Movements and Mortality of American White Pelicans Fledged in Three Montana Colonies .:ri&-' .-. ^r'.,,. -'w "■■ ,. -■■':''r'r.'" t ;_ .4^ ' ■. ■■. ^^-'-Z - r!t'-,^ -;r.._-^b Prepared for: U.S. Fish and WUdlife Sei-vice Benton Lake National Wildlife Refuge Prepared by: Panl Hendricks and Robert F. Johnson Montana Natural Heiitage Program Natuial Resoufce Information System Montana State Library June 2002 MONTANA Natural Heritage Program Movements and Mortality of American White Pelicans Fledged in Three Montana Colonies Prepared for : U.S. Fish and WUdMe Sei-vice Benton Lake National WUdlife Rffuge 922 Bootlegger Trail Great FaDSjMT 59404-6133 Agreement #61510-0-0400 Prepared by: Panl Hendricks and Robert F. Jolinson j^" MONTANA ;)J Natural Heritage Program ^j^T^tate "§*JVi Natural Resource ^^ Library ^^ISs^ information System © 2002 Montana Natural Heritage Program P.O. Box 201800 • 1515 East Sixth Ave • Helena, MT 59620-1800 •406-444-3009 This document should be cited as follows: Hendricks, P., andR. F. Jolinson. 2002. Movements and mortality of American White Pehcans fledged in three Montana colonies. Report to the U.S. Fish and Wildlife Service. Montana Natural Heritage Program, Helena. 17 pp. Executive Summary Montana is home to a significant number of breeding American White Pehcans (Pefecatuts erythrorlmichos), a species of conservation concern with a global population breeding in fewei' than 70 colonies. The four breeding colonies in Montana have giown to approximately 10,000 breeding paiis within the last ten years. Banding of juveniles has occuired at each colony at one time or anothei\ yet band return data have never been examined to identify patterns in migratoiy movements and mortality tliatmay affect management. We analyzed data for 276 band recoveries of American White Pelican banded as juveniles at tliree Montana colonies east of tlie Continental Divide. The majority of non-Montana returns for tlie Bowdoin colony in far-eastern Montana indicated most pelicans migiated south and east to the Gulf of Mexico and eastern Mexico. Returns for tlie Canyon Feiry and Arod colonies near tlie Continental Divide showed a different migratory movement, with tlie majority of biids crossing tlie Divide and moving west and south to southern California and western Mexico. We found little evidence for movement of birds between Montana colonies, and only one case of a bird bom in one Montana colony attempting to breed in another. Overall mortality rates ranged from 30% for Canyon Feny and Bowdoin to 39% for Arod; botli Arod and Canyon Feny had adult mortality rates >35%, which are unusually high for diis species. Birds ''found dead" of unknown causes accounted for 39.0-53.0% of recoveries from all colonies. Shot birds accounted for 21.7 % of returns for Bowdoin, whereas disease (botulism in particular) was the attributed mortality factor for 47.5 % of Canyon Feiry returns and 34.3 % of Arod Lakes' returns. The difference in significant mortality factors among the colonies can be attributed largely to differences in years when banding occurred and diffeient colony migration patterns. Over 40% of the recoveries for tlie Arod and Canyon Feny colonies weie from the Salton Sea of California, which expeiienced a botulism outbreak in tlie late 1990 's. Botulism affected all age classes, whereas shooting was more likely to affect juvenile pelicans (first- and second-year birds). We conclude that tliere are two relatively distinct colony groupings of breeding pelicans in Montana: one (Canyon Feny and Arod Lakes colonies) using tlie Intermountain West and Pacific flyways, and tlie otlier (Bowdoin, with tlie Medicine Lake colony) tlie Central and Mississippi flyways. However, a coordinated and simultaneous color-marking program at all Montana colonies would better document migratoiy movements, areas of non-breeding concentiation, and the amount of inter-colony dispersal. Conservation and management of the Montana colonies needs to include consideration of the distinctive flyway affiliations of tlie colonies and threats to habitat in areas wheie birds concenti'ate during the non- breeding season, as well as direats to the individual breeding colonies diemselves. A regional pei'spective is also desirable to coordinate management among colonies demonstrating inteniiixing of the breed ing m enib er s hips. IV Acknowledgements Field work in 2000 and production of this report was funded by a Challenge Cost-Share agreement between the U.S. Fish and Wildlife Service (USFWS), and the Montana Natural Heritage Program (MTNHP), which is part of the Montana State Libraiy's Natural Resource hiformation System (NRIS). Banding at tlie Arod Lakes and Canyon Feiry colonies was largely the result of volunteer efforts initiated in 1991 by Dave Gentei' and later co-sup ems ed by Jim Reichel (botli formerly witli MTNHP) and conducted in cooperation with personnel at Benton Lake National Wildlife Refuge and Canyon Feiry Wildlife Management Area. We tliank the many volunteers involved in six years of banding at tliese two colonies, especially Cedron Jones, Margaret Beer, Melony Bruhn, Kathy Jurist, Debbie Dover, Mike