) for the summer-fall period is almost 6 times larger than the estimates of 49 indi- viduals (CV=0.43) from the 2008 survey (Barlow, 2010) and of 36 whales (CV=0.51) from pooled 1991-2005 surveys (Barlow and Forney, 2007); however, just over half of our on-effort humpback sightings came from the 2007 summer cruise near Point Conception, where and when zooplankton abundance was notably high (Munger et ah, 2009), indicating that an unusually large proportion of the population shifted into this area to take advantage of available prey. Because Southern California represents the southern end of the hump- back whale’s feeding range, such annual variation in available prey could strongly affect the abundance of Douglas et al. Seasonal distribution and abundance of cetaceans off Southern California 213 124°W 123°W 122°W 121°W 120°W 119°W 118°W 117°W 116°W Figure 5 Map of on-effort encounters with the fin whale ( Balaenoptera physalus) by season, recorded during the 16 shipboard line-transect surveys conducted quarterly during 2004-08 as part of the Cali- fornia Cooperative Oceanic Fisheries Investigation. The color of the triangle indicates the season: blue=winter, green=spring, red=summer, and yellow=fall. this species in our study area. Humpback whales off California, Oregon, and Washington migrate season- ally to wintering grounds off Baja, California, main- land Mexico, and Central America (Steiger et ah, 1991; Calambokidis et ah, 2000; Urban et ah, 2000). Clapham et al. (1997) and Forney and Barlow (1998) noted that in waters off California, a significantly greater propor- tion of the humpback whale population was found far- ther offshore in winter than in summer. We also found that more humpback whales occurred farther offshore and in deeper water in spring than during the sum- mer-fall seasonal period (Table 6). The thirdmost frequently encountered and abun- dant baleen whale within the study area, the blue whale, showed a distinct seasonal presence, a result that concurs with the findings from year-round aerial and ship-based surveys off Southern California. Forney and Barlow (1998) and Larkman and Veit (1998) found the greatest abundance of blue whales during August- October. Our abundance estimate of 288 blue whales (CV=0.23) was much lower than Barlow and Forney’s (2007) estimate of 842 individuals (CV=0.20) off South- ern California. The discrepancy between these esti- mates is likely due to a few factors, including interan- nual differences in proportion of the population found within the study area and our lack of a correction fac- tor for transect-line detection probability. Barlow and Forney’s (2007) estimate included surveys completed in 1991, 1993, and 1996, all years when much of the blue whale population was thought to be feeding along the California coast; however, in more recent years, evi- dence has indicated that blue whales are using more northerly, southerly, and offshore waters (Calamboki- dis and Barlow, 2004; Barlow and Forney, 2007; Calam- bokidis et ah, 2009). No blue whales were encountered in the study area during the winter— spring period — a finding that corresponds to their known migration pat- tern of feeding off California from May to November and migrating south to spend winter and spring off Mexico (Calambokidis et ah, 1990; Mate el al., 1999; Stafford et ah, 1999) and as far south as 6°N at the Costa Rica Dome (Wyrtki, 1964). Although known to be present year-round (Dohl et al.5; Forney et ah, 1995; Barlow3), minke whales 5 Dohl, T. P., R. C. Guess, M. L. Duman, and R. C. Helm. 1983. Cetaceans of central and northern California, 1980-1983: status, abundance, and distribution. Part of investigator’s fi- 214 Fishery Bulletin 112(2-3) 300 E 1 200 O CD o c ro t 100 0 | Winter Spring Summer Fall Season Figure 6 Box-and-whisker plot showing distance to land by season for fin whales (Balaenoptera physalus) within the study area for line- transect surveys conducted off Southern California during 2004- OS for 16 quarterly California Cooperative Oceanic Fisheries In- vestigation cruises. In each box, the middle horizontal line shows the median value and the upper and lower lines show the 75th and 25th percentiles. Ends of the upper and lower whiskers indi- cate the minimum and maximum data values; an * indicates the outlier and the vertical lines extend to a maximum of 1.5 times the interquartile range. * (Balaenoptera acutorostrata) are difficult to sight even in very good sea conditions. The sample size of this species in the 6 CalCOFI surveys was insufficient for an abundance estimate, but it is worth noting that we encountered minke whales in low numbers from spring to fall, and a peak in encounter rates occurred during the spring (Table 4) that cannot be explained by bet- ter sea conditions in spring. Although sei whales were historically the fourth-most commonly captured whale along coastal California during whaling activity in the 1950s and 1960s (Rice, 1974), they now are considered rare in California waters (Dohl et al.5; Mangels and Gerrodette, 1994; Forney et al., 1995; Barlow3). Our results support findings that they are not commonly encountered off southern California with only a sin- gle sighting of a sei whale and a sighting of one other individual that was either a sei whale or a Bryde’s whale. Odontocetes We encountered 16 species of odontoce- tes, with sufficient sightings of 8 species to calculate seasonal abundance and density and examine seasonal trends. The most commonly encountered odontocete species along Southern California are present year- nal report, Marine Mammal and Seabird Study, central and northern California, Contract No. 14-12-0001-29090. Pre- pared by Center for Marine Sciences, Univ. California, Santa Cruz, for the Pacific OCS Region, Minerals Management Service, OCS Study MMS 84-0045, 284 p. round, although some of them undergo sea- sonal shifts in abundance; the Dali’s porpoise and Risso’s dolphin have been recognized as moving seasonally into Southern California waters during the winter months. Such sea- sonal shifts of abundance out of Southern California waters during winter months in- creases the likelihood that these species were regionally underrepresented in previous es- timates (Barlow and Forney, 2007; Carretta et al., 2011b) of density and abundance that were generated from sighting data collected during summer-fall ship-based surveys. Sufficient sample size allowed for density and abundance estimation of sperm whales; however, mean group size (2.7 individuals) was significantly lower than the 8.1 indi- viduals reported off Southern California from pooled sightings collected over 5 years of SWFSC surveys (Barlow and Forney, 2007). In our study, group-size estimates were very likely negatively biased by the constraints of conducting a survey in passing mode, in- stead of using the protocol for the SWFSC line-transect surveys of conducting multiple counts over 90 min to enumerate asynchro- nously diving whales (Barlow and Taylor, 2005; Barlow and Forney, 2007). We encoun- tered sperm whales year-round and in both depth categories, but we observed this spe- cies primarily during the summer-fall period in depths >2000.5 m — findings similar to earlier analyses of year- round survey effort (Dohl et al.5; Barlow, 1995; Forney et al., 1995). Even with our relatively high number of common dolphin sightings that could not be identified to spe- cies, we found that short-beaked common dolphins were the most abundant and widely distributed cetacean in our study area — a finding that is consistent with previ- ously published results from cetacean survey effort off Southern California (Leatherwood et al., 1982; Dohl et al., 1986; Smith et al., 1986; Barlow, 1995; Forney et al., 1995). Moreover, our stratified abundance estimates provide clear evidence of seasonal shifts in habitat use. We found that, during the summer-fall period, short- beaked common dolphins were fairly evenly spread throughout the study area, and, during the winter- spring period, there was a surge in abundance of this species into offshore waters (mean group size: 127.7 in- dividuals; abundance: 101,900 individuals [CV=0.45]). The greatest seasonal abundance estimate (170,151 in- dividuals [CV=0.14]) was from the summer-fall period, a level that is very close to Barlow and Forney’s (2007) estimate for that seasonal period of 165,400 individu- als (CV=0.19). From aerial and ship-based line-transect surveys, the abundance of short-beaked common dol- phins off California has been shown to change on sea- sonal and interannual times scales (Dohl et al., 1986; Barlow, 1995; Forney et al., 1995). Douglas et al. Seasonal distribution and abundance of cetaceans off Southern California 215 Long-beaked common dolphins were the fourthmost commonly encountered and secondmost abundant small cetacean in the study area. Distribution of long-beaked common dolphins was limited to waters near the Cali- fornia coast or Channel Islands — a result that is consis- tent with findings that this species is commonly found within ~93 km of the coast and ranges from Baja Cali- fornia to central California, with the highest densities observed during warm-water events throughout their range (Heyning and Perrin, 1994). The uncorrected abundance estimate from summer-fall surveys at shal- low depths for long-beaked common dolphins (49,118 individuals [CV=0.52]) was about 3 times higher than Barlow’s (2010) abundance estimate determined from pooled data from line-transect surveys conducted during 1991-2008, but our mean group size (217.2 individuals [CV=0.52]) and abundance estimate were much lower than his mean group size and abundance estimates from the 2009 line-transect survey, where corrected mean group size was 481.0 individuals and abundance was 111,738 individuals (CV=0.44) (Carretta et al., 2011a). Our estimates are likely negatively biased, given the relatively large number of common dolphin sightings that were not identified to species. However, the 2009 es- timates were much greater than the results from earlier surveys, and there was an indication that the moderate El Nino event in 2009 may have caused an influx of dolphins from the south. Our surveys, conducted during 2004-08, show that this species is present year-round but increases 22-fold in abundance during the summer- fall period, indicating that dolphins are shifting south for the winter and spring. Although the number of sightings was insufficient from the winter-spring period to quantify year-round seasonality of long-beaked common dolphins, this study is the first to provide evidence of seasonal habitat use for the 2 common dolphin species found along Southern California. For previous publications that have docu- mented seasonality, aerial surveys were used for cold- water seasonal surveys; however, at the time of those studies, there was not an effective method for distin- guishing the 2 species from an aerial platform (Dohl et al., 1986; Forney et al., 1995; Forney and Barlow, 1998). In marked contrast to the ratio of encounters of short- and long-beaked common dolphins reported here (6:1), Carretta et al. (2011a) encountered the 2 species at a 1:1 ratio in 2009; their observation supports the hypothesis of a dramatic shift or pronounced interan- nual variability from the preceding years off Southern California. Pacific white-sided dolphins were encountered in all seasons, with the greatest abundance estimate (14,898 individuals [CV=0.21]) from both depth categories com- bined in the winter-spring seasonal period. Although density was markedly different between the shallow and deep categories during the winter-spring season, abundance was fairly constant throughout the entire study area. During the summer-fall period, we found that density and abundance (9.24 individuals/1000 km'2; 2204 individuals [CV=0.35]) decreased by almost 15% from the previous winter-spring period, with greater abundance in shallow waters than in deep waters. Bar- low and Forney (2007) published a similar pooled abun- dance estimate of 2196 individuals (CV=0.39) for sur- veys conducted in all depths during the summer-fall period during 1991-2005. From the data on encounters by season, we found that a significant shift into deep water occurred during the winter-spring period (Table 7, Fig. 7). Along the coast of California, the 2 forms of Pacific white-sided dolphins are primarily found in wa- ters over the continental shelf and slope (Forney, 1994). The northern form is thought to enter coastal South- ern California waters during the winter months and to congregate with the southern form (Walker et al., 1986; Lux et al., 1997; Soldevilla et ah, 2011). Because we were unable to differentiate between the 2 forms, it is possible that the increase in observed abundance dur- ing the winter-spring season was a result of capturing both forms that use the study area rather than captur- ing only the southern form. Risso’s dolphins were encountered year-round in shallow water, with abundance estimates of 2546 indi- viduals (CV=0.36) for the winter-spring period and of 279 individuals (CV=0.55) for the summer— fall period. Our findings agreed with those from visual surveys that found high seasonal variability in occurrence and distribution of this species off California (Shane, 1994; Forney and Barlow, 1998; Kruse et ah, 1999; Benson et ah, 2002; Barlow and Forney, 2007) and that their abundance along the California coast could be an or- der of magnitude higher during the winter than dur- ing the summer (Forney and Barlow, 1998). However, further research is needed to understand our results in relation to the findings of Soldevilla et al. (2010), who found peak Risso’s dolphin echolocation activity off Southern California in the fall. On the basis of genetics and morphology, bottlenose dolphins along the coast of California and elsewhere worldwide are split into offshore and coastal popula- tions (Hansen, 1990; Carretta et ah, 1998; Defran and Weller, 1999; Bearzi et ah, 2009; Perrin et ah, 2011). The Southern California coastal population typically is encountered within 500 m of shore (this species was sighted within that boundary 99% of the time during a previous study; Hanson and Defran [1993]), and the offshore population is found outside of a few kilome- ters from the mainland. The mean distance from a land mass that bottlenose dolphins were recorded in this study was 34 km; the minimum distance was just over 2 km. The study area did not include nearshore waters sufficiently to encounter coastal bottlenose; therefore, we assume that our abundance estimate is for the offshore bottlenose dolphin population. For our stratum of the summer-fall period and shallow depth, the abundance estimate (2879 individuals [CV=0.69]) is greater than Barlow and Forney’s (2007) abundance estimate (1831 individuals [CV=0.47]) for this popula- tion off Southern California during the same period. 