Schwitters, Kiisti Dubois, Bert Lindler, Tom Carlsen, and Rick Nortlirup. Qui' apologies to those whose names we failed to mention. Return or colony-size data for tlie Canyon Feiry, Medicine Lake, and Bowdoin colonies were shared by Tom Carlson, Beth Madden, Dwain Prellwitz, and Jeff Marks, or provided by tlie U. S.D.I. Bird Banding Laboratoiy, Laurel, Maryland. Pam Ganettson (USFWS) made a heroic effort to obtain old return location data from Mexico for the Bowdoin colony. Tommy King (USDA National Wildlife Center, Mississippi State, Mississippi) provided the most recent global estimates on thenumber of breeding colonies. Cedron Jones and Whitney Weber (both MTNHP) produced tlie maps. This report benefited from the comments of Tom Carlsen (Canyon Ferry Wildlife Management Area), Dwain Prellwitz (Bowdoin National Wildlife Refuge), Betli Madden (Medicine Lake National Wildlife Refuge), John Carlson (MTNHP), Sue Crispin (MTNHP) and Joy Lewis (MTNHP). Terrie Kenney (MTNHP) provided comments and was invaluable in the final preparation of tliis report. Table of Contents Executive Summaiy iv Acknowledgements v Inti'oduction 1 Study Areas 1 Metliods 2 Results 6 Discussion 12 Literatuie Cited 15 Table 1. Band recovery locations for fledged American White Pelicans banded at three colonies in Montana. Colonies are listed from east to west 7 Table 2. Circumstances associated with liands recovered from fledged American White Pelicans banded at tliree Montana colonies. Colonies are listed east to west 9 Table 3. Age distiibution of recovered American White Pelicans banded as flightless young at tliree Montana colonies. Colonies are listed east to west. Year 1 is tlie first year following banding, and so on. "Band only" recoveries are not included in tlie calculations 9 Table 4. Distiibution of band recoveries by age (excluding ''band only" returns) for tlie combined Arod Lake and Canyon Feiry colonies of American White Pelican in Montana. Numbers in bold show the elevated numbei' of recoveries in 1996 for each age class, mostly from the Salton Sea, California... 13 Figure 1. Distribution of American White Pelican returns for the Bowdoin Lake colony, Phillips County, Montana. Five of twelve Mexico returns are not shown because specific location data weie not available (states of recovery are listed in Table (1) 3 Figure 2. Distribution of American White Pelican returns for tlie Canyon Feiry Reservoir colony, Broadwater County, Montana 4 Figure 3. Distribution of Ameiican White Pelican returns for the Arod Lakes colony, Teton County, Montana 5 Figure 4. Seasonal distribution of band returns for immature (A) and adult (B) American White Pelicans from Arod Lakes (open fill), Canyon Feiiy (right- diagonal fill), and Bowdoin (cross-hatched fill) colonies. Numbeis of band returns are indicated above tlie bars 11 VI Introduction The American White Pelican {Pelecanus erythrorhynchos) is widespread throughout much of western and southeastern North America, and is frequently encountered at lakes, reservoirs, rivers, and coastal regions, often in large numbers. Most breeding colonies in Canada and tlie United States lie west of tlie Mississippi River and north of 40° N latitude (Vemieer 1970, Sidle et al. 1985, Evans and Knopf 1993), witli a few isolated colonies in soutliem Texas and nortliem Mexico (Chapman 1988, Evans and Knopf 1993). The number of breeding birds for all colonies has exceeded 100,000 in recent decades (Sidle et al. 1985), witli an additional unknown number of n on -breeding individuals, but the global breeding population is concenti'ated now in fewer tlian 50 colonies (T. King personal communication). Pelicans disperse from colonies following breeding, eventually wintering in two general areas: in the West from central California south to Cential America (often along tlie Pacific Coast}, and in the East tliroughout the southern United States, along tlie Gulf of Mexico, and thioughout much of the interior of Mexico. Pelicans dispersing from colonies east of tlie Continental Divide migrate primarily southward and eastward to the Gulf of Mexico coast. Birds dispeising from colonies west of the Continental Divide migrate primarily to California and the western coast of Mexico (Behle 1958, Diem and Condon 1967, Stiait and Sloan 1975, Veimeer 1977, Ryder 1981, Evans and Knopf 1993, Keith and O^Neill 2000, King and Grewe2001). The few American White Pelican colonies in Montana weie small in tlie early decades of the 20^' Centuiy (Saunders 1921), and tlie state was not thought an important breeding ground for pelicans at the time of die first comprehensive status assessment (Thompson 1933). Following Thompson's survey, numbei's of breeding birds increased in Montana such tliat tlie state was a major breeding area by tlie time of die next range-wide status assessment in 1964 (Lies and Behle 1966). Montana has remained a significant breeding ground for American White Pelicans through the end of the 20 Century (Sloan 1982, Sidle etal. 1985). Over the last several decades a few smaller breeding colonies of American White Pelican have formed in eastern Montana, but most of these were abandoned after a year or two of activity. However, two newer colonies have now remained active for over ten continuous years. This study examined tliese two newer colonies in more detail, emphasizing dispersal/migration of banded fledglings, natal colony fidelity, and mortality of marked young. We compared attributes of tliese newer colonies with the older Montana colony at Bowdoin National Wildlife Refuge and considered implications for conservation and management of American White Pelican colonies in Montana. Study Areas Montana's four American White Pelican colonies are scatteied from near the North Dakota State line to tlie Rocky Mountain Front, just 60 miles east of tlie Divide. We compared data from the two western- most, recently established colonies at Arod Lakes and Canyon Feiry Reseivoir with data from the Bowdoin colony in eastern Montana. Analyses of returns for Medicine Lake, the eastern-most colony in Montana (in Sheridan County), will be presented elsewhere (B. Madden, personal communication), and are discussed in this report only in general tenns. For migratoiy patterns and mortality factors, comments pertaining to tlie Bowdoin colony apply generally to Medicine Lake as well. Arod (or Eyraut^ Lakes: This colony in Teton County (47° 59' N, 112° OT W) was discovered in 1990, when 300 nests weie estimated to be present. Since tlien, nest estimates have varied from 300-500 annually, but no routine counts have been conducted. Annual numbers of young fledged at diis colony has varied dramatically from about 60 in 2000 to over 700 in 1993. Between 1991-1996, 2220 young weie banded, but no additional banding has occuned at diis colony since 1996. Ctaiy on Ferry Reservoir: This colony in Broadwater County (46° 24' N, 111'' 30' W) was discoveied in 1989 when 13 nests were counted. It grew dramatically since tlien to 1804 nests in 2000. Banding also occmred at tliis colony during 1991-1996, witli 1537 young banded. As witli tlie Arod colony, no banding has occmred at Canyon Ferry since 1996. Bowdoin Lake: This colony, on a National Wildlife Refuge in Phillips County (48° 24' N, 107° 41' W), has been known since 1903 when George Willett (1907) reported a small nesting colony on one of the islands in tlie lake. At least 336 nests weie present by 1935 (Weydemeyer and Marsh 1936). Estimates of nest numbers have generally increased since then (Lies and Behle 1966, Sidle et al. 1985), and during the 1990's ranged from 1552 nests in 1999 to 2882 nests in 1993. Banding has occurred intermittently at Bowdoin: 1932-1934, 1936-1940, 1942-1943, 1946-1947, and 1958. No banding has been undertaken at tliis colony in recent decades. Methods We obtained band recovery data for the Bowdoin, Canyon Feny, and Arod Lakes colonies from the Bird Banding Laboratoiy (BBL) in Laurel, Maryland. Additional reports of banded birds from Arod Lakes and Canyon Feny Resei'voir were made directly to tlie Montana Natural Heritage Program (MTNHP), which supervised the banding at tliese two colonies. Our tabular analyses included only recoveiies reported through the BBL. We have noted the additional MTNHP reports in tlie results, however, to supplement BBL data and address specific points. Distribution maps of band retums were generated from latitude-longitude data (10' blocks) provided by tlie BBL. Mexico retums for tlie Bowdoin colony weie not poitiayed in tlie distiibution map (Figure 1) because data were not provided in latitude- longitude format. Retums from Mexican States are listed in Table 1, however. We define recovery sites west of tlie Continental Divide as including British Columbia, those states designated by tlie BBL with tlie Pacific Flyway code 4 (Washington, Oregon, California, Idaho, Nevada, Utah, and Arizona), all Mexican States bordeiing the Pacific coast, and Cential American locations on the Pacific coast. All othei' recovery sites are treated as east of the Continental Divide. All bii'ds weie banded witli standard aluminum bands while they were flightless juveniles. During the marking program at Arod Lakes and Canyon Feny Reservoir aluminum bands weie supplemented with colored plastic bands (red-Arod, blue-Canyon Feny); in all, 575 (25.9%) of 2220 birds at Arod Lakes, and 763 (49.6%) of 1537 birds at Canyon Feny Reservoir were color banded. Juveniles considered too small to retain bands were not marked. We revisited one colony each year in 1992-1995 to look for lost bands. Band losses before fledging each year ranged from 0.9-1.3% (1.03 ± 0.19%, N = 4). Ciixumstances identified with band retums (found dead, shot, etc.) were taken from coded data provided by the BBL. We assigned returns to one of four season categories based on the annual cycle (see Evans and Knopf 1993). Spring (February, March, April) is tlie season when birds are beginning to move to breeding colonies, summer (May, June, July) is when breeding birds and first-year birds are mostly associated witli the breeding colonies, autumn (August, Septembei, October) tends to be the period when birds move to wintering areas, and winter (November, December, January) is the period when birds remain on wintering grounds prior to moving nortli towards breeding colonies. Bowdoin Lake Colony n Colony Site Number of Recoveries o 1 • 2-3 • >10 ^:^?