216 Fishery Bulletin 112(2-3) 124°W 123°W 122°W 121°W 120°W 119°W 118°W 117°W 116°W i i i i i i i i i Figure 7 Map of on-effort encounters with the Pacific white-sided dolphin ( Lagenorhynchus obliquidens ) by season during the 16 shipboard line-transect surveys conducted quarterly during 2004-08 as part of the California Cooperative Oceanic Fisheries Investigation. The color of the triangle indicates the season: blue=winter, green=spring, red=summer, and yellow=fall. In addition to the high CV value associated with our abundance estimate, a likely cause of this discrepancy between the 2 studies is the difference in estimated group size, where we observed an average of 40.5 in- dividuals in a group and Barlow and Forney (2007) re- ported 13.4 individuals in a group. The Northern right whale dolphin and Dali’s porpoise are known to favor cold waters, and we found both spe- cies to have the greatest abundance estimates during the winter-spring period over all depths. Although encoun- ters with northern right whale dolphins in the summer- fall period were few, an increase in density during the winter-spring surveys in shallow water was observed — a finding that is consistent with earlier records that found this species beyond the continental slope for warm-water seasons and in shelf waters of the Southern California Bight for the cold-water season (Barlow, 1995; Forney et al., 1995; Forney and Barlow, 1998). Although seasonally abundant, Dali’s porpoises are often initially sighted when they react to survey ves- sels, thereby biasing abundance estimates upward. To compensate for vessel attraction, Barlow and Forney (2007) included only Dali’s porpoise sightings made in sea states of 0-2 — an approach that they noted limited sample size. On the basis of the detection model for Dali’s porpoises (Fig. 2), which showed an even taper- ing of sightings with distance from the vessel, we in- cluded sightings in sea states of 0-5, assuming that it would be better to have a greater number of sight- ings than an insufficient number to estimate abun- dance. Spatially, our analysis of encounters with Dali’s porpoises in the CalCOFI study area agrees with the finding of Morejohn (1979) that Dali’s porpoises were commonly seen in small groups along the shelf and slope and in offshore waters. Dali’s porpoises were con- sistently found in recently upwelled waters near shore (Peterson et al., 2006). In the CalCOFI study area, the highest encounter rates of Dali’s porpoises occurred in spring, when upwelling waters were active. As with the Dali’s porpoise, many of the delphinids are known to react to a vessel before visual observers can detect them; this behavior is especially a concern when the naked eye and low-power binoculars are used in the search method, as they were in the CalCOFI surveys used in this study. Although reaction to ves- sel cannot be ruled out as a factor in our results, our decision to keep all on-effort sightings in the analyses was based on the detection model for delphinids (Fig. Douglas et al. Seasonal distribution and abundance of cetaceans off Southern California 217 3) that showed an even tapering of sightings with dis- tance from vessel. Our results on seasonal occurrence of the 6 fre- quently occurring delphinid species and the Dali’s porpoise are consistent with prior findings. The bottle- nose dolphin, long-beaked common dolphin, and short- beaked common dolphin generally favor warm-water (summer-fall) periods along the California coast (Dohl et al., 1986; Barlow, 1995; Forney et al., 1995). Dali’s porpoise, the Pacific white-sided dolphin, the northern right whale dolphin, and Risso’s dolphin commonly are found during the cold-water (winter— spring) periods off Southern California, and these species tend to migrate north into central California or Oregon and Washing- ton during the warm-water periods (Forney, 1994; For- ney and Barlow, 1998). These species have exhibited abundance shifts associated with oceanographic vari- ability on both seasonal and interannual time scales (Perrin et al., 1985; Heyning and Perrin, 1994; Forney, 1997; Forney and Barlow, 1998; Becker, 2007). There were only 6 sightings of beaked whales, but all 3 genera ( Ziphius , Berardius, and Mesoplodon) known to be present off Southern California were de- tected. The single sighting of a Mesoplodon could not be confirmed to species. A single encounter with a group of Baird’s beaked whales ( Berardius bairdii) near the shelf break during a survey in the summer is consis- tent with other sightings of this species in continental slope waters from late spring to early fall (Balcomb, 1989; Carretta et al., 2011b). Of the 11 dolphin species encountered, 5 species were represented by only 1-3 sightings per species: killer whale, false killer whale, short-finned pilot whale ( Glo - bicephala macrorhynchus), rough-toothed dolphin, and striped dolphin. Of these 5 species, only the killer whale is commonly found year-round off Southern California, with 2 U.S. stocks (Eastern North Pacific Transient and Eastern North Pacific Offshore [Carretta et al., 2011b]) that use the area. We were unable to confirm which stocks were represented in the 2 sightings of this spe- cies. The rough-toothed dolphin and false killer whale are considered rare off California, with no known current or historical populations along the West Coast of the United States; therefore, our sightings likely represent extralimital movements from populations farther south. There were too few encounters with striped dolphins in the study area to look at seasonal shifts in habi- tat; however, it is worth noting that the 3 sightings of this species occurred in surveys conducted in the sum- mer-fall period, in the deepest mean water depth, and at the greatest mean distance to land of any species observed in the study area (Table 6). Season, distance to shore, and depth of striped dolphin encounters cor- respond with those of previous surveys conducted in summer and fall and with habitat models that revealed the presence of striped dolphins in tropical to warm- temperate pelagic waters, with a continuous distri- bution outside upwelling coastal waters of California (Perrin et al., 1985; Jefferson et al., 1993; Mangels and Gerrodette, 1994; Archer and Perrin, 1999; Becker et al., 2012; Forney et al., 2012). Short-finned pilot whales were encountered com- monly off Southern California before the El Nino event in 1982-83 (Dohl et al.1); on the basis of numerous sur- veys, including this one, it is apparent that this spe- cies now uses these waters only infrequently (Carretta and Forney, 1993; Shane, 1994; Barlow3; Forney, 2007). The single encounter of false killer whales in the study area occurred during the 2008 winter cruise at a depth of 300 m and within 5 km of Santa Rosa Island. False killer whales are normally found in tropical to warm- temperate oceans; however, sightings have been made occasionally in cold-temperate areas as well (Stacey and Baird, 1991; Baird, 2008). Conclusions We collected sighting data from seasons and years that have not been reported previously, generated density and abundance estimates for 11 species of cetaceans off Southern California, and documented shifts in seasonal distribution for fin whales and Pacific white-sided dol- phins. In recent years, interest has increased in the development of predictive models to forecast near real- time marine mammal distribution as a way to inform, mitigate, and decrease the effect of potentially harmful human activities in the marine environment (Becker et al., 2012; Forney et al., 2012; Thompson et al., 2012; Henderson et al., 2014). Although our data set spans a 5-year period that ends in 2008, visual and acoustic data on detections of marine mammals continue to be collected with corresponding oceanographic data, both physical and biological, during CalCOFI cruises. As the CalCOFI data set grows, it potentially could become one of the most valuable collections of data both for monitoring and creating year-round habitat models of cetacean species and their environment off Southern California. Acknowledgments F. Stone, E. Young, and L. Petitpas and the Office of Naval Research provided funding and project manage- ment. We appreciate the efforts of all who were in- volved in the CalCOFI surveys in 2004-08: the cap- tains and crews of the New Horizon, David Starr Jor- dan, and Roger Revelle and the scientists, especially D. Wolgast, J. Wilkinson, A. Hays, R. Baird, S. Yin, M. Smith, A. Miller, L. Morse, V. Iriarte, E. Vazquez, N. Rubio, K. Merkens, J. Burtenshaw, E. Oleson, and E. Henderson. We also thank K. Forney and R. Baird for manuscript review. Finally, the authors would like to honor S. Claussen, whose presence and laughter is greatly missed. 218 Fishery Bulletin 112(2-3) Literature cited Amante, C., and B. W. Eakins. 2009. ETOPOl 1 Arc-Minute Global Relief Model: proce- dures, data sources and analysis. NOAA Tech. Memo. NESDIS-NGDC-24, 19 p. Archer, F. I., II, and W. F. Perrin. 1999. Stenella coeruleoalba. Mamm. Species 603:1-9. Baird, R. W. 2008. False killer whale, Pseudorca crassidens. In En- cyclopedia of marine mammals, 2nd ed. (W. F. Perrin, B. Wiirsig, and J. G. M. Thewissen, eds.), p. 405-406. Aca- demic Press, Amsterdam. Balcomb, K. C., III. 1989. Baird’s beaked whale Berardius bairdii Stejneger, 1883: Arnoux’s beaked whales Berardius arnuxii Du- vernoy, 1851. In Handbook of marine mammals, vol. 4: river dolphins and the larger toothed whales (S. H. Ridgway and R. Harrison, eds.), p. 261-288. Academic Press, New York. Barlow, J. 1988. Harbor porpoise, Phocoena phoeoena, abundance estimation in California, Oregon, and Washington: I. Ship surveys. Fish. Bull. 86:417-432. 1995. The abundance of cetaceans in California waters. Part I: Ship surveys in summer and fall of 1991. Fish. Bull. 93:1-14. 2010. Cetacean abundance in the California Current estimated from a 2008 ship-based line-transect sur- vey. NOAA Tech. Memo. NMFS-SWFSC-456, 19 p. Barlow, J., and K. A. Forney. 2007. Abundance and population density of cetaceans in the California Current ecosystem. Fish. Bull. 105:509-536. Barlow, J., T. Gerrodette, and J. Forcada. 2001. Factors affecting perpendicular sighting distances on shipboard line-transect surveys for cetaceans. J. Cetacean Res. Manage. 3:201-212. Barlow, J., and B. L. Taylor. 2005. Estimates of sperm whale abundance in the north- eastern temperate Pacific from a combined acoustic and visual survey. Mar. Mamm. Sci. 21:429-445. Bearzi, M., C. A. Saylan, and A. Hwang. 2009. Ecology and comparison of coastal and offshore bottlenose dolphins (Tursiops truncatus) in Califor- nia. Mar. Freshw. Res. 60:584—593. Becker, E. A. 2007. Predicting seasonal patterns of California ceta- cean density based on remotely sensed environmental data. Ph.D. diss., 303 p. Univ. California, Santa Bar- bara, CA. Becker, E. A., D. G. Foley, K. A. Forney, B. Barlow, J. V. Red- fern, and C. L. Gentemann. 2012. Forecasting cetacean abundance patterns to en- hance management decisions. Endang. Species Res. 16:97-112. Benson, S. R, D. A. Croll, B. B. Marinovic, F. P. Chavez, and J. T. Harvey. 2002. Changes in the cetacean assemblage of a coastal upwelling ecosystem during El Nino 1997-98 and La Nino and La Nina 1999. Prog. Oceanogr. 54:279-291. Berman-Kowalewski, M., F. M. D. Gulland, S. Wilkin, J. Calambokidis, B. Mate, J. Cordaro, D. Rotstein, J. St. Leger, P. Collins, K. Fahy, and S. Dover. 2010. Association between blue whale (Balaenoptera musculus ) mortality and ship strikes along the Califor- nia coast. Aquat. Mamm. 36:59-66. Bograd, S. J., D. A. Checkley, and W. S. Wooster. 2003. CalCOFI: a half century of physical, chemi- cal, and biological research in the California Cur- rent System. Deep Sea Res. (II Top. Stud. Oceanogr.) 50:2349-2353. Buckland, S. T., D. R. Anderson, K. P. Burnham, J. L. Laake, D. L. Borchers, and L. Thomas. 2001. Introduction to distance sampling: estimating abundance of biological populations, 448 p. Oxford Univ. Press, Oxford, UK. Buckland, S. T., D. R. Anderson, K. P. Burnham, and J. L. Laake. 1993. Distance sampling: estimating abundance of bio- logical populations, 446 p. Chapman and Hall, London. Burnham, K. R, and D. R. Anderson. 2002. Model selection and multimodel inference: a prac- tical information-theoretic approach, 2nd ed., 488 p. Springer- Verlag, New York. Calambokidis, J., and J. Barlow. 2004. Abundance of blue and humpback whales in the eastern North Pacific estimated by capture-recapture and line-transect methods. Mar. Mamm. Sci. 20:63-85. Calambokidis, J., J. Barlow, J. K. B. Ford, T. E. Chandler, and A. B. Douglas. 2009. Insights into the population structure of blue whales in the eastern North Pacific from recent sight- ings and photographic identifications. Mar. Mamm. Sci. 25:816-832. Calambokidis, J., G. H. Steiger, J. C. Cubbage, K. C. Balcomb, C. Ewald, S. Kruse, R. Wells, and R. Sears. 1990. Sightings and movements of blue whales off cen- tral California 1986-88 from photo-identification of individuals. Rep. Int. Whal. Comm, (special issue) 12:343-348. Calambokidis, J., G. H. Steiger, K. Rasmussen, J. Urban R., K. C. Balcomb, P. Ladron de Guevara R, M. Salinas Z., J. K. Jacobsen, C. S. Baker, L. M. Herman, S. Cerchio, and J. D. Darling. 2000. Migratory destinations of humpback whales that feed off California, Oregon and 'Washington. Mar. Ecoi. Prog. Ser. 192:295—304. Carretta, J. V., S. J. Chivers, and W. L. Perryman. 2011a. Abundance of the long-beaked common dolphin (Delphinus capensis ) in California and western Baja California waters estimated from a 2009 ship-based line-transect survey. Bull. South. Calif. Acad. Sci. 110:152-164. Carretta, J. V., and K. A. Forney. 1993. Report of the two aerial surveys for marine mam- mals in California coastal waters utilizing a NOAA DeHavilland twin otter aircraft March 9 April 7, 1991 and February 8 April 6, 1992. NOAA-TM-NMFS-SWF- SC-185, 77 p. Carretta, J. V., K. A. Forney, and J. L. Laake. 1998. Abundance of southern California coastal bot- tlenose dolphins estimated from tandem aerial sur- veys. Mar. Mamm. Sci. 14:655-675. Carretta, J. V., K. A. Forney, E. Oleson, K. Martien, M. M. Muto, M. S. Lowry, J. Barlow, J. Baker, B. Hanson, D. Lynch, L. Carswell, R. L. Brownell, J. Robbins, D. K. Mattila, R. L. Brownell Jr., J. Robbins, D. K. Mattila, K. Ralls, and M. C. Hill. 2011b. U.S. Pacific marine mammal stock assessments: 2011. NOAA Tech. Memo. NMFS-SWFSC-488, 356 p. Douglas et al. Seasonal distribution and abundance of cetaceans off Southern California 219 Clapham, P. J., S. Leatherwood, I. Szczepaniak, and R. L. Brownell. 1997. Catches of humpback and other whales from shore stations at Moss Landing and Trinidad, California, 1919-1926. Mar. Mamm. Sci. 13:368-394. Defran, R. H., and D. W. Weller. 1999. Occurrence, distribution, site fidelity, and school size of bottlenose dolphins ( Tursiops truncatus) off San Diego, California. Mar. Mamm. Sci. 15:366-380. Dohl, T. P., M. L. Bonnell, and R. G. Ford. 1986. Distribution and abundance of common dolphin, Delphinus delphis , in the Southern California Bight: a quantitative assessment based upon aerial transect data. Fish. Bull. 84:333-344. Forney, K. A. 1994. Recent information on the status of odontocetes in Californian waters. NOAA Tech. Memo. NMFS-SWF- SC-202, 87 p. 1997. Patterns of variability and environmental models of relative abundance for California cetaceans. Ph.D. diss., 260 p. Scripps Inst. Oceanography, Univ. Califor- nia, San Diego, CA. 2007. Preliminary estimates of cetacean abundance along the U.S. West Coast and within four National Marine sanctuaries during 2005. NOAA Tech. Memo. NMFS-SWFSC-406, 27 p. Forney, K. A., and J. Barlow. 1998. Seasonal patterns in the abundance and distribu- tion of California cetaceans, 1991-1992. Mar. Mamm. Sci. 14:460-489. Forney, K. A., J. Barlow, and J. V. Carretta. 1995. The abundance of cetaceans in California waters. Part II: aerial surveys in winter and spring of 1991 and 1992. Fish. Bull. 93:15-26. Forney K. A., M. C. Ferguson, E. A. Becker, P. C. Fiedler, J. V. Redfern, J. Barlow, I. L. Vilchis, and L. T. Ballance. 