=^^^=^^j^ ■1000 kilo meter? Figure 1. Distribution of American White Pelican returns for the Bowdoin Lake colony, Phillips County, Montana. Five of twelve Mexico returns are not shown because specific location data were not available (states of recovery are listed in Table (1). Canyon Ferry Colony D Colony Site Number of Recoveries o 1 . 2-3 . >10 1000 Kilometers Figure 2. Distribution of American White Pelican returns for tlie Canyon Feiry Reservoir colony, Broadwater County, Montana. Arod Lake Colony D Colony Site Number of Recoveries o 1 . 2-3 . >10 1000 Kilometers Figure 3. Distribution of American White Pelican returns for the Arod Lakes colony, Teton County, Montana. Mortality rates for immature and adult pelicans, and overall colony means, were calculated from composite life tables based on band returns with definite dates of recoveiy (see Ryder 1981). This method is likely to overestimate tnie mortality rates (Seber 1972, Botkin and Miller 1974) but is used heie for comparisons among colonies. Where we used statistical tests in analyses of frequency distributions of recovery data we followed standard procedures (Sokal and Rohlf 1981). Results Recovery location: The 276 band returns provided by tlie BBL for tlie Bowdoin, Canyon Ferry, and Arod Lakes American White Pelican colonies are summarized by geopolitical recovery location in Table 1 and Figs. 1-3. Pelicans banded at these thiee Montana colonies were reported from 55° 50' N (Albeita) to 13° 10' N (Honduras). Recoveries witliin Montana ranged from 20.3-25.2% of total returns reported by tlie BBL for tlie tliree colonies. The difference among tlie thiee pelican colonies in the proportion of in-state recoveries was not statistically significant (Gj = 0.562, P> 0.5). Out-of-Montana recoveries west of tlie Continental Divide (Table 1) accounted for 14.0% of returns for Bowdoin, 87.2% for Canyon Feny, and 69.6% for Arod Lakes. The differences between colonies in proportions of recoveries east and west of the Continental Divide were highly significant (G? = 91.390, P « 0.001), with the fewest western returns for the colony fartliest east from the Divide. For the Arod Lakes and Canyon Feny colonies, the greatest number of out-of-state returns (45.1% and 64.4%, respectively) came from California and Idaho (Table 1). California and Idaho accounted for < 6% of total recoveries for tlie Bowdoin colony. By conti'ast, midwestern and central states (North Dakota, South Dakota, Nebraska, Kansas, Oklahoma and Texas) accounted for 32.2% of total returns for Bowdoin, but only 5.1% and 7. 8% for Canyon Feiry and Arod Lakes, respectively. Ten of 12 Mexican recoveries for Bowdoin occuired in states east of the Continental Divide. The opposite pattern was evident for tlie Canyon Ferry and Arod colonies combined, with 66.7% (6 of 9) of recoveries from states west of the Continental Divide. Routes of movement: Recoveiy locations are plotted on maps for each colony (Figs. 1-3), excluding Mexico recoveries for Bowdoin (see Table 1). These maps suggest the primary routes of migration used by members of each colony. Bowdoin pelicans apparently moved to the south and east to the Gulf of Mexico, where tliey spread out along die coast (Fig. 1). They seemed to use a broad migratoiy patliway west of tlie Mississippi River to reach tlie Gulf. The minority tliat moved west traveled to the Snake River in soutliem Idaho, and then appeared to move south to the Colorado River and western Mexico. Canyon Ferry and Arod Lakes pelicans moved mostly south and west (Table 1, Figs. 2-3) to the Snake River in southern Idaho and along some unknown route south to tlie Pacific coast or the Salton Sea in southern California. Others moved to Oregon and northern California, and may have continued to southern California and western Mexico. Pelicans that moved east from tlie Canyon Feny and Arod colonies appeared to pass thiough Wyoming to the Nortli Platte River and on to tlie Mississippi River south to the Gulf of Mexico. Some biids appeared to ti'avel soutli through Kansas, Oklahoma, and east Texas. Table 1. Band recovery locations for fledged American White Pelicans banded at tliree colonies in Montana. Colonies are listed from east to west. Recovery location Bowcloiii Colony Canyon F«^y Arod Canada Alberta Manitoba Saskatchewan United States (E of Divide) Arkansas Colorado Florida Iowa Kansas Louisiana Minnesota Missouri Montana Nebraska Noith Dakota Oklalioma South Dakota Texas Wisconsin Wyoming United States (W of Divide) California Idaho Oregon Utaii Wasliington Mexico (E of Divide) Chihualiua Coaliuila Mexico Nuevo Leon Puebla Tainaulipas Vera Cniz Mexico (W of Divide) Baja California Chinas Jalisco Sinaloa Central America Honduras 2 1 6 1 4 1 1 7 6 3 29 1 4 9 2 14 1 1 1 5 1 3 1 1 1 2 2 3 12 31 7 1 1 3 1 1 2 2 2 23 5 1 4 41 5 2 1 3 Total recoveries 115 59 102 The Salton Sea was a particularly important wintering area for American White PeHcans from tlie Canyon Ferry and Arod colonies. Forty-one percent of 161 band recoveries for the two colonies (66 total: 30 from Canyon Feny and 36 