2012. Habitat-based spatial models of cetacean density in the eastern Pacific Ocean. Endang. Species Res. 16:113-133. Forney, K. A., D. A. Hanan, and J. Barlow. 1991. Detecting trends in harbor porpoise abundance from aerial surveys using analysis of covariance. Fish. Bull. 89:367-377. Hansen, L. J. 1990. California coastal bottlenose dolphins. In The bottlenose dolphin (S. Leatherwood and R. R. Reeves, eds.), p. 403-420. Academic Press, San Diego, CA. Hanson, M. T., and R. H. Defran. 1993. The behaviour and feeding ecology of the Pacific coast bottlenose dolphin, Tursiops truncatus. Aquat. Mamm. 19:127-142. Henderson, E., K. A. Forney, J. P. Barlow, J. A. Hildebrand, A. B. Douglas, J. Calambokidis, and W. J. Sydeman. 2014. Effects of fluctuations in sea-surface tempera- ture on the occurrence of small cetaceans off Southern California: implications for climate change. Fish. Bull. 112:159-177. Heyning J. E., and W. F. Perrin. 1994. Evidence for two species of common dolphins (ge- nus Delphinus) from the eastern North Pacific. Nat. Hist. Mus. Los. Ang. Cty., Contrib. Sci. 442:1-35. Jefferson T. A., S. Leatherwood, and M. A. Webber. 1993. Marine mammals of the world. FAO species iden- tification guide, 320 p. FAO, Rome. Jensen, A. S., and G. K. Silber. 2003. Large whale ship strike database. NOAA Tech. Memo. NMFS-OPR-25, 37 p. Julian, F., and M. Beeson. 1998. Estimates of marine mammal, turtle, and seabird mortality for two California gillnet fisheries: 1990— 95. Fish. Bull. 96:271-284. Kinzey, D., and T. Gerrodette. 2003. Distance measurements using binoculars from ships at sea: accuracy, precision and effects of refrac- tion. J. Cetacean Res. Manage. 5:159-171. Kruse, S., D. Caldwell, and M. Caldwell. 1999. Risso’s dolphin. In Handbook of marine mam- mals, vol. 6: the second book of dolphins and porpoises (S. H. Ridgway and R. Harrison, eds.), p. 183-212. Aca- demic Press, London. Laist, D. W., A. R. Knowlton, J. G. Mead, A. S. Collet, and M. Podesta. 2001. Collisions between ships and whales. Mar. Mamm. Sci. 17:35-75. Larkman, V. E., and R. R. Veit. 1998. Seasonality and abundance of blue whales off of Southern California. CalCOFl Rep. 39:236-239. Leatherwood, S., R. R. Reeves, W. F. Perrin, and W. E. Evans. 1982. Whales, dolphins and porpoises of the eastern North Pacific and adjacent Arctic waters: a guide to their identification. NOAA Tech. Rep. NMFS Circular 444, 245 p. Lux, C. A., A. S. Costa, and A. E. Dizon. 1997. Mitochondrial DNA population structure of the Pacific white-sided dolphin. Rep. Int. Whal. Comm. 47:645-652. Mangels, K. F., and T. Gerrodette. 1994. Report on cetacean sightings during a marine mammal survey in the eastern Pacific Ocean and the Gulf of California aboard the NOAA Ships McAr- thur and David Starr Jordan , July 28-November 6, 1993. NOAA Tech. Memo. NMFS-SWFSC-211, 88 p. Marques, F. F. C., and S. T. Buckland. 2003. Incorporating covariates into standard line tran- sect analyses. Biometrics 59:924-935. Mate, B. R., B. A. Lagerquist, and J. Calambokidis. 1999. Movements of North Pacific blue whales during the feeding season off southern California and southern fall migration. Mar. Mamm. Sci. 15:1246-1257. McDonald, M. A, J. A. Hildebrand, and S. M. Wiggins. 2006. Increases in deep ocean ambient noise in the northeast Pacific west of San Nicolas Island, Califor- nia. J. Acoust. Soc. Am. 120:711-8. Moore, J. E., and J. Barlow. 2011. Bayesian state-space models of fin whale abun- dance trends from a 1991-2008 time series of line-tran- sect surveys in the California Current. J. Appl. Ecol. 48:1195-1205. Morejohn, G. V. 1979. The natural history of Dali’s porpoise in the North Pacific Ocean. In Behavior of marine mammals, vol. 3: cetaceans. Current perspectives in research (H. E. Winn and B. L. Olla, eds.), p. 45-83. Plenum Press, New York. Munger, L. M., D. Camacho, A. Havron, G. Campbell, J. Calam- bokidis, A. Douglas, and J. Hildebrand. 2009. Baleen whale distribution relative to surface tem- perature and zooplankton abundance off southern Cali- fornia, 2004-2008. CalCOFl Rep. 50:155-168. 220 Fishery Bulletin 112(2-3) Ohman, M. D., and E. L. Venrick. 2003. CalCOFI in a changing ocean. Oceanography 16:76-85. Perrin, W. F., M. D. Scott, G. J. Walker, and V. L. Cass. 1985. Review of geographical stocks of tropical dolphins (Stenella spp. and Delphinus delphis ) in the eastern Pa- cific. NOAA Tech. Rep. NMFS 28, 28 p. Perrin, W. F., J. L. Thieleking, W. A. Walker, F. I. Archer, and K. M. Robertson. 2011. Common bottlenose dolphins (Tursiops truncatus ) in California waters: cranial differentiation of coastal and offshore ecotypes. Mar. Mamm. Sci. 27:769-792. Peterson, W. T., R. Emmet, R. Goericke, E. Venrick, A. Manty- la, S. J. Bograd, F. B. Schwing, R. Hewitt, N. Lo, W. Watson, J. Barlow, M. Lowry, S. Ralston, K. A. Forney, B. E. La- vaniegos, W. J. Sydeman, D. Hyrenbach, R. W. Bradely, P. Worzybok, F. Chavez, K. Hunter, S. Benson, M. Weise, and J. Harvey. 2006. The state of the California Current, 2005-2006: warm in the north, cool in the south. CalCOFI Rep. 47:30-74. Reeves, R. R., B. S. Stewart, P. J. Clapham, and J. A. Powell. 2002. Guide to marine mammals of the world, National Audibon Field Guide Series, 527 p. Alfred A. Knopf, New York. Reid, J. L., G. I. Roden, and J. G. Wyllie. 1958. Studies of the California Current system. Contri- butions from the Scripps Institution of Oceanography, New Series No. 998. In Calif. Coop. Ocean. Fish. In- vest. Prog. Rep., 1 July 1956 to 1 January 1958, p. 28- 56. [Available from http://www.calcofi.org/publications/ calcofireports/v06/CalCOFI_Rpt_Vol_06_ 1958. pdf.] Rice, D. W. 1974. Whales and whale research in the eastern North Pacific. In The whale problem: a status report (W. E. Schevill, ed.), p. 170-195. Harvard Press, Cambridge, MA. Rosel, P. E., A. E. Dizon, and J. E. Heyning. 1994. Genetic analysis of sympatric morphotypes of common dolphins (genus Delphinus). Mar. Biol. 119:159-167. Shane, S. H. 1994. Occurrence and habitat use of marine mam- mals at Santa Catalina Island, California from 1983- 1991. Bull. South. Calif. Acad. Sci. 93:13-29. Smith, R. C., P. Dustan, D. Au, K. S. Baker, and E. A. Dunlap. 1986. Distribution of cetacean and sea-surface chloro- phyll concentrations in the California Current. Mar. Biol. 91:385-402. Soldevilla, M. S., S. M. Wiggins, and J. A. Hildebrand. 2010. Spatial and temporal patterns of Risso’s dolphin echolocation in the Southern California Bight. J. Acoust. Soc. Am. 127:124-132. Soldevilla, M. S., S. M. Wiggins, J. Calambokidis, A. Douglas, E. M. Oleson, and J. A. Hildebrand. 2006. Marine mammal monitoring and habitat inves- tigations during CalCOFI surveys. CalCOFI Rep. 47:79-91. Soldevilla, M. S., S. M. Wiggins, J. A. Hildebrand, E. M. Ole- son, and M. C. Ferguson. 2011. Risso’s and Pacific white-sided dolphin habitat modeling from passive acoustic monitoring. Mar. Ecol. Prog. Ser. 423:247-260. Stacey, P. J., and R. W. Baird. 1991. Status of the false killer whale, Pseudorca crassidens, in Canada. Can. Field-Nat. 105:189-197. Stafford, K. M., S. L. Nieukirk, and C. G. Fox. 1999. An acoustic link between blue whales in the east- ern tropical Pacific and the northeast Pacific. Mar. Mamm. Sci. 15:1258-1268. Steiger, G. H., J. Calambokidis, R. Sears, K. C. Balcomb, and J. C. Cubbage. 1991. Movement of humpback whales between Califor- nia and Costa Rica. Mar. Mamm. Sci. 7:306-310. Thomas, L., S. T. Buckland, E. A. Rexstad, J. L. Laake, S. Strindberg, S. L. Hedley, J. R. B. Bishop, T. A. Marques, and K. P. Burnham. 2010. Distance software: design and analysis of distance sampling surveys for estimating population size. J. Appl. Ecol. 47:5-14. Thompson, S. A., W. J. Sydeman, J. A. Santora, B. A. Black, R. M. Suryan, J. Calambokidis, W. T. Peterson, and S. J. Bograd. 2012. Linking predators to seasonality of upwelling: us- ing food web indicators and path analysis to infer tro- phic connections. Prog. Oceanogr. 101:106-120. Urban R., J., A. Jaramillo L., A. Aguayo L., P. Ladron de Gue- vara P, M. Salinas Z., C. Alvarez F., L. Medrano G., J. K. Jacobsen, K. C. Balcomb, D. E. Claridge, J. Calambokidis, G. H. Steiger, J. M. Straley, O. von Ziegesar, J. M. Waite, S. Mizroch, M. E. Dahlheim, J. D. Darling, and C. S. Baker. 2000. Migratory destinations of humpback whales win- tering in the Mexican Pacific. J. Cetacean Res. Man- age. 2:101-110. Walker, W. A., S. Leatherwood, K. R. Goodrich, W. F. Perrin, and R. K. Stroud. 1986. Geographical variation and biology of the Pacific white-sided dolphin, Lagenorhynchus obliquidens, in the north-eastern Pacific. In Research on dolphins (M. M. Bryden and R. Harrison, eds.), p. 441-465. Claren- don Press, Oxford, UK. Weilgart, L. S. 2007. The impacts of anthropogenic ocean noise on ceta- ceans and implications for management. Can. J. Zool. 85:1091-1116. Wyrtki, K. 1964. Upwelling in the Costa Rica Dome. Fish. Bull. 63:355-372. 221 Abstract— Our study goal was to characterize the demographics of the population of Hickory Shad (Alosa mediocris ) in the Albemarle Sound-Roanoke River watershed during a period of population in- crease and to assess its susceptibil- ity to harvest. Adults were collected from gillnet surveys and a river rec- reational fishery from February to May 1996. The male-to-female ratio was similar between the Albemarle Sound (0.73:1) and the spawning grounds in Roanoke River (0.76:1). Ages were 2-7 years, but most sam- pled fish were age 3 or 4. The von Bertalanffy growth equation was Lt = 460 (1 - e~°-24u + L63)), where Lt was predicted length at time t for sexes combined. Total mortality ( Z ) was 1.43 for males age 3-5, 1.76 for females age 4-6, and 1.40 for sexes combined. Sexual maturity in both sexes was essentially complete by age 4. Repeat spawning was com- mon: 46.8% of males were virgin, 45.5% had spawned once, and 7.7% had spawned 2 or 3 times. For fe- males, 24.9% were virgin, 45.5% had spawned once, and 29.6% showed ev- idence of spawning 2, 3, or 4 times. Mesentery fat in both sexes de- creased from the prespawning aggre- gation (staging) area in the sound to the river spawning grounds, indicat- ing that both sexes feed extensively in ocean waters before the inland portion of the spawning migration. The short lifespan of Hickory Shad, combined with an early age to ma- turity and an anadromous migration pattern, indicates that mature indi- viduals are very susceptible to recre- ational and commercial harvest and are removed by exploitation or natu- ral mortality within 1 or 2 seasons. Manuscript submitted 31 January 2014. Manuscript accepted 27 May 2014. Fish. Bull. 112:221-236 (2014). doi:10.7755/FB.112.2-3.8 The views and opinions expressed or implied in this article are those of the author (or authors) and do not necessarily reflect the position of the National Marine Fisheries Service, NOAA. Population demographics of Hickory Shad (Alosa mediocris ) during a period of population growth Roger A. Rulifson (contact author) Christopher F. Batsavage Email address for the contact author: rulifsonr@ecu.edu Institute for Coastal Science and Policy, and Department of Biology East Carolina University Greenville, North Carolina 27858 Hickory Shad ( Alosa mediocris) is 1 of 4 anadromous Alosa species native to the East Coast of North America. The other species are American Shad (A. sapidissima) and the river her- rings Blueback Herring (A. aestiva- lis) and Alewife (A. pseudoharengus) . Often, Hickory Shad is confused with American Shad, and they commonly appear together in local fish mar- kets. Hickory Shad ranges from Cape Cod, Massachusetts, to the St. Johns River, Florida (Robins at ah, 1986) and there is no evidence of spawning populations north of Maryland (Rich- kus and DiNardo1). It is assumed that this species returns to natal streams to spawn as does American Shad (Melvin et ah, 1986), but homing has not been documented. Hickory Shad typically are 30-45 cm in fork length (FL) and 0. 5-1.0 kg in weight — size ranges that are intermediate between the larger American Shad and small- er river herrings (Robins et ah, 1986). The center of abundance for Hick- ory Shad is thought to be in North Carolina because historically the North Carolina commercial fishery landed the greatest number of Hick- ory Shad among the fisheries along the U.S. eastern seaboard (Atlantic States Marine Fisheries Commis- 1 Richkus, W. A., and G. DiNardo. 1984. Current status and biological character- istics of the anadromous alosid stocks of the eastern United States: American shad, hickory shad, alewife, and blue- back herring, 248 p. Atlantic States Marine Fisheries Commission, Washing- ton, D.C. sion [ASMFC]-). In 1902, the Hick- ory Shad harvest from North Caro- lina through Florida was 351,970 kg (775,962 lb) and worth $37,709 (ap- proximately $900,000 in 2011 dol- lars); North Carolina fisheries landed 88.3% of the total and represented 90.0% of the dockside value (Alex- ander, 1905). By 2001, the species was an incidental catch in various North Carolina gillnet fisheries in Albemarle and Pamlico sounds and in the coastal Atlantic Ocean (North Carolina Division of Marine Fisher- ies [NCDMFp). Hickory Shad also were landed from pound nets, haul seines, and the nearshore ocean win- ter trawl fishery (Street et al.2 3 4). 2 ASMFC (Atlantic States Marine Fisher- ies Commission). 1999. Amendment 1 to the interstate fishery management plan for shad and river herring. Fish- ery Management Report No. 35, 76 p. ASMFC, Washington, D.C. [Available from http://www.asmfc.org/uploads/file/ shadaml.pdf] 3 NCDMF (North Carolina Division of Marine Fisheries). 2001. Assessment of North Carolina commercial finfisher- ies, 1997-2000. Final performance re- port for Award Number NA 76 FI 0286, segments 1-3, 365 p. [Available from Division of Marine Fisheries, North Carolina Department of Environment and Natural Resources, 3441 Arendell St., Morehead City, NC 28557.] 4 Street, M. W., R P. Pate, B. F. Holland Jr., and A. B. Powell. 