from Arod Lakes) were reported from the Salton Sea, representing all cohorts (1991-1996) from each colony. Some pelicans from each of tlie Montana colonies moved north into adjacent Canadian provinces (Table 1, Figs. 1-3). The routes followed and time of year when birds fiist dispelled nortliward is uncertain. The location of recoveries suggested tliat many birds didnot follow river corridors. Two (15.4%) of 13 recovei'ies from tliis group (one ''band only") were for first year birds, both recoveied in September. Other recoveries in Canada were usually from late spring and sunimei; ages of tliese birds ranged from 3-9 years (5.4 ± 2.2 years, N = 1 1). Fidelity to natal colony: There are few data on natal colony fidelity for two of tlie Montana colonies. Recoveries at natal colonies (not including first-year juveniles and*'band only" returns) totaled sixbiids at Bowdoin (ages 2, 2, 4, 5, 6, 9) and tliiee birds at Canyon Ferry (ages 2, 4, 7). The Arod Lakes colony was searched for banded birds several times by volunteei' observers and MTNHP personnel, who reported 15 tliird-year birds, 9 fourtli-year birds, 5 fifth-year birds, 3 seventli-year birds, and 1 eighth-year biid. Of tliese, 14 third-year and 6 fourdi-year birds were killed in 1997 by a late- sunimei' hailstomi (J. and D. Rhinehart, personal communication). In-state band recoveries for birds that moved between colonies were rare. The band of one bird marked at Medicine Lake in 1967 was found at Bowdoin in 1983 (B. Madden personal communication). A sixth-year bii'd banded at Bowdoin was found dead following a storm in July 1942 at Medicine Lake. Bii'ds banded at Bowdoin or Medicine Lake have yet to be reported at Canyon Feny or Arod Lakes, but banding ceased at these two eastern colonies well before the western colonies were founded, so this result is expected. Of more than 3500 birds banded at Canyon Ferry and Arod Lakes during 1991-1996, only one has been reported from tlie two older colonies in eastern Montana, and only two otliers have been reported moving between the two newer colonies. One sixth-year bird from Canyon Feny was found dead at Bowdoin in August 1998, and one fourth-year bird was observed paired in May 2000 at Arod Lakes. One sixtli-year bird from Arod Lakes was found injuied in September 1997 at Canyon Feny. Cifcumstaiice of recoveries: Most of the returns from Bowdoin and Arod were from biids ''found dead"— 53.0% and 42.2%, respectively (Table 2). For Canyon Feny, 39.0% of all returns weie from birds found dead. Proportions of birds found dead (excluding band-only returns) differed significantly among colonies (G2 = 6.142, P < 0.05) but remain a substantial unknown in tlie analysis of mortality for all Montana colonies. Shooting and disease were the second most frequently identified causes of band recoveiy across all colonies. Shot pelicans represented 21.7% of recoveries for Bowdoin, but only 0.0% and 2.0% of recoveries for Canyon Feiry and Arod Lakes, respectively. Disease accounted for 47.5% and 34.3% of returns for tlie Canyon Ferry and Arod colonies, respectively, but was undocumented among Bowdoin returns. Rates of shooting and disease retums (excluding band-only) differed very significantly between colonies (G^= 105.932, P « 0.001). The three mortality 'Tactors" combined (found dead, shot, and disease) accounted for 74.8-86.4% of the total recoveries for the three Montana colonies (Gj = 3.358, P > 0. 1). Table 2. Circumstances associated with bands recovered from fledged American White Pelicans banded at tliree Montana colonies. Colonies are listed east to west. Colony Recovery circumstance Bowdoin Canyon Ferry Arod Found dead 61 23 43 Shot 25 — 2 Disease — 28 35 Injured 8 1 4 Caught — 2 2 Weatlier 7 Entangled 2 1 2 Collision 1 — 2 Band only 11 3 5 Sight record — 1 7 Total lecoveiics 115 59 102 Mortality rates attd age {Itstribution of recoveries: Age of recovery for Montana pelicans from the three colonies (Table 3) ranged from 1-20 years; the oldest bird was from Bowdoin. Adult mortality rates ranged from 23.0% at Bowdoin to 57.6% at Arod (Canyon Feiry was 36.5%). Oveiallmean annual mortality rates ranged from 29.9% at Canyon Ferry to 39.3% at Arod (Bowdoin was 30.0%). The adult estimates for Arod and Canyon Feiry are unusually high for this species (see Ryder 1981, Evans and Knopf 1993), as is the overall estimate for Arod. Table 3. Age distiibution of recovered American White Pelicans banded as flightless young at three Montana colonies. Colonies are listed east to west. Year 1 is tlie first year following banding, and so on. ''Band only" recoveiies are not included in the calculations. Colony Year Bowdoin Canyon Ferry Arod 1 2 3 4 5 6 7 8 9 10 11 12 >12 Total recoveries 104 56 97 50.0% (52) 21.4% (12) 38.1%o (37) 7.7% (8) 23.2% (13) 18.6% (18) 6.7% 0) 8.9% (5) 12.4% (12) 3.8% (4) 21.4% (12) 15.5% (15) 8.7% (9) 7.1% (4) 8.2%. (8) 7.7% (8) 12.5% '(7) 5.2% (5) 2.9% (3) 3.6% (2) 1.0% (1) .._ 1.8% (1) 1.0%o (1) 8.7% (9) — — 1.9% (2) 1.0% (1) _.. _._ 1.0% (1) — — The majority of band recoveries from Bowdoin and Arodweie for birds of all ages ''found dead" of unknown causes (Table 2). Typical for pelicans (Sloan 1982, Evans and Knopf 1993) abouthalf of tlie recovei'ies for each colony, excluding ''band only" returns, were immatures (birds < 3 years). Annual mortality rates of immature pelicans ranged from 24.2% at Canyon Feiry to 38.5% at Bowdoin (Arod was 35.4%). Inimatuie (first- and second-year) birds comprised the majority shot. Nearly two thirds of shot birds from Bowdoin weie immatures. Both of the shot pelicans from tlie Arod colony were second- year immatuies; tliere were no recoveries from Canyon Feny linked to shooting. Pelicans of all ages appeared equally susceptible to botulism. Large proportions (two thirds or more) of second- and fourth-year birds in recoveries from the Canyon Feny colony (Table 3) were related to a botulism outbreak in the Salton Sea, California. Similar peaks in recoveries of tliese two age groups from the Arod colony were evident, and attributable to tlie same botulism outbreak in 1996. Furthermore, disease contiibuted to increased mortality in each cohort (Table 4). Pooled recovery data for tlie Canyon Ferry and Arod colonies showed a noticeable pattern in peak recoveries for each age class, beginning witli first-year returns in tlie 1996 cohort (15 retuins) and running diagonally tlirough tlie data matiix in Table 4 to tlie upper right (bold numbeis), ending with sixth-year birds of the 1991 cohort (4 returns). Botulism-related recoveries at Salton Sea for the six cohorts (1991-1996) accounted for 55 of 76 post- July returns in 1996. Percentage of returns attjibuted to botulism were 67% of first-year 1996 birds, 83% of second-year 1995 birds, 86% of tliird-year 1994 birds, 85%>offourtli-year 1993 biids, 71%) of fiflli-year 1992 birds, and 100%> of sixth-year 1991 birds. Season of recovery: Over half of the band returns for the three colonies combined (not including "band only" returns) occurred in autumn (Fig. 4). This pattern was evident for immature pelicans (Fig. 4a), witlino significant difference in the seasonal distribution of returns for the three colonies (G^= 1.114, P > 0.9). Adult returns for the Arod and Canyon Feny colonies also showed an autumn peak in returns similar to tlie pattern for immatures (Fig. 4b). Bowdoin, however, showed equal proportions of adult returns for all seasons, in contiast to the otlier two colonies (Gs = 24.036, P < 0.001; Arod and Canyon Ferry pooled). The difference in seasonal distiibution of adult recoveries was probably related to disease-caused mortality of migrant and wintering Arod and Canyon Ferry pelicans at the Salton Sea. 10 Spring Summer Autumn Winter Figure 4. Seasonal distribution of band returns for immature (A) and adult (B) American White Pelicans from Arod Lakes (open fill), Canyon Feny (right- diagonal fill), and Bowdoin (crosshatched fill) colonies. Numbers of band returns are indicated above tlie bars. 11 Discussion Colony isolation {utd/lywi^ affiliation: Data gathered from juvenile American White Pelicans banded at tlie Bowdoin, Canyon Feriy, and Arod colonies suggest tliat migratory movements fall into two groupings. Birds from the westernmost colonies (Canyon Ferry, Arod Lakes) migrate largely west across the Continental Divide via Idaho to southern California and western Mexico. Birds from the eastern colony (Bowdoin) migrate largely south and east along the Missouri and Mississippi rivers, or tlirough the Midwest, to the Gulf of Mexico and eastern Mexico. The Medicine Lake colony follows the Bowdoin pattern of dispersal (B. Madden personal communication). The two western-most American White Pelican colonies in Montana (Canyon Ferry and Arod Lakes) are functionally a part of the Intermountain West and Pacific flyways, which include colonies in British Columbia, Utah, Oregon, Nevada, California, and Wyoming (see Behle 1958, Diem and Condon 1967, Veiineei' 1977, Keitli and O'Neill 2000). The two eastern-most pelican colonies in Montana (Bowdoin and Medicine Lake) are functionally a part of the Cential and Mississippi flyways (see Stiait and Sloan 1975, Vermeer 1977, Ryder 1981), where the global majority of American White Pelican colonies occur in the uppei' Midwest and northern Great Plains. While tlie recovery data indicate ample opportunity for in-state immigiation among Montana pelican colonies, documented cases are extiemely rare. There are only two records of colony exchange between Bowdoin and Medicine Lake, separated by 250 km, and only two records of colony exchange between Canyon Feny and Arod Lakes, located about 200 km apart. Because many pelicans were color-marked at tliese latter two colonies, increasing the probability of documenting inter-colony movements, we would expect more records of inter- colony exchange if it were occuixing with any regularity. Theie is also only one report of a pelican from one colony group appearing in the otlier (a Canyon Feny bird reported at Bowdoin). This suggests tliat each colony is relatively isolated repro due lively despite tlieir close proximity and tlie dispersal capabilities of the birds. Two other factors may account for tlie apparent rarity of inter-colony movements. Fiist, there has been no banding at tlie eastern colonies since the western colonies became established. Second, colony surveys are not routinely conducted during tlie pre- and post-breeding periods, tliereby reducing the likelihood of observing banded pelicans or recovering their remains. The two colony gioups in Montana may represent parts of two larger metapopulations (Hanski 1996), witli occasional immigration and emigration among other colonies along tlie migration routes of the Montana biids. Little hard data exists showing such a reproductive connection between tlie various colonies during the breeding season, altliough a few pelicans from Montana have been recoveied in sunimei' from colonies in otlier states and provinces. Nevertlieless, pelicans establish new colonies and reestablish old ones (Sloan 1982, Findholt and Diem 1988, Paullin et al. 1988, Bryan 1991). The natal pelican colony for the founders of tlie Arod and Canyon Feny colonies is unknown; migratory behavior of birds from tliese two western Montana colonies indicates the founders probably weie birds with ties to colonies in states west of the Continental Divide. 12 Table 4. Distiibution of band recoveries by age (excluding '*band only" returns) for tlie combined Arod Lake and Canyon Feiry colonies of American White Pelican in Montana. Numbers in bold show the elevated numbei' of recoveries in 1996 for each age class, mostly from the Salton Sea, California. Year No. Yf :ar of lecoverv after banding Total banded banded 1 2 3 4 5 6 7 8 returns % 1991 200 2 4 6 3.0 1992 668 10 1 2 1 7 1 1 1 24 3.6 1993 1059 10 2 4 20 1 4 2 1 44 4 2 1994 668 4 4 7 ___ 1 1 ___ ___ 17 2.5 1995 791 8 23 2 5 1 2 ___ ___ 41 5.2 1996 371 15 1 2 1 2 — — — 21 5.7 Total 3757 49 31 17 27 12 12 3 2 154 4.1 Implications for management: Conservation of American White Pelican populations requires coordinated management at widely separated breeding and wintering sites. Pelicans from the Arod and Canyon Ferry colonies in Montana clearly are linked to wintering sites in California. Botulism outbreaks in the Salton Sea of soutliern California could significantly affect adult survivorship of the two Montana colonies (Fig. 4, Tables 2 and 4) as well as other western colonies (Colin 2000). All age classes of pelicans from the Arod Lakes and Canyon Feny colonies (Tables 3-4 and tlieir interpretation) weie affected by tlie 1996 botulism outbreak at tlie Salton Sea; catastiophic events of tliis magnitude could alter the age-stnicture of tlie breeding membership at each colony. Because the Salton Sea is likely to undergo massive changes in tlie 21^' centuiy (Jehl 1994), tlie results of future management activities at this site will be of importance to the pelicans of tlie western colonies in Montana. The Snake River in southern Idaho is anothei' important stopping spot in the non-breeding season for members of tliese two colonies (Table 1, Figs. 2-3) experiencing increased pressure from human recreation and agiiculture that could negatively affect pelican use (personal obseivation). The Bowdoin and Medicine Lake colonies are clearly linked primarily to coastal winteiing sites along and near tlie Gulf of Mexico, ratlier than tlie Salton Sea and other saline lakes of the Great Basin. Shooting in fomier decades, mostly of immature birds in migration and at wintering sites, was a significant mortality factor for the membeiship of tlie Bowdoin colony (Table 2), and Medicine Lake as well (B. Madden personal communication). Similar mortality from shooting in fomier decades has also been documented for otlier eastern colonies such as Chase Lake, North Dakota (Strait and Sloan 1975), and Riverside Reservoir, Colorado (Ryder 1981); at Marsh Lake, Minnesota, where banding was begun more recently tlian at Chase Lake and Riverside Reservoii, shooting mortality was responsible for a smaller percentage of returns, but most of these came from Mexico and Central America where enforcement of protection may be more difficult to implement (King and Grewe 2001). Whether shooting remains a major contributor to mortality is unknown; protection for pelicans has increased in recent decades, but concunent band return data are usually not available. Data from the Arod and Canyon Feny colonies suggest shooting is not a serious thieat, at least nortli of Mexico, but tliis could be due to their migiatory routes in the hitemiountain West and Pacific flyways ratlier than a range-wide decline in shooting mortality. In addition, tlie preponderance of "found dead" returns for all Montana colonies could include a significant number of shot birds reported as "found dead" to avoid prosecution (King and Grewe 2001). One newer thieat for pelicans winteiing in tlie Gulf Coast states (which includes tlie Bowdoin and Medicine Lake birds) is a rapidly expanding catfish aquaculture 13 industiy. There are increasing numbers of pelicans foraging and loafing near tliese aquacultiire facilities (King and Grewe 2001) tliat likely will result in more human-pelican conflicts and tlie possible implementation of letlial measures to control pelican depredation. In conclusion, we concur with Pamell et al. (1988) that management of pelican colonies involves more tlian just