1975. Anadromous fisheries research program, northern coastal region. North Carolina. Final report for Project AFCS-8, 210 p. Divi- sion of Marine Fisheries, North Carolina Department of Natural and Economic Resources, Morehead City, NC. 222 Fishery Bulletin 112(2-3) Figure 1 Commercial harvest of Hickory Shad (Alosa mediocris) and American Shad (Alosa sapidissima) in North Carolina for the period of 1972-2010. The study was conducted after a period of population growth for Hickory Shad from 1990 to 1996. Baseline data for the 20th century are those from 1890 (when landings of Hickory Shad were 104,780 kg [231,000 lb]) to 1902 (when land- ings had increased to 310,711 kg [685,000 lb] [Alexander, 1905]). During the 1990s, Hickory Shad populations in- creased, whereas populations of the other 3 Alosa spe- cies of the eastern seaboard decreased (Rulifson, 1994; Waldman and Limburg, 2003; Watkinson, 2004). This trend was evident in the commercial landings data for North Carolina (Fig. 1). Federal and state landings data for shads are sometimes difficult to interpret because often Hickory Shad are not separated from landings of American Shad. However, personnel of state fisheries agencies and recreational fishermen have noted these increases through much improved springtime opportu- nities (e.g., catches and abundance) in the recreational fishery throughout the range of the Hickory Shad. In North Carolina and other mid-Atlantic states, sportfish- ing for Hickory Shad is now common during February, March, and April, when adults ascend rivers to spawn before the other 3 Alosa species; this shad is also popu- lar as a secondary target in the spring sport fisheries for White Perch ( Morone americana ) and Striped Bass (M. saxatilis). The Roanoke River watershed just down- stream of the last dam at Roanoke Rapids, North Caro- lina, and the tributary Cashie River near the town of Windsor are popular areas for Hickory Shad sportfish- ing (Fig. 2). Angler harvest in the Roanoke River wa- tershed increased from a 1968 estimate of 143 Hickory Shad caught by rod and reel and 2377 fish caught by special devices, such as dip nets and gill nets (Baker* * * 5), 5 Baker, W. D. 1968. A reconnaissance of anadromous fish to a 1996 estimate of 58,621 fish caught by hook and line that did not include the significant harvest by bank anglers (Kornegay6 7). In 1996, concerns about over- harvesting caused the North Caro- lina Wildlife Resources Commission (NCWRC) to classify Hickory Shad as a game fish in inland waters (Fig. 1). Since then, the bag limit has been 10 shads in aggregate per day (but only 1 American Shad) in inland, estuarine, and coastal wa- ters (ASMFC2). Subsequently, the recreational fishery for Hickory Shad and Striped Bass in the Roa- noke River has turned into a mul- timillion-dollar activity (McCargo et al. ). In 2006, anglers expended 14,065 hours (standard error of the mean [SE] 11,589), primarily in March and April. An estimated 81% of the shad were released; the remainder was harvested, but only 1.4% of that remainder were Amer- ican Shad — indicating the impor- tance of Hickory Shad to the sport fishery. Similar trends have been observed in the nearby Neuse River watershed, which has supported a long-standing shad sport fishery (Marshall, 1977; Hawkins8; Manooch, 1984). A comprehensive review of Hickory Shad popula- tions in South Atlantic coastal states was conducted by Rulifson et al. (1982), who documented that many of the life history aspects of this species were un- known. Since then, life history aspects of the Hickory Shad have been studied in Virginia rivers by Watkin- son (2004), the Roanoke River by Batsavage (1997) and Harris and Hightower (2010, 2011), the Tar- Pamlico River by Smith (2006), Murauskas (2006), and Murauskas and Rulifson (2009, 2011), and the Neuse River by Burdick and Hightower (2006), al- runs into the inland fishing waters of North Carolina. Final report for Project AFS-3, 38 p. [Available from Division of Inland Fisheries, North Carolina Wildlife Resources Commis- sion, 1751 Varsity Dr., Raleigh, NC 27606.] 6 Kornegay, J. W. 1996. Unpubl. data. North Carolina Wildlife Resources Commission, 1751 Varsity Dr., Raleigh, NC 27606. 7 McCargo, J. W., K. J. Doekendorf, and C. D. Thomas. 2007. Roanoke River recreational angling survey, 2005-2006. Final report. Coastal Fisheries Investigations, Federal Aid in Fish Restoration Project F-22, 67 p. [Available from Division of Inland Fisheries, North Carolina Wildlife Resource Commis- sion, 1751 Varsity Dr., Raleigh, NC 27606.] 8 Hawkins, J. H. 1980. Investigations of anadromous fishes of the Neuse River, North Carolina. Special Scientific Report. No. 34, 111 p. Division of Marine Fisheries, North Carolina Department of Natural Resources and Community Develop- ment, Morehead City, NC. Rulifson and Batsavage: Population demographics of Alosc mediocris 223 Figure 2 Map of the lower Roanoke River watershed and Albemarle Sound in North Carolina, showing the general locations where adult Hickory Shad ( Alosa mediocris) were collected during Feb- ruary-May 1996 from 2 independent gillnet surveys in the western end of Albemarle Sound, in the Roanoke River National Wildlife Refuge (RRNWR; indicated with dotted rectangle), and in the recreational fishery on the spawning grounds of Hickory Shad near the city of Weldon, North Carolina. though none except Batsavage (1997) were focused on age and growth. At the southern end of its range in Florida, the St. Johns River population was studied early by Walberg (1960) and Williams and Bruger9. In North Carolina, no directed sampling by state agen- cies has been conducted since 1993, but the NCWRC has collected Hickory Shad data for the 4 major North Carolina coastal rivers (Roanoke, Tar- Pamlico, Neuse, and Cape Fear) between 2000 and 2010 with annual monitoring (Dockendorf10 * *). Understanding key aspects of the life history, as well as the stock status of individual populations, is critical for species management. The ASMFC has long- identified life history aspects and the stock status of Hickory Shad as priorities for future research (Richkus and DiNardo1; ASMFC1112-13). 9 Williams, R. O., and G. E. Bruger. 1972. Investigations on American shad in the St. Johns River. Technical Series No. 66, 49 p. Florida Department of Natural Resources, St. Petersburg, FL. [Available from http://research.myfwc.com/ publications/publication_info.asp?id=29934.] 10Dockendorf, K. 2013. Personal commun. North Carolina Wildlife Resources Commission, Raleigh, NC 27606. nASMFC (Atlantic States Marine Fisheries Commission). 1985. Fishery management plan for the anadromous alosid stocks of the eastern United States: American shad, hickory shad, alewife, and blueback herring. Phase II in interstate management planning for migratory alosids of the Atlantic The goal of our study was to characterize the de- mographics of Hickory Shad during a known period of stock rebuilding with the Albemarle Sound-Roanoke River watershed as the focus population because of the important commercial and recreational fisheries there that target Hickory Shad. We describe the age, size, sex ratio, fecundity, age to maturity, growth, and mortal- ity of adult Hickory Shad in the spring prespawning population in Albemarle Sound and during the spawn- ing run near the spawning region in the Roanoke River near Weldon, North Carolina. Results of this study pro- vide important life history information for future man- agement plan development. coast. Fisheries Management Report No. 6, 347 p. ASMFC, Washington, D.C. [Available from http://www.asmfc.org/up- loads/file/1985FMP.pdf.] 12ASMFC (Atlantic States Marine Fisheries Commission). 2007. American shad stock assessment report for peer re- view, vol. 3. Stock Assessment Report No. 07-01 (Supple- ment), 489 p. ASMFC, Washington, D.C. [Available from http://www.asmfc.org/uploads/file/2007ShadStockAssmtRepo rtVolumeIII.pdf.) 13ASMFC (Atlantic States Marine Fisheries Commis- sion). 2009. Review of the Atlantic States Marine Fish- eries Commission fishery management plan for shad and river herring (Alosa spp.) 2009, 11 p. ASMFC, Washing- ton, D.C. [Available from http://www.asmfc.org/uploads/ file/2009ShadFMPReview.pdf.] 224 Fishery Bulletin 112(2-3) Materials and methods Study area Albemarle Sound is an extensive estuarine habitat in northeastern North Carolina, measuring 88.5 km long east to west and 4.8-22.5 km wide north to south (Street et al.4; Fig. 2). Known spawning populations of Hickory Shad are located in 3 of the 15 tributaries — the Roanoke, Cashie, and Chowan Rivers — all of which are situated at the extreme western end of Albemarle Sound. The estuary is relatively shallow, with depths ranging from 5.5 to 7.6 m, and is bordered by cypress swamps and small sand beaches. The sound is essen- tially freshwater through its western and central por- tions and brackish in the eastern part. Access to the Atlantic Ocean is at Oregon Inlet between Bodie and Hatteras islands, which are parts of the Outer Banks barrier island system. The Roanoke River is the largest tributary to Albemarle Sound in terms of freshwater input. Only the last 220.5 km of the river are accessi- ble to anadromous fishes; upriver portions are blocked by a series of impoundments ending with the Roanoke Rapids Reservoir upstream from Weldon (Rulifson and Manooch, 1991). The coastal plain portion of the wa- tershed downstream of the last dam has an extensive floodplain that consists of hardwood forest, backwater swamps, oxbow lakes, and small creeks (Zincone and Rulifson, 1991), which are connected to the river by natural and anthropogenic openings in the natural river levee (Walsh et al., 2005). Field collection Adult Hickory Shad were collected during 2 indepen- dent gillnet surveys and the Roanoke River recreation- al fishery. Albemarle Sound and its tributaries were sampled in the NCDMF Independent Gill Net Survey of Albemarle Sound Striped Bass from 19 February to 1 May 1996. Anchored, experimental gill nets in both floating and sinking configurations were 36.6 m long and constructed of monofilament with stretched mesh sizes ranging from 64 to 178 mm in 13-mm increments; additional nets of 203-mm and 254-mm stretched mesh were also used (Dilday and Winslow14). The lower Roa- noke River at the Roanoke River National Wildlife Ref- uge (RRNWR; Fig. 2) was sampled during an indepen- dent gillnet survey conducted by personnel from the National Marine Fisheries Service and RRNWR from 30 March to 17 April 1996. The single-mesh gill nets ranged from 3.6 m long and 1.5 m deep to 12.2 m long and 2.3 m deep; stretched mesh sizes ranged from 63 mm to 76 mm (Settle et al., 1996). During the spawn- ing run, fish from the sport fishery at Weldon were 14Dilday, J. L., and S. E. Winslow. 2000. North Carolina striped bass monitoring. Annual report, Grant F-56, seg- ment 7, 43 p. [Available from Division of Marine Fisheries, North Carolina Department of Environment and Natural Re- sources, 3441 Arendell St., Morehead City, NC 28557.] obtained at access points (primarily boat ramps) and examined fresh; fish from the gillnet surveys were fro- zen and transported to the laboratory for examination. Each fish was measured for both FL and total length (TL) in millimeters and weighed to the nearest gram. Gonads were removed from the fresh fish and weighed to the nearest gram, and ovaries were preserved in 10% cold buffered formalin for later examination. Chi- square tests were used to determine significant differ- ences in adult sex compositions between the 3 collec- tion sites, and regression analyses were used to estab- lish length-weight relationships. Age analysis Both scales and sagittal otoliths were used for aging adult Hickory Shad. From the left side above the lat- eral line and below the dorsal fin, 10-20 scales were removed. Scales were soaked in soapy water to remove dirt, mucus, and residual pigment and then dried. For examination under a microfiche reader equipped with a 24x lens, scales were mounted between 2 glass slides. Whole otoliths were removed, then aged by placing each in a watch glass containing distilled water and viewed under a dissecting microscope at 30x magnifi- cation. Otoliths were not sectioned for aging because their thin nature allowed their rings to be visible on their external portions. Both scales and otoliths were aged by 3 independent readers; each determination was considered successful when either the scale or otolith ages of at least 2 read- ers agreed. For scale aging, the traditional techniques and criteria of Gating (1953), Judy (1961), Street and Adams15, and Pate (1972) were used. Otolith aging techniques used criteria by Kornegay (1977) and Libby (1985). Results for fish aged with both scales and oto- liths were used to determine agreement between the 2 aging methods. Otoliths were used to back calculate growth because erosion of scale margins during the spawning migra- tion precludes the necessary relationship between fish length and scale radius (DeVries and Frie, 1996). To determine the relationship of otolith radius to FL, we used 75 fish, of which all fish <250 mm FL and nearly all fish >350 mm FL; 8 of those larger fish had otoliths that were unreadable. The 2 dominant length classes (250-299 and 300-349 mm FL) were subsampled to minimize bias associated with the effect that dominant size classes can have on linear regression calculations. Otolith images were measured on a video screen con- nected to a dissecting microscope at 16x power; otolith annuli were measured vertically from the nucleus to the ventral margin. 15Street, M. W., and J. G. Adams. 1969. Aging of hickory shad and blueback herring in Georgia by the scale method. Contribution Series No. 18, 13 p. Marine Fisheries Division, Georgia Game and Fish Commission, Brunswick, GA. Rulifson and Batsavage: Population demographics of Alosa mediocris 225 Two separate methods were used to estimate length at age. FL at age was estimated from the von Berta- lanffy growth equation (Cailliet et al., 1986), which was calculated with mean back-calculated FL at age (sexes combined). Back calculations also were computed by the Dahl-Lea direct proportion method (DeVries and Frie, 1996) with the following equation: L[ = (S[ / Sc) Lc, (1) where Lj = back-calculated FL( mm) of the fish at forma- tion of the ith increment; Lc = FL (mm) at capture; Sc = otolith radius at capture; and S; = otolith radius at the ith increment. Mortality estimates Mortality was estimated for ages where recruitment was more than 95% complete (on the basis of catch curves) — for males ages 3-5, females ages 4-6, and sex- es combined ages 3-6 to eliminate age classes not fully recruited to the spawning population. Total instanta- neous mortality ( Z ) was calculated by least-squares re- gression, and by estimating the slope of the line from the catch curve of a single season. Annual total mortal- ity (A) was estimated by taking the inverse natural log of -Z and subtracting the value from 1 (Ricker, 1975): A = 1 - e-z. (2) Spawning history Spawning history for both sexes was determined by counting the number of spawning marks on the scales; spawning marks are formed by erosion of the scale margin from lack of feeding during the spawning mi- gration (Cating, 1953; Pate, 1972). Spawning marks are thicker and more visible than the winter annuli that form before a fish matures sexually. The presence of these marks on scales is indicative of repeat spawners in a population, and the percentage of repeat spawn- ers can be calculated. The percentage of the population that was sexually mature was calculated by sex by di- viding the number of fish with developed gonads by the total number of fish examined. Fecundity and gonadosomatic index When this research was completed in 1996, gonad- al maturity and fecundity were not well understood. We understand now that the Hickory Shad is a batch spawner (Murauskas and Rulifson, 2011), a character- istic that requires special consideration in the estima- tion of fecundity (Olney et al., 2001; Murua and Sabo- rido-Rey, 2003). However, in 1996, sexual maturity was assigned by visual inspection of the gonads. We present the gonadosomatic index (GSI) here as documentation and for comparison with other limited studies of this species in other watersheds. For sexually mature indi- viduals, the number of ova present in each ovary was estimated by the gravimetric method. Each preserved whole ovary was blotted with a wet paper towel and then weighed to the nearest 0.01 g. Three subsamples of ovarian tissue, each -0.50 g, were taken from each ovary at the anterior, medial, and posterior regions. Each subsample was weighed, the ova were counted, and the number of ova per grain of ovarian tissue was calculated. The mean number of ova present in the 3 subsamples was multiplied by ovary weight to estimate the total number of ova in that ovary; the sum of the ova in the 2 ovaries was consid- ered the estimate of total potential fecundity. The GSI was estimated by dividing the gonad weight by somatic body weight (no gonads or gastrointestinal tract), and then multiplying the quotient by 100. The mean GSI was calculated by week so that temporal changes in the GSI could be identified. Two-sample t-tests with assumed unequal variances were used to detect dif- ferences between left and right ovaries in weight and total counts of ova. To detect significant differences in the number of ova per gram of ovarian tissue between the 3 regions of the ovary, analysis of variance was performed with Statistical Analysis System16 software. Regression analysis was used to predict potential fe- cundity on the basis of FL, somatic weight, and age. Significance was assigned an alpha (a) value of 0.05. Mesentery fat and gut content analyses The few literature references available on this topic indicate that Hickory Shad usually do not feed dur- ing the spring spawning migration (White and Curtis17; Curtis18; Perkins and Dahlberg, 1971; Pate, 1972; Har- ris et al., 2007). However, we observed Hickory Shad in the Roanoke River (during this study) and Neuse River (Murauskas and Rulifson, 2011) with full stomachs, and Harris et al. (2007) found similar trends in the St. Johns River population. To determine whether feeding or fat reserves were more important during the phase of inland spawning migration, we removed mesentery fat from the viscera and weighed it to the nearest 0.01 g. Food items removed from the stomach and intestine were identified to the lowest practical taxon, enumer- ated, and weighed to the nearest 0.01 g. T-tests were used to test for significant differences in mesentery fat between males and females and between fish collected in Albemarle Sound and fish collected in the Roanoke River. Relationships between mesentery fat and somat- 16Mention of trade names or commercial companies is for iden- tification purposes only and does not imply endorsement by the National Marine Fisheries Service, NOAA. 1 'White, M. G., Ill, and T. A. Curtis. 1969. Anadromous fish survey of the Black River and Pee Dee River watersheds. Project AFS-2-4, 73 p. South Carolina Wildlife Resources Department, Charleston, SC. 18Curtis, T. A. 1970. Anadromous fish survey of the Ashley River watershed. Project AFS-2, 91 p. South Carolina Wild- life Resources Department, Charleston, SC. 226 Fishery Bulletin 112(2-3) Figure 3 Length-frequency distributions for male (black bars) and female (white bars) Hick- ory Shad ( Alosa mediocris) collected from the Albemarle Sound-Roanoke River watershed in North Carolina during the spawning run in the spring of 1996. Fre- quencies are expressed as percentages, and fork lengths are delineated in 10-mm increments. ic weight and between mesentery fat and gonad weight were estimated by regression analyses. Males: loge BWT = 3.09(loge FL) - 11.76; and Females: loge BWT = 2.94(loge FL) - 10.80. (3) (4) Results Sex ratios, lengths, and age The male-to-female ratio of 0.73:l(n=266) for Albemar- le Sound was similar (%2=0.064, n= 532, df=l, P>0.05) to the ratio for the spawning grounds on the Roanoke River at Weldon (0.76:1, n= 266). Although there were more female Hickory Shad collected in the Albemarle Sound than on the spawning grounds in the river, the sex ratios were similar at both areas; however, in the RRNWR, the sex ratio was significantly skewed to- ward males. The independent gillnet survey conducted in the RRNWR was biased by the gear characteristics used (e.g., mesh sizes), yielding a male-to-female ra- tio of 4.29:1 (n = lll) — a value significantly different from that of the other 2 sites (x2=54.28, n=643, df=2, PcO.OOl). The mean size and range of lengths were larger for females than for males. Females were 280- 402 mm FL, and males were 257-376 mm FL. Domi- nant size classes (10-mm increments) were 330-339 and 340-349 mm FL for females (41.5%) and 280-289 and 290-299 mm FL for males (47.3%) (Fig. 3). Body weight, or Loge body weight (BWT) measured in grams, increased with length, or Loge FL measured in millimeters, for both males (coefficient of determi- nation [r2] =0.78) and females (7-2=0.73). The following equations were used to calculate these relationships: Because gonad weight varied considerably for both sexes, the length-weight relationship was calculated for somatic weight (loge SWT), improving the linear fit. for males (r2=0.81) and females (r2=0.76). The following equations were used to determine these relationships: Males: loge SWT = 3.01(loge FL) - 11.34; and (5) Females: loge SWT = 2.78(loge FL) - 9.98. (6) As expected, the relationship between FL and TL was highly correlated (r2=0.99). The following equation was used to establish this relationship: Loge TL = 0.99(loge FL) + 0.19. (7) Ages determined from scales and otoliths showed only a 57% agreement (n=478 pairs). On the basis of results of similar otolith-scale comparisons reported for Hickory Shad (Murauskas, 2006) and for other spe- cies (e.g., Kornegay, 1977; Paramore, 1998; Paramore and Rulifson, 2001), we assumed that otolith ages were accurate. With that assumption, scales generally overestimated the age of younger fish and underesti- mated the age of older fish (Table 1). However, scale and otolith ages never differed by more than 2 years for any given fish. There was no agreement between age-2 scales and otoliths. Agreement of age-3 scales and otoliths was 62%, age-4 scales and otoliths had a 61% agreement, and only 26% of scales age 5 and older agreed (Table 1). Rulifson and Batsavage: Population demographics of A/osa mediocns 227 Table 1 Percent agreement of ages, measured in years, from scales and otoliths of Hickory Shad ( Alosa mediocris ) collected in Albemarle Sound from February to May 1996. Data represent percent agreement calculated with otolith age as the standard; bold numbers indicate one-to-one cor- respondence (100% accuracy). Scale age Total number of Otolith age 2 3 4 5 6 7 otoliths examined 2 0 65 35 0 0 0 20 3 4 62 30 4 0 0 242 4 0 23 61 15 1 0 192 5 0 15 55 25 5 0 20 6 0 0 0 0 50 50 2 7 0 0 0 0 50 50 2 Total number of scales examined 10 209 208 44 5 2 478 Because of the deciduous nature of Alosa scales, oto- liths were used for age composition analyses, mortal- ity estimates, and back calculations of length at age. Scales that agreed in age with their respective otoliths were used for spawning mark analysis. The Hickory Shad population in the Albemarle Sound and Roanoke River ranged from age 2 to age 7, but the majority were age-3 and age-4 fish (Table 2). Age-3 males were dominant (66%), and the majority of females (55%) were age 4. Females were larger at age than were males, but size ranges (Table 2) and weights (Table 3) at age for each sex showed some degree of overlap. There was a strong relationship between otolith ra- dius and FL. The following equations were used to ex- press this relationship: Males: FL = 133.32 (otolith radius) - 62.35, (r2 [coefficient of determination] =0.95); (8) Females: FL = 116.54 (otolith radius) - 31.18, (r2=0.92); and (9) Sexes combined: FL = 117 ,02(otolith radius) - 29.20, (r2=0. 93). (10) The von Bertalanffy growth equation was Lt = 460 (1 - e-°-24(' + 163)). (11) In general, the mean back calculations of length at age were shorter than the observed lengths for younger fish and longer than observed lengths for older fish (Ta- ble 2). With the proportional method, back-calculated lengths for male Hickory Shad of ages 2-4 were shorter Table 2 Observed means and ranges of fork lengths at age, measured in millimeters and years, respectively, and back-calculated estimates of fork lengths at age of male and female Hickory Shad ( Alosa mediocris) collected from the Albemarle Sound-Roa- noke River watershed in 1996. Sexes combined represent the predicted size-at-age from the von Bertalanffy growth function (VBGR). Standard deviations (SD) are provided in parentheses. n= number of fish sampled. Age Males Females Sexes combined (VBGR) n Mean (SD) Range 1 Calculated n Mean (SD) Range Calculated 1 0 206 0 212 215 2 16 293 (9.3) 278-314 247 9 304 (7.0) 292-313 263 268 3 178 288 (12.9) 257-328 287 78 313 (18.4) 280-360 306 309 4 69 319 (11.9) 283-354 293 133 339 (15.3) 296-390 345 341 5 4 332 (16.4) 318-355 355 18 343 (18.8) 320-397 363 366 6 1 376 1 402 402 386 7 0 2 397 (4.2) 394-400 394 402 Total 268 241 228 Fishery Bulletin 112(2-3) Table 3 Observed means and ranges of body and somatic weights at age, measured in grams and years, respectively, of male and female Hickory Shad (Alosa mediocris) collected from the Roanoke River-Albemarle Sound watershed in 1996. Standard deviations (SD) are provided in parentheses. n=number of fish sampled. Age n Males Females Body weight Somatic weight Body weight Somatic weight Mean Range Mean Range n Mean Range Mean Range 2 16 330 (41.7) 273-411 310 (35.8) 256-388 9 391 (27.3) 358-446 343 (15.8) 325-379 3 178 319 (54.1) 210-548 300 (57.8) 197-525 78 440 (85.4) 291-839 390 (71.1) 280-612 4 69 451 (70.2) 316-698 422 (59.8) 297-640 133 591 (101.1) 359-839 505 (83.2) 318-705 5 4 452 (65.2) 403-542 430 (69.6) 385-532 18 639 (113.9) 447-908 542 (84.6) 417-710 6 1 651 638 1 1031 871 7 2 946 (192.0) 810-1082 779 (145.4) 676-881 Total 268 241 and lengths for age-5 males were longer than observed values. Females were similar to males except for age- 7 fish, which had back-calculated lengths that were shorter than observed values. Predicted FL values from the von Bertalanffy growth equation were less than the observed lengths for age-2 fish and greater than the observed lengths for fish of ages 5-7. Predicted lengths for age-3 and age-4 fish fell between the mean observed lengths for age-3 and age-4 males and females (Table 2). Females were larger and heavier at age than males (Tables 2 and 3). Mortality, maturity, and fecundity Mortality estimates were lower for males than for fe- males. Total instantaneous mortality (Z) was 1.43 for males of ages 3-5, 1.76 for females of ages 4-6, and 1.40 for both sexes combined. Annual total mortality (A) was 0.76 for males, 0.83 for females, and 0.75 for both sexes combined. Between 36% and 38% of both male and female Hickory Shad were sexually mature by age 2, most (>93%) were mature by age 3, and al- most all were mature by age 4 (Table 4). Virgin males represented 46.8% of the male population; an addition- al 45.5% had spawned once, and 7.7% had spawned at least 2 or more times (Table 4). No males exhibited more than 3 spawning marks. Virgin females composed only about one-fourth (24.9%) of the sample, 45.5% of females had spawned once before, and 29.1% of them showed evidence of spawning 2 or 3 times. One age-7 female had 4 spawning marks (Table 4). Slowly increasing trends in the mean GSI were ob- served for Hickory Shad from both Albemarle Sound and Roanoke River through March, whereas mean GSI slowly decreased through the week of April 7-13 and then decreased quickly thereafter (Fig. 4). The mean number of ova per gram of ovarian tissue ranged from more than 1500 to less than 4000, and the anterior por- tions of both ovaries tended to have higher ova counts per gram of ovarian tissue than the posterior region. This relationship was significant for the left ovary (n=47, F=4.68, P=0.011) but not for the right ovary (F=1.21, P=0.303). The left ovary was significantly greater in weight and mean total egg counts than the right ovary (n = 186, £=3.686, P<0.001). Mean left ovary weight was 42.88 g, and mean right ovary weight was 35.98 g. Mean left egg count was 111,037, and the right ovary contained an average of 93,630 ova. These means were not significantly different for the left and right ovaries (n=47, £=-1.746, P=0.840). Potential fecundity (PF) of female Hickory Shad generally increased with fish length, body weight, somatic weight, and age class (Table 5). Fecundity estimates ranged from 80,290 to 478,944 ova (n=47). We used the following prediction equations: Log,, PF = 3.90(loge FL) - 10.46 (r2=0.63); (12) Loge PF = 1.33(loge BWT) + 3.70 (r2=0.76); (13) Loge PF = 1.39(loge SWT ) + 3.59 (r2=0.67); and (14) Loge PF = 0.30(Age) + 10.97 (r2=0.52) (15) Feeding and mesentery fat Fish collected from the Roanoke River had significant- ly less mesentery fat reserves than Albemarle Sound fish (males: n=6 2, £=-3.050, P=0.005; females: n=110, £=-4.54, P<0.0001). Also, male fish from the Roanoke River had significantly less remaining fat than Roa- noke River females (n=98, £=-2.140, P=0.030), but this sex difference was not observed for fish collected in Al- Rulifson and Batsavage: Population demographics of Alosa mediocris 229 Table 4 Proportion (%) of mature fish by sex, determined from visual inspection of gonads, and number of spawning marks by age class for male and female Hickory Shad ( Alosa mediocris ) collected from the Roanoke River-Albemarle Sound watershed in 1996. n=number of fish examined. Age Males Females Percent mature Number of spawning marks n Percent mature Number of spawning marks n 0 1 2 3 0 1 2 3 4 2 36.1 12 12 38.5 7 7 3 97.9 92 56 148 93.9 38 24 62 4 99.6 4 50 14 68 98.6 6 69 48 123 5 100 1 0 1 2 4 100 2 4 9 3 18 6 100 0 0 0 1 1 100 0 0 0 1 0 1 7 100 0 0 1 0 1 2 n examined 233 109 106 15 3 233 213 53 97 58 4 1 213 Proportion (%) of total population 46.8 45.5 6.4 1.3 24.9 45.5 27.2 1.9 0.5 bemarle Sound (n= 74, f=-1.570, P=-0.120), indicating that both sexes feed extensively in ocean waters be- fore entering the phase of inland spawning migration. There were weak positive relationships between somat- ic weight and mesentery fat for both males (r2=0.17) and females (r2=0.29) in the Albemarle Sound, but these relationships essentially disappeared (males: r2=0.14; females: r2=0.02) on the spawning grounds in the Roanoke River (Fig. 5). Interestingly, a similar set of relationships was ob- served between gonad weight and mesentery fat in males and females (males: P2=0.46; females: P2=0.21) in the Albemarle Sound — relationships that disap- peared (males: P2=0.20; females: f?2=0.05) on the spawning grounds (Fig. 6), indicating that fish were using mesentery fat during their upsti’eam migration for metabolic energy and not for increasing gonad size. For gut analysis, the stomachs of 212 fish were ex- amined. Of the fish collected from the Albemarle Sound and Roanoke River, 26% (62) and 28% (110), respec- tively, contained identifiable items. In fish from both locations, 83% of stomach items found fitted into 5 categories: fish (Clupeidae), parasites (Isopoda), seeds, wood, and plastic. Insects, a sixth category, were found only in stomachs of Roanoke River fish. Discussion Adult sex ratios The sex ratios in our study indicated that there was no sex-selective harvest by anglers in 1996. The male-to- female ratios from near the Roanoke River spawning grounds at Weldon (0.76:1) indicated that slightly more females than males were sampled. A similar result (0.73:1) was obtained from the NCDMF Independent Gill Net Survey of Striped Bass in the Albemarle Sound , but the independent gillnet survey in the RRNWR se- lected for male fish (4.29:1). The RRNWR survey used small mesh sizes, causing bias toward smaller males; the gillnet survey in Albemarle Sound for Striped Bass used a wide range of mesh sizes to allow for collection of the full size range of both sexes. In some cases, males were more abundant than fe- males because a greater proportion of males reach ma- turity at an earlier age; moreover, differential arrival periods of males and females on the spawning grounds, as in the Chesapeake Bay, can affect the sex ratios found in samples (Klauda et al. 1991a). Pate (1972) found the male-to-female ratio to be 4:1 for Hickory Shad sampled by a nonselective haul seine in the Neu- se River, North Carolina. This ratio could have been the result of recruitment of a large proportion of virgin males to the spawning population (47.3% of the males were age 2). Skewness of true sex ratios from increased mortal- ity of a targeted sex likely plays a role in population rebuilding. For Alosa species, females are obviously the limiting factor, and older age classes have greater reproductive potential. Higher rates of survival to re- peat-spawning age and a sex ratio closer to 1:1 should lead to accelerated population rebuilding, as opposed to the rebuilding potential of a population with far more males and virgin spawners. Some alosine fisheries (e.g., American Shad) target females for their roe (Rulifson et al., 1982), and such activity will shift the true sex ratio. 