the colony itself, altliough events at individual colonies remain exti'emely important in determining year-to-year viability; Montana pelicans experienced significant declines in tlie 1960's and 1970's due to water level fluctuations at their breeding colonies (Sloan 1982). However, because most colonies are probably dependent on larger regional populations, effective management needs to address factors tliat affect recniitment of sub-adults into breeding populations tliroughout regional landscapes. A coordinated color-marking program at each of tlie Montana colonies, witli routine pre-and post- breeding suiveys for marked birds could clarify the isolation of tlie two colony groupings, tlie degree of dispersal among the individual colonies witliin each grouping, and better document colony dispersal during tlie non-breeding season. Similar marking and monitoring programs would yield valuable data for breeding colonies in otlier regions. The results of this and further studies would help in developing regional management plans that 1) identify current and impending thieats to breeding and wintering populations linked thiough dispersal, 2) provide for coordinated responses to those thieats, and 3) support the maintenance of viable populations of American White Pelicans tliroughout tlieir range. 14 Literature Cited Behle,W.H. 1958. The bud life of Great Salt Lake. University of Utah Press. Salt Lake City, UT. 230 pp. Botkin, D. B., and R. S. Miller. 1974. Mortality rates and survival of birds. American Naturalist 108:181-192. Biyan, S. 1991. Pelicans nesting on Lake Nipigon. Ontario Birds 9:58-63. Chapman, B. R. 1988. History of the White Pelican colonies in south Texas and nortliem Tamaulipas. Colonial Wateitirds 11:275-283. Cohn.J. P. 2000. Saving the Salton Sea. BioScience 50:295-301. Diem, K. L., and D. D. Condon. 1967. Banding studies of water birds on die Molly Islands, Yellowstone Lake, Wyoming. Yellowstone Libraiy and Museum Association. Yellowstone National Park, WY. 41 pp. Evans, R. M, and F. L. Knopf. 1993. Ameiican White Pelican (Pelecamis erythrorhyiwhos). In The Bii'ds of Nortli America, No. 57 (A. Poole and F. Gill, Eds.). Philadelphia: The Academy of Natural Sciences; Washington, D.C.: The American Ornithologists' Union. 24 pp. Findholt, S. L., and K. L. Diem. 1988. Status and distribution of American White Pelican nesting colonies in Wyoming: an update. GreatBasinNaturalist 48:285-289. Hanski, I. 1996. Metapopulation ecology. Pp. 13-43 ///Population dynamics in ecological space and time (O. E. Rhodes, Jr., R. K. Chesser, andM, H. Smitli, eds.). University of Chicago Press, Chicago, IL. Jehl, J. R., Jr. 1994. Changes in saline and alkaline lake avifaunas in vi^estem North America in the past 150 years. Studies in Avian Biology 15:258-272. Keith, J. O., and E. J. O'Neill. 2000. Movements of juvenile American White Pelicans from breeding colonies in California andNevada. Waterbirds 23:33-37. King, D. T., and A. H. Grewe, Jr. 2001. Movements and mortality of American White Pelicans banded at Marsh Lake, Minnesota. North American Bird Bander 26:57-60. Lies,M. F., andW. H. Behle. 1966. Status of the White Pelican in the United States and Canada tlirough 1964. Condor 68:279-292. Pamell, J. F.,D. G. Ainley, H. Blokpoel,B. Cain, T. W. Custer, J. L. Dusi, S. Kress, J. A. Kushlan, W. E. Southern, L. E. Stenzel, and B. C. Thompson. 1988. Colonial waterbird management in Nortli America. Colonial Waterbirds 11:129-169. Paullin, D. G., G. L. Ivey, and C. D. Littlefield. 1988. The re-establishment of Ameiican White Pelican nesting in the Malheur-Harney Lakes Basin, Oregon. Muixelet 69:61-64. 15 Ryder, R. A. 1981. Movements and mortality of White Pelicans fledged in Colorado. Colonial Waterbirds 4:72-76. Saunders, A. A. 1921. A distiibutional list of the birds of Montana. Pacific Coast Avifauna Number 14. 194 pp. Seber, G. A. F. 1972. Estimating survival rates from bird-band returns. Journal of Wildlife Management 36:405-413. Sidle, J. G., W. H. Koonz, and K. Roney. 1985. Status of the American White Pelican: an update. Amei'ican Birds 39:859-864. Sloan, N. F. 1982. Status of breeding colonies of White Pelicans in tlie United States tlirough 1979. Amei'ican Birds 36:250-254. Sokal, R. R., and F. J. Rohlf. 1981. Biometry, second edition. W. H. Freeman and Company. San Francisco, CA. 859 pp. Strait, L. E., and N. F. Sloan. 1975. Movements and mortality of juvenile White Pelicans from North Dakota. Wilson Bulletin 87:54-59. Thompson, B. H. 1933. History and present status of the breeding colonies of tlie White Pelican (Peleccmits eryihrorimtchos). National Park Service, Contiibution of the Wildlife Division, Occasional Papei'No. 1. 85 pp. Venneei', K. 1970. Distribution and size of colonies of White Pelicans, Pelecamis erythroriiyncbos, in Canada. Canadian Journal of Zoology 48:1029-1032. Veiineei', K. 1977. Comparison of White Pelican recoveiies from colonies east and west of the Canadian Rocky Mountains. Murrelet 58:79-82. Weydemeyer, W., and V. L. Marsh. 1936. The bird life of Lake Bowdoin, Montana. Condor 38:185- 198. Willett, G. 1907. Summer birds of a prairie lake. Condor 9:105-106. 16