230 Fishery Bulletin 1 12(2-3) 30 25 - 20 - M 15 -I O 1 5 - 30 -| CO 15 O 13 10 - r Lr r t I Figure 4 The mean (dashed line), median (solid line), 5-95% percentiles (box), 10th and 90th percentiles (whiskers), and outliers (dots) of gonadosomatic index (GSI) values for Hickory Shad (Alosa mediocris) collected during the pre- spawning and spawning periods during February-May 1996: (A) females and (B) males sampled from the Albemarle Sound as part of the North Carolina Divison of Marine Fisheries Independent Gill Net Survey and (C) females and (D) males from the Roanoke River sampled as part of a gillnet survey in the Roanoke River National Wildlife Refuge and from a recreational fishery at sites near the city of Weldon, North Carolina. Age analysis, otolith back calculations, and mortality estimates The 57% agreement between scale and otolith ages in our study in the Albemarle Sound-Roanoke River wa- tershed is similar to results reported by Harris et al. (2007), who found a 57.3% agreement for fish from St. Johns River; 96% of the ages were in agreement of one year. Our study found no more than 2 years disagree- ment for any given fish. Kornegay (1977) reported a similar agreement for Alewife from Albemarle Sound, but for Blueback Herring his agreement was approxi- mately 68%. Kornegay’s (1977) Alewife scale ages never deviated by more than 2 years from otolith ages, but, for 2 Blueback Herring, scale ages deviated by up to 3 years from otolith ages. A difference of 1 or 2 years between scale age and otolith age is a relatively large de- viation for a fish with a lifespan of only 7 or 8 years. Likewise, the agreement level of 57% between scale and otolith ages is low. Alosa scales are commonly regenerated, and spawning marks sometimes obscure annuli near the scale mar- gin. The first annulus is sometimes confused with the freshwater zone, which is a false annulus formed when juvenile Alosa first enter the marine environment, and the first annulus is not always visible on the scale (Cating, 1953; Judy, 1961; Kornegay, 1977). In addition, the Hickory Shad, among Alosa spe- cies, has scales considered to be the most difficult to use for age analysis (Richkus and DiNardo1). Therefore, otoliths should be used whenever possible for aging Hickory Shad. The few published studies on age composition of Hickory Shad and other Alosa species generally show 1-3 dominant year classes (Street and Adams15; Pate, 1972; Street et al.4; Kornegay, 1977; Winslow19’20; NCDMF21; Harris et ah, 2007). In our study, ages 3 and 4 were the dominant age classes, with male Hickory Shad contributing the ma- jority of the younger age classes (ages 2 and 3) and female Hickory Shad contributing the majority of the older age classes (ages 4-7) (Ta- ble 2). For the adjacent Neuse Riv- er, Murauskas and Rulifson (2011) reported that both sexes averaged 3 years of age, although a larger pro- portion of females were in older age classes: 25% of females were of ages 4, 5, and 6, whereas 14% of males were of age 4 only. Overlapping lengths at age made it difficult to ac- curately determine age structure from length fre- 19Winslow, S. E. 1989. North Carolina alosid fisheries man- agement program. Completion report for Project AFC-27, 102 p. Division of Marine Fisheries, North Carolina Depart- ment of Natural Resources and Community Development, Morehead City, NC. 20Winslow, S. E. 1990. Status of American Shad, Alosa sapi- dissima (Wilson), in North Carolina. Completion report for Job 5, Project AFC-27, 94 p. + appendix. Division of Marine Fisheries, North Carolina Department of Natural Resources and Community Development, Morehead City, NC. 21NCDMF. 2001. North Carolina shad and river herring compliance report, 2000, 66 p. [Available from Division of Marine Fisheries, North Carolina Department of Environ- ment and Natural Resources, 3441 Arendell St., Morehead City, NC 28557.] Rulifson and Batsavage: Population demographics of Alosa mediocris 231 Table 5 Number of ova, estimated gravimetrically and from regressions developed for age, in ovaries of female Hickory Shad (Alosa mediocris ), collected in Albemarle Sound and the Roanoke River in 1996, and mean values of the following variables: fork length (FL) measured in millimeters, body weight (BWT) measured in grams, and somatic weight (SWT) measured in grams at age, measured in years. «=number of ovaries examined. Age n Gravimetric counts (observed) Counts on basis of age Counts by mean FL at age Counts by mean BWT at age Counts by mean SWT at age 2 i 85,803 108,012 138,195 113,366 121,109 3 14 137,523 147,267 154,849 132,642 144,776 4 19 223,576 200,787 211,380 196,382 207,344 5 3 294,798 273,758 221,275 217,874 228,758 6 1 478,944 373,249 410,929 411,646 442,326 7 2 250,918 508,897 391,352 367,134 378,751 quencies. Other studies on Hickory Shad and other Alosa species also showed a significant degree of overlap of lengths at age (Street and Adams15; Pate, 1972; NCD- MF21). Our results indicate that the mean FLs at age for both sexes from age 3 and older were smaller than the means reported from ear- lier investigations. This dif- ference could be the result of 1) the capture method used by previous investiga- tors, who collected Hickory Shad in gill nets with large mesh sizes that were set for American Shad (Street et al.4; Hawkins8), 2) the scales used for determining age, or 3) both. Pate (1972) ex- amined Hickory Shad captured in a nonselective haul seine and found that the largest Hickory Shad of both sexes were the oldest fish sampled (ages 5-7). The mean FLs from back calculations with both the Dahl-Lea direct proportion method and the von Berta- lanffy growth equation indicate that the smaller age-2 Hickory Shad were not part of the spawning migra- tion. The largest differences between mean observed FLs and mean back-calculated FLs from the von Ber- talanffy growth equation occurred at age 2 (Table 2). The only age-2 fish sampled were the ones that were sexually mature. Analysis of spawning marks showed that approximately 41% of age-2 Hickory Shad (sexes combined) were mature at this age, indicating that the majority of age-2 Hickory Shad were not sampled. It is presumed that age-1 fish and many age-2 fish remain at sea, but this portion of the lifecycle is unknown. Annual mortality rates were higher in this study (0.75, sexes combined) than in previous studies in Al- bemarle Sound; rates from other studies ranged from 0.40 to 0.65 (Street et al.4; Johnson et al.22). However, it should be noted that the annual mortality in those studies was calculated with the Robson and Chapman (1961) method that computes survival instead of us- ing the catch curve to estimate mortality. Our mortal- ity rates were higher for females (0.83) than for males (0.76). Because the catch curve was generated from just 1 year of data, it is difficult to determine the cause for higher mortality estimates. Moderate fluctuations 22 Johnson, H. B„ D. W. Crocker, B. F. Holland Jr., J. W. Gil- liken, D. L. Taylor, M. W. Street, J. G. Loesch, W. H. Rrete Jr., and J. G. Travelstead. 1978. Biology and management of mid-Atlantic anadromous fishes under extended jurisdiction. Completion Report AFCS-9-2, 175 p. Division of Marine Fisheries, North Carolina Department of Natural Resources and Community Development, Morehead City, NC, and Vir- ginia Institute of Marine Science, Gloucester Point, VA. in annual recruitment are common in fish populations, but catch curves derived from 2 or more years of data can reduce the effects of variable recruitment (Ricker, 1975). Spawning history and reproductive analysis The short lifespan of Hickory Shad, combined with an early age to maturity and an anadromous migra- tion pattern, indicates that adult fish in the Albemarle Sound-Roanoke River population are subject to rec- reational (sound and inland waters) and commercial harvest (sound and ocean waters) for 1 or 2 seasons before they are removed by harvest or natural mortal- ity. Approximately one-third of both sexes are sexually mature as early as age 2, >93% of the population is mature by age 3, and essentially 100% of the popula- tion is mature by age 4 (Table 4). One or 2 spawning marks on the scales examined were common, but 3 or more marks were rare. These results were similar to findings for Hickory Shad in the Altamaha River, Geor- gia (Street, 1970), the Neuse River (Pate, 1972), and more recently the Tar- Pamlico River (Murauskas and Rulifson, 2011). On the basis of age to maturity and spawning pat- terns, Hickory Shad and American Shad are exploited similarly in the Albemarle Sound region, but the level of exploitation for these species differs south of Cape Hatteras. American Shad in Albemarle Sound usually reach sexual maturity by age 3 or age 4 for males and by age 4 or age 5 for females. Both sexes spawned up to 3 times (Winslow19’20). American Shad show a lati- tudinal gradient between semelparity and iteroparity throughout its range (Leggett and Carscadden, 1978). In contrast, individual American Shad in populations south of Cape Hatteras seldom spawn more than once, and adults in populations in New York and Connecticut 232 Fishery Bulletin 112(2-3) Figure 5 Relationship between mesentery fat and somatic weight, both mea- sured in grams, of (A) female and (B) male Hickory Shad ( Alosa me- diocris) collected during February-May 1996 in Albemarle Sound and on the Roanoke River spawning grounds near Weldon, North Carolina. r2=coefficient of determination; f?2=coefficient of multiple determination. spawn up to 5 times. Hickory Shad appear to be iter- oparous south of Cape Hatteras as indicated by repeat spawners in the Neuse River (Pate, 1972; Hawkins8), Altamaha River (Street, 1970), and St. Johns River (Harris et al., 2007). Because of the short life spans and limited number of spawning events (i.e., repeat spawning) exhibited by Hickory Shad and other Alosa species, successive years of poor recruitment could re- sult in relatively quick population declines. Therefore, our estimate of 0.75 for annual total mortality, sexes combined, is possible. State landings data for Hickory Shad after 1996 indicate that such a mortality esti- mate may have been real (Fig. 1). Landings stabilized in the 2000s decade. Mean GSI values were similar between fish caught in the Albemarle Sound and fish captured in the Roa- noke River. The spawning season for Hickory Shad in the Albemarle Sound-Roanoke River watershed lasts for about 4 weeks in March and April; therefore, female Hickory Shad will be at differing de- grees of gonadal development (i.e., pre- spawning, running ripe, partially spent, postspawning) for any given week dur- ing the spawning season, and this vari- ation will result in a large variability in GSI values (Fig. 4). Murauskas and Rulifson (2011) observed multiple batch spawning of Hickory Shad in the Tar- Pamlico River watershed over several weeks, and these events were related to water temperature. This observation is supported by our ovary data, which revealed significantly different states of maturity between anterior and poste- rior oocytes in the ovaries. Fecundity estimates are important in population modeling and also for hatch- ery managers who attempt to spawn and rear Hickory Shad for the purpose of stock restoration. The Maryland De- partment of Natural Resources has been rearing and stocking larval and early juveniles of both Hickory Shad and American Shad in at least 6 Chesa- peake Bay watersheds and tributaries (Richardson et al.23). Although Olney et al. (2001) and Murauskas and Rulifson (2011) classified both species as batch spawners, Maryland hatchery person- nel do not mention this aspect in their methodology. Both males and females received an intramuscular implant of leutinizing-hormone-releasing hormone analog (LHRHa) in the dorsal muscu- lature at the collection site and were returned to the hatchery for spawning. Very little information exists on fecun- dity estimates of Hickory Shad, and es- timates include age-2 fish. Pate (1972) reported 44,556 to 347,610 eggs per female from the Neuse River. Hickory Shad in the Altamaha River showed increased fecundity with age and size, with estimates ranging from 252,693 to 730,213 eggs per female and a mean of 500,519 eggs per female (Street, 1970). St. Johns River females exhibited low correlation between fe- cundity and weight, length, and age. Fecundity ranged from 168,000 to 591,000 eggs per female with a mean of 363,000 eggs per female (Williams and Bruger9). Our study of the Roanoke River population (egg counts from 23Richardson, B. M., C. P. Stence, M. W. Baldwin, and C. P. Mason. 2009. Hickory shad restoration in three Maryland rivers. F-57 Segment 9 Progress Report, 48 p. Maryland Department of Natural Resources, Oxford, MD. [Available from http://www.dnr.state.md.us/irc/docs/00014544.pdfl Rulifson and Batsavage: Population demographics of Alosa mediocris 233 80,290 to 478,944) found good correla- tion of egg counts with weight, length, and age (Table 5). It appears that suitable spawn- ing habitats differ among watersheds. Spawning activity of Hickory Shad in the Neuse River, North Carolina, and the Altamaha River, Georgia, was con- fined primarily to flooded bottomlands and tributaries away from the main stem of each river (Street, 1970; Pate, 1972; Burdick and Hightower, 2006). Smith (2006) noted Hickory Shad spawning in small tributaries of the Tar-Pamlico River watershed. Mansueti (1962) found Hickory Shad spawning in the main stem of the Patuxent River, Maryland, upstream of American Shad spawning sites. Hickory Shad have been found to spawn in both the main stem and tributaries of rivers in Virginia (Klauda et ah, 1991b). In the Roanoke River upstream of our study area, Har- ris and Hightower (2011) conducted a study of spawning habitat for Hickory Shad, and they determined that adults generally avoided spawning in areas with very low (<0.1 m/s) or no water velocity, especially when substrate sizes were small. When water velocities were low (<0.1 m/s), spawning occurred only on bedrock substrates. When water ve- locities were higher (>0.1 m/s), spawning occurred on a variety of substrate types, including gravel and occasionally sand. Although we did not survey spawning habitat of Hickory Shad in the Roanoke River in 1996, both ripe and partially spent adults were collected from tribu- taries of the Roanoke River in the RRN- WR and at Weldon. The higher flows of the Roanoke River in the spring flood the backwater tributaries and swamps; therefore, maintenance of a flow regime similar to the natural springtime flows probably is needed to ensure suitable spawning habitat for Hickory Shad in this watershed. Why spawning runs of Hickory Shad in the 1990s far exceeded the spawning runs of American Shad in the Roanoke River-Albemarle Sound watershed remains a mystery. Historically, American Shad dominated har- vests of anadromous shad in every major watershed in mid-Atlantic and South Atlantic states. Near the turn of the 20th century in 1890, North Carolina landings of American Shad totaled 2.616 million kg (5.768 million lb), increasing to 4.066 million kg (8.963 million lb) in 1897 and dropping to 2.979 million kg (6.567 million lb) in 1902 (Alexander, 1905). Hickory Shad landings were only 104,780 kg (231,000 lb) in 1897 and 310,711 kg (685,000 lb) in 1902. At the end of the 20th century (in 1996, the year of our study), commercial harvests of both species were nearly equal: 90,554 kg (199,638 lb) of American Shad and 85,244 kg (187,887 lb) of Hickory Shad (NCDMF database, http://portal.ncdenr. org/web/mf/statistics/comstat; Fig. 1). Interpretation of landings beyond 1996 becomes more difficult because the species was declared a game fish in inland wa- ters, and harvest restrictions were subsequently put in place for the recreational fishery. How this designation of a game fish may have affected commercial harvest is unknown. The Roanoke River fishery, once dominated by thriv- ing commercial fisheries that targeted anadromous species American Shad, Alewife, Blueback Herring, and 234 Fishery Bulletin 112(2-3) Striped Bass, is now a multimillion-dollar recreational fishery best known for Striped Bass and Hickory Shad (McCargo et al.7). Habitat loss and fragmentation, along with overharvesting the species, are considered major factors in the reduction of alosine stocks to rem- nant populations in this watershed (Walsh et ah, 2005; McCargo et al.7) and elsewhere in the North Atlan- tic (Limburg and Waldman, 2009). Restoration of the American Shad population in the Roanoke River has been ongoing since 1988 (Waters24), but adult abun- dance remains low despite the stocking of 43 million American Shad fry in the Roanoke River as of 2010 (Dockendorf10). Populations of Hickory Shad in upper Chesapeake Bay tributaries are experiencing resurgence and are supporting an active catch-and-release recreational fishery. This resurgence also means better access to brood fish for hatchery programs, and the state of Maryland now has implemented stock restoration ef- forts for Hickory Shad in 3 rivers: the Patuxent, Choptank, and Nanticoke (Richardson et al.23). Mary- land agencies hope to establish increased fishing op- portunities for targeting Hickory Shad, believing that restoration of this species has the potential to occur over a shorter time frame (because of its earlier age at maturity) than the period needed for American Shad restoration (Richardson et al.23). Conclusions Our findings clearly indicate that the short lifes- pan of the Hickory Shad, combined with an early age to maturity and an anadromous migration pat- tern, means that adult individuals of the population will be subjected to recreational and commercial har- vest in inland waters for 1 or 2 seasons before they are removed by exploitation or natural mortality. Our data were collected before the implementation of the 10-fish bag limit on shads. North Carolina fisheries agencies hope that a daily 10-fish limit for shads (only 1 fish can be American Shad in the Roanoke River) will protect current population size while maintain- ing the interest of fishermen in this lucrative fishery. The study presented here is the most recent on this species for North Carolina; data collected during creel surveys by the NCWRC have included only recorded catches but not samples for lengths, weights, or age. We recommend that new data be collected on age and growth since this regulation went into effect to deter- mine whether incidences of repeat spawning events have increased in this population. This growing popu- lation has a sex ratio slightly dominated by females both in the prespawning staging area in Albemarle 24Waters, C. T. 2000. Summary of activities in 1998 and 1999 for restoring American Shad to Roanoke River. [Avail- able from North Carolina Wildlife Resources Commission, 1751 Varsity Dr., Raleigh, NC 27606.] Sound in January and on the spawning grounds in the Roanoke River. Continued research on the poorly un- derstood life history of this species will increase our understanding and, perhaps, provide insight on its success in relative abundance compared with that of American Shad. Acknowledgments We thank the staff of the North Carolina Division of Marine Fisheries, especially H. Johnson, S. Trowell, S. Winslow, and field technicians of the Elizabeth City of- fice for their assistance in fish collections; the National Marine Fisheries Service (Beaufort Laboratory, South- east Fisheries Science Center) sampling crew headed by D. Peters; and the staff of the Roanoke River Na- tional Wildlife Refuge led by J. Holloman. P. Kornegay of the North Carolina Wildlife Resources Commission provided the 1996 Hickory Shad recreational harvest data for the Roanoke River. C. Manooch III, and J. Potts of the Beaufort Laboratory assisted with otolith preparation and reading. We thank K. Dockendorf, W. Patrick, J. Murauskas, A. Dell’Apa, C. Bangley, and the anonymous referees for critical review of the manu- script. Funding was provided, in part, by the North Carolina Fishery Resource Grant program (through the NC Marine Fisheries Commission), Project No. M-6057, to R. Rulifson. Literature cited Alexander, A. B. 1905. Statistics of the fisheries of the South Atlantic states, 1902, 67 p. Division of Statistics and Methods of the Fisheries, U.S. Fish Commission. GPO, Washing- ton, D. C. Batsavage, C. F. 1997. Life history aspects of the hickory shad (Alosa me- diocris) in the Albemarle Sound/Roanoke River water- shed, North Carolina. M.S. thesis, 100 p. East Caro- lina Univ., Greenville, NC. Burdick, S. M., and J. E. Hightower. 2006. Distribution of spawning activity by anadromous fishes in an Atlantic Slope drainage after removal of a low-head dam. Trans. Am. Fish. Soc. 135:1290-1300. Cailliet, G. M., M. S. Love, and A. W. Ebeling. 1986. Fishes: a field and laboratory manual on their structure, identification, and natural history, 194 p. Wadsworth Publ., Belmont, CA. Gating, J. P. 1953. Determining age of Atlantic shad from their scales. Fish. Bull. 54:187-199. DeVries, D. R., and R. V. Frie. 1996. Determination of age and growth. In Fisheries techniques, 2nd ed. (B. R. Murphy and D. W. Willis, eds.), p 483-508. Am. Fish. Soc., Bethesda, MD. Harris, J. E., R. S. McBride, and R. O. Williams. 2007. Life history of hickory shad in the St. Johns River, Florida. Trans. Am. Fish. Soc. 136:1463-1471. Rulifson and Batsavage: Population demographics of Alosa mediocris 235 Harris, J. E., and J. E. Hightower. 2010. Evaluation of methods for identifying spawning sites and habitat selection for alosines. N. Am. J. Fish. Manage. 30:386-399. 2011. Spawning habitat selection of hickory shad. N. Am. J. Fish. Manage. 31:495-505. Judy, M. H. 1961. Validity of age determination from scales of marked American shad. Fish. Bull. 61:161-170. Klauda, R. J., S. A. Fischer, L. W. Hall, and J. A. Sullivan. 1991a. Alewife and blueback herring. In Habitat re- quirements for Chesapeake Bay living resources, 2nd ed. (S. L. Funderburk, J. A. Mihursky, S. J. Jordan, and I). Riley, eds.), p 10-1-10-29. Chesapeake Research Consor- tium, Solomons, MD. 1991b. American Shad Alosa sapidissima and Hickory Shad Alosa mediocris. In Habitat requirements for Chesapeake Bay living resources, 2nd ed. (S. L. Funder- burk, J. A. Mihursky, S. J. Jordan, and D. Riley, eds.), p 9-1-9-27. Chesapeake Research Consortium, Solo- mons, MD. Kornegay, J. W. 1977. A comparison of the scale and otolith methods of ageing alewife ( Alosa pseudoharengus) and blueback herring ( Alosa aestivalis). M.S. thesis, 79 p. East Carolina Univ., Greenville, NC. Leggett, W. C., and J. E. Carscadden. 1978. Latitudinal variation in reproductive character- istics of American shad (Alosa sapidissima ): evidence for population specific life history strategies in fish. J. Fish. Res. Board Can. 35:1469-1478. Libby, D. A. 1985. A comparison of scale and otolith aging meth- ods for the alewife, Alosa pseudoharengus . Fish. Bull. 83:696-701. Limburg, K. E., and J. R. Waldman. 2009. Dramatic declines in North Atlantic diadromous fishes. Bioscience 59:955-965. Manooch, C. S., III. 1984. Fisherman’s guide to fishes of the southeastern United States, 362 p. North Carolina State Museum of Natural History, Raleigh, NC. Mansueti, R. J. 1962. Eggs, larvae, and young of the hickory shad, Alosa mediocris, with comments on its ecology in the estu- ary. Chesapeake Sci. 3:173-205. Marshall, M. D. 1977. Status of hickory shad in North Carolina. In Pro- ceedings of a workshop on American shad had (R. St. Pierre, ed.), p. 33-45. U.S. Fish and Wildlife Service and National Marine Fisheries Service, Amherst, MA. Melvin, G. D., M. J. Dadswell, and J. D. Martin. 1986. Fidelity of American shad, Alosa sapidissima (Clu- peidae), to its river of previous spawning. Can. J. Fish. Aquat. Sci. 43:640-646. Murauskas, J. G. 2006. Investigating the reproductive migration of adult hickory shad, Alosa mediocris. M.S. thesis, 121 p. East Carolina Univ., Greenville, NC. Murauskas, J. G., and R. A. Rulifson. 2009. A comprehensive approach to understanding diad- romy at the species level: learning from the spawning migration of hickory shad. In Challenges for diadro- mous fishes in a dynamic global environment (A. Haro, K. L Smith, R. A. Rulifson, C. M. Moffitt, R. J. Klauda, Michael J. Dadswell, R. A. Cunjak, J. E. Cooper, K. L. Beal, and T. S. Avery, eds), p. 837-840. Am. Fish. Soc. Symp. 69, Bethesda, MD. 2011. Reproductive development and related obser- vations during the spawning migration of hickory shad. Trans. Am. Fish. Soc. 140:1035-1048. Murua, H., and F. Saborido-Rey. 2003. Female reproductive strategies of marine fish spe- cies of the North Atlantic. J. Northwest Atl. Fish. Sci., 33:23-31. Olney, J. E., S. C. Denny, and J. M. Hoenig. 2001. Criteria for determining maturity stage in female Alosa sapidissima and the mystery of partial spawn- ing. Bull. Fr. Peche Piscic. 362/363:881-901. Paramore, L. M. 1998. Age, growth, and life history characteristics of striped bass, Morone saxatilis, from the Shubenacadie- Stewiacke River, Nova Scotia. M.S. thesis, 91 p. East Carolina Univ., Greenville, NC. Paramore, L M., and R. A. Rulifson. 2001. Dorsal coloration as an indicator of different life history patterns for striped bass within a single watershed of Atlantic Canada. Trans. Am. Fish. Soc. 130:663-674. Pate, P P 1972. Life history aspects of the hickory shad, Alosa mediocris (Mitchill), in the Neuse River, North Caro- lina. M.S. thesis, 66 p. North Carolina State Univ., Raleigh, NC. Perkins, R. J., and M. D. Dahlberg. 1971. Fat cycles and condition factors of Altamaha River shads. Ecology 52:359-362. Ricker, W. E. 1975. Computation and interpretation of biological sta- tistics of fish populations. Bull. Fish. Res. Board Can. 191, 382 p. Robins, C. R., G. C. Ray, and J. Douglass. 1986. A field guide to Atlantic Coast fishes of North America, 354 p. Houghton Mifflin, Boston, MA Robson, D. S., and D. G. Chapman. 1961. Catch curves and mortality rates. Trans. Am. Fish. Soc. 90:181-189. Rulifson, R. A. 1994. Status of anadromous Alosa along the East Coast of North America. In Anadromous Alosa symposium, Tidewater Chapter. (J. E. Cooper, R. T. Eades, R. J. Klau- da, and J. G. Loesch, eds.), p. 134-158. Am. Fish. Soc., Bethesda, MD Rulifson, R. A., and C. S. Manooch, III (eds.). 1991. Roanoke River Water Flow Committee report for 1990. NOAA Tech. Memo. NMFS-SEFC-291, 433 p. Rulifson, R. A., M. T. Huish, and R. W. Thoesen. 1982. Anadromous fish in the southeastern United States and recommendations for development of a man- agement plan, 525 p. U.S. Fish and Wildlife Service, Atlanta, GA. Smith, M. C. 2006. Habitat use of early Alosa spp. and striped bass, Mo- rone saxatilis, in the lower Tar River, North Carolina. M.S. thesis, 165 p. East Carolina Univ., Greenville, NC. Street, M. W. 1970. Some aspects of the life histories of hickory shad, Alosa mediocris (Mitchill), and blueback herring, (Alo- sa aestivalis ) (Mitchill), in the Altamaha River, Geor- gia. M.S. thesis, 89 p. Univ. Georgia, Athens, GA. 236 Fishery Bulletin 112(2-3) Walberg, C. H. 1960. Abundance and life history of the shad, St. Johns River, Florida. Fish. Bull. 60:487-501. Waldman, J. R., and K. E. Limburg. 2003. The world’s shads: summary of their status, conser- vation, and research needs. In Biodiversity, status, and conservation of the world’s shads (K. E. Limburg and J. R. Waldman, eds.), p. 363-369. Am. Fish. Soc. Symp. 35, Bethesda, MD. Walsh, H. J., L. R. Settle, and D. S. Peters. 2005. Early life history of blueback herring and alewife in the lower Roanoke River, North Carolina. Trans. Am. Fish. Soc. 134:910-926. Watkinson, E. R. 2004. Age, growth, and fecundity of hickory shad (Alosa mediocris) in a Virginia coastal river. M.S. thesis. Virginia Commonwealth Univ.. Richmond. VA. Zincone, L. H., and R. A. Rulifson. 1991. Instream flow and striped bass recruitment in the lower Roanoke River, North Carolina. Rivers 2:125-137. 237 Errata Fishery Bulletin 112:49-70. Munyandorero, Joseph In search of climate effects on Atlantic Croaker (Mi- cropogonias undulatus) stock off the U.S. Atlantic coast with Bayesian state-space biomass dynamic models Corrections: On page 56, Equation 6 is missing an equal sign in the calculation of the deviance statistic D. The equation should read as follows: D(e) = -21og L(0) = -2 log [P(0 | 0)] On page 70, the age symbol in Equation A2 should be a, instead of a (alpha). Equation A2 should read as follows: Fa=/3{L„[l-exp(-K(a-a0))]}y 238 Fishery Bulletin 112(2-3) National Marine Fisheries Service Best The award for best publication of the year is given to authors who are employees of the National Marine Fisheries Service and whose article is judged to be the most noteworthy of those published in Fishery Bulletin and Marine Fisheries Review. The selections were made by scientists at all NOAA Fisheries Science Centers and thus represent well-deserved recognition by peers. Authors from the National Marine Fisheries Service are noted in bold font. The winners for Fishery Bulletin Rose, Craig S., Carwyn F. Hammond, Allan W. Stoner, J. Eric Monk, and John R. Gauvin Quantification and reduction of unobserved mortality rates for snow, southern Tanner, and red king crabs ( Chionoecetes opilio, C. bairdi [ and Paralithodes camtschaticus ) after encoun- ters with trawls on the seafloor Fish. Bull. 111:42-53. This clearly written paper presents an excellent study of an important topic for ecosystem-based fisheries management. The problem of unobserved mortality of bycatch is a consequential yet unaccounted for aspect of numerous fishery operations and is of worldwide concern. The paper efficiently describes a well-thought- out and well-executed approach to quantification of un- observed and unintended mortalities of commercially valuable crab species that encounter bottom trawls tar- geting ground-fish in the Bering Sea. Care was taken in the design of the experiments and in the thorough- ness of the research and this paper provides a practi- cal, low-cost, and innovative method for reducing crab bycatch mortality rates while maintaining catch lev- els for target fish species. Moreover, the authors went beyond quantification of unobserved crab mortalities and incorporated the design and testing of modified trawl gear, which — if adopted by commercial fishing fleets — could mitigate bycatch mortalities. This kind of research can be tedious, but it is only through just such studies that a real understanding of the effects of trawling and the extent of unobserved mortality rates can be obtained. This work addresses a breadth of questions and provides practical management insights (based on a well-defined model) that could be expanded for the study and management of other species. Paper Awards for 2013 The winner for Marine Fisheries Review Castro, Jose Historical knowledge of sharks: ancient science, earliest Amer- ican encounters, and American science Mar. Fish. Rev. 74(4): 1-26. This impressive, particularly well-written and re- searched compendium of shark information demon- strates how sharks have interacted with humans over time and how our understanding and study of these often misunderstood creatures have progressed since antiquity. This paper is of wide interest and utility to those who study the biology of sharks and their use by humans. It is the result of a monumental undertaking that required numerous years of research to amass the knowledge and material that is presented. This type of effort is time consuming and sometimes underappre- ciated. The paper highlights the value of older docu- ments that are not available to any but the most schol- arly of scientists with avid interest in fisheries history. For example, rare pictures and engravings of sharks show progressively more sophisticated and detailed depictions through several centuries up to the present time. This paper will keep alive such past works that are difficult to find and allow them to still contribute to current scholarly works and research efforts. The detailed explanation of how sharks have been utilized over time and how fisheries operated in the past passes on information that is available to few. Dr. Castro, for example, has one of the only surviving documents from the Ocean Leather Company. His discussion of the his- tory of this company and how it is intertwined with past U.S. fisheries and research will enlighten even the most seasoned shark researcher. This article is truly an interesting and detailed work by one of the world’s foremost experts on sharks. 239 Fishery Bulletin Guidelines for authors Manuscript preparation Contributions published in Fishery Bulletin describe original research in marine fishery science, fishery en- gineering and economics, as well as the areas of ma- rine environmental and ecological sciences (including modeling). Preference will be given to manuscripts that examine processes and underlying patterns. Descriptive reports, surveys, and observational papers may occa- sionally be published but should appeal to an audience outside the locale in which the study was conducted. Although all contributions are subject to peer review, responsibility for the contents of papers rests upon the authors and not on the editor or publisher. Submission of an article implies that the article is original and is not being considered for publication elsewhere. Articles may range from relatively short contributions (10-15 typed, double-spaced pages [tables and figures not in- cluded]) to extensive contributions (20-30 typed pages). Manuscripts must be written in English; authors whose native language is not English are strongly advised to have their manuscripts checked by English-speaking colleagues before submission. Title page should include authors’ full names and mailing addresses and the senior author’s telephone, fax number, and e-mail address. Abstract should be limited to 250 words (one-half typed page), state the main scope of the research, and emphasize the authors conclusions and relevant findings. Do not review the methods of the study or list the contents of the paper. Because abstracts are circulated by abstracting agen- cies, it is important that they represent the research clearly and concisely. General text must be typed in 12-point Times New Roman font throughout. A brief introduction should convey the broad significance of the paper; the remain- der of the paper should be divided into the following sections: Materials and methods, Results, Discus- sion, Conclusions, and Acknowledgments. Headings within each section must be short, reflect a logical se- quence, and follow the rules of subdivision (i.e., there can be no subdivision without at least two subhead- ings). The entire text should be intelligible to interdisci- plinary readers; therefore, all acronyms, abbreviations, and technical terms should be written out in full the first time they are mentioned. For general style, follow the U.S. Government Print- ing Office Style Manual (2008) [available at http://www. gpoaccess.gov/stvlemanual/index.htmI] and Scientific Style and Format: the CSE Manual for Authors , Edi- tor's, and Publishers (2014, 8th ed.) published by the Council of Science Editors. For scientific nomenclature, use the current edition of the American Fisheries So- ciety’s Common and Scientific Names of Fishes from the United States, Canada, and Mexico and its compan- ion volumes (Decapod Crustaceans, Mollusks, Cnidaria and Ctenophora, and World Fishes Important to North Americans). For species not found in the above men- tioned AFS publications and for more recent changes in nomenclature, use the Integrated Taxonomic Informa- tion System (ITIS) (available at http://itis.gov/). or, sec- ondarily, the California Academy of Sciences Catalog of Fishes (available at http://researcharehive.ca!academv. org/research/ichthvologv/catalog/fishcatmain.asp) for species names not included in ITIS. Citations must be given of taxonomic references used for the identification of specimens. For example, “Fishes were identified by using Collette and Klein-MacPhee (2002); sponges were identified by using Stone et al. (2011).” Dates should be written as follows: 11 November 2000. Measurements should be expressed in metric units, e.g., 58 metric tons (t); if other units of measure- ment are used, please make this fact explicit to the reader. Use numerals, not words, to express whole and decimal numbers in the general text, tables, and fig- ure captions (except at the beginning of a sentence). For example: We considered 3 hypotheses. We collected 7 samples in this location. Use American spelling. Re- frain from using the shorthand slash ( / ), an ambiguous symbol, in the general text. Word usage and grammar that may be useful are the following: Aging For our journal the word aging is used to mean both age determination and the aging process (se- nescence). The author should make clear which meaning is intended where ambiguity may arise. Fish and fishes For papers on taxonomy and biodiver- sity, the plural of fish is fishes, by convention. In all other instances, the plural is fish. Examples: The fishes of Puget Sound [biodiversity is indicated]; The number of fish caught that season [no emphasis on biodiversity]; The fish were caught in trawl nets [no emphasis on biodiversity]. The same logic applies to the use of the words crab and crabs, squid and squids, etc. Sex For the meaning of male and female, use the word sex, not gender. Participles As adjectives, participles must modify a specific noun or pronoun and make sense with that noun or pronoun. Incorrect: Using the recruitment model, estimates of age-1 recruitment were determined. [Estimates did not use the recruitment model.] Correct: Using the recruitment model, we deter- mined age-1 estimates of recruitment. [The participle now modifies the word we, those who were using the model.] 240 Fishery Bulletin 112(2-3) Incorrect: Based on the collected data, we concluded that the mortality rate for these fish had increased. [We were not based on the col- lected data.] Correct: We concluded on the basis of the collected data that the mortality rate for these fish had increased. [Eliminate the participle and replace it with an adverbial phrase.] Equations and mathematical symbols should be set from a standard mathematical program (MathType) or tool (Equation Editor in MS Word). LaTex is accept- able for more advanced computations. For mathemati- cal symbols in the general text (a, x2, rc, ±, etc.), use the symbols provided by the MS Word program and itali- cize all variables. Do not use photo mode when creating these symbols in the general text. Number equations (if there are more than 1) for fu- ture reference by scientists; place the number within parentheses at the end of the first line of the equation. Round all values to 2 decimal points. Literature cited section comprises published works and those accepted for publication in peer- reviewed journals (in press). Follow the name and year system for citation format in the “Literature cited” section (that is to say, citations should be listed alphabetically by the authors’ last names, and then by year if there is more than one citation with the same authorship. A list of abbreviations for citing journal names can be found at http://oliver.ross.p.luminv.univ- amu.fr/iournal abbrevs/abbreva.htm Authors are responsible for the accuracy and com- pleteness of all citations. Literature citation format: Author (last name, followed by first-name initials). Year. Title of article. Abbreviated title of the journal in which it was published. Always include number of pages. If there is a sequence of citations in the text, list chrono- logically: (Smith, 1932: Green. 1947; Smith and Jones, 1985). Digital object identifier (doi) code ensures that a publication has a permanent location online. Doi code should be included at the end of citations of published literature. Cite all software and special equipment or chemical solutions used in the study within parenthe- ses in the text (e.g., SAS, vers. 6.03, SAS Inst., Inc., Cary, NC). Footnotes are used for all documents that have not been formally peer reviewed and for observations and communications. These types of references should he cited sparingly in manuscripts submitted to the journal. All reference documents, administrative reports, inter- nal reports, progress reports, project reports, contract reports, personal observations, personal communica- tions, unpublished data, manuscripts in review, and council meeting notes are footnoted in 9 pt font and placed at the bottom of the page on which they are first cited. Footnote format is the same as that for formal literature citations. A link to the online source (e.g., [http://www/ , accessed July 2007.]), or the mail- ing address of the agency or department holding the document, should be provided so that readers may ob- tain a copy of the document. Tables are often overused in scientific papers; it is seldom necessary or even desirable to present all the data associated with a study. Tables should not be ex- cessive in size and must be cited in numerical order in the text. Headings should be short but ample enough to allow the table to be intelligible on its own. All un- usual symbols must be explained in the table legend. Other incidental comments may be footnoted with italic numeral footnote markers. Use asterisks only to indi- cate significance in statistical data. Do not type table legends on a separate page; place them above the table data. Do not submit tables in photo mode. Figures must be cited in numerical order in the text. Graphics should aid in the comprehension of the text, but they should be limited to presenting patterns rather than raw data. Figures should not exceed one figure for every four pages of text. Figures must be la- beled with the number of the figure. Place labels A, B, C, etc. within the upper left area of graphs and photos. Avoid placing labels vertically (except for the y axis). Figure legends should explain all symbols and abbre- viations seen in the figure and should be double-spaced on a separate page at the end of the manuscript. Color is allowed in figures to show morphological differences among species (for species identification), to show stain reactions, and to show gradations in temperature con- tours within maps. Color is discouraged in graphs, and for the few instances where color may be allowed, the use of color will be determined by the Managing Editor. • Notate probability with a capital, italic P. • Provide a zero before all decimal points for values less than one (e.g., 0.07). • Capitalize the first letter of the first word in all la- bels within figures. • Do not use overly large font sizes in maps and for units of measurements along axes in figures. • Do not use bold fonts or bold lines in figures. • Do not place outline rules around graphs. • Use a comma in numbers of five digits or more (e.g., 13,000 but 3000). • Place a North arrow and label degrees latitude and longitude (e.g., 170°E) in maps. • Use symbols, shadings, or patterns (not clip art) in maps and graphs. Failure to follow these guidelines and failure to correspond with editors in a timely manner will delay publication of a manuscript. Copyright law does not apply to Fishery Bulletin, which falls within the public domain. However, if an author reproduces any part of an article from Fishery Bulletin in his or her work, reference to source is con- sidered correct form (e.g., Source: Fish. Bull. 97:105). Guidelines for authors 241 Submission Submit manuscript online at http://mc.manuscriptcentral. com/fishervbulletin. Commerce Department authors should submit papers under a completed NOAA Form 25-700. For further details on electronic submission, please contact the Associate Editor, Kathryn Dennis, at kathryn.dennis@noaa.gov When requested, the text and tables should be submit- ted in Word format. Figures should be sent as PDF files (preferred), Windows metafiles, TIFF files, or EPS files. Send a copy of figures in the original software if con- version to any of these formats yields a degraded ver- sion of the figure Questions? If you have questions regarding these guidelines, please contact the Managing Editor, Sharyn Matriotti, at sharyn.matriotti@noaa.gov Questions regarding manuscripts under review should be addressed to Kathryn Dennis, Associate Editor. Fishery Bulletin Subscription form Superintendent of Documents Publications Order Form *5178 I 1 YES, please send me the following publications: Subscriptions to Fishery Bulletin for $32.00 per year ($44.80 foreign) The total cost of my order is $ . Prices include regular domestic postage and handling and are subject to change. (Company or Personal Name) (Please type or print) (Additional address/attention line) (Street address) (City, State, ZIP Code) (Daytime phone including area code) (Purchase Order No.) Charge your order. IT’S EASY! Please Choose Method of Payment: ] Check payable to the Superintendent of Documents ] GPO Deposit Account -□ ] VISA or MasterCard Account To fax your orders (202) 512-2104 (Credit card expiration date) (Authorizing Signature) Mail to: Superintendent of Documents P.O. Box 979050, St. Louis, Missouri 63197 Tel. no. (U.S. callers): 866-512-1800 Tel. no. (Int. callers): 1-202-512-1800 Thank you for your order! Also available online at http://bookstore.gpo. gov/actions/GetPublication.do?stocknumber=703-023-00000-2 email: ContactCenter@gpo.gov