FLOOD TOLERANCE OF PLANT SPECIES IN BOTTOMLAND FORESTS OF THE SOUTHEASTERN UNITED STATES BY RUSSELL FRANCIS THERIOT A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1992 UNIVERSITY OF FLORIDA USSAS1ES Digitized by the Internet Archive in 2011 with funding from University of Florida, George A. Smathers Libraries with support from LYRASIS and the Sloan Foundation To my parents http://www.archive.org/details/floodtoleranceofOOther ACKNOWLEDGMENTS This research effort was greatly aided by innumerable people. Foremost among them is Dr. Dana Sanders, who assisted me in all phases of the study. His advice and encouragement throughout this effort are greatly appreciated. I thank Drs . Ken Rodgers , Dan Evans, and Tom Hein- eke for their assistance in identifying the plant species in this study. Blake Parker was invaluable in helping me interpret the soils in the study. Don Hill conducted the geodetic surveys for all of the sites. Phil Jones and Jeff Irvin taught me all I know about hydrology. Phil was especially helpful in correcting the water surface elevations between the study sites and the gauging stations, and Jeff interpreted the hydrologic program into FORTRAN language. I am also grateful to Dr. Dara Wilber, who assisted me in the statistical analyses of the data. Many people provided helpful discussion and advice on various aspects of the study, including Drs. Donal Hook, Bill Patrick, Helen Leitman, Sandra Brown, and Bill Mitsch. I gratefully acknowledge the support and understanding of Dr. Bob Engler, who allowed a work schedule flexible enough to finish the manuscript. I especially want to thank my wife and children, who encouraged me to go on when it would have been easy to quit. iii Finally, I would like to thank my major professor, Dr. Jerome Shireman, for his advice, encouragement, and support in guiding my degree program, and also my committee for their helpful suggestions and patience . IV TABLE OF CONTENTS Page ACKNOWLEDGMENTS iii LIST OF TABLES vii LIST OF FIGURES x ABSTRACT xil INTRODUCTION 1 Plant Community Organization 2 Bottomland Forest Community Organization 3 Zonation of Bottomland Forests 5 Purpose and Objectives 7 METHODS 8 Study Area 8 Site Selection 10 Determining Hydrologic Zone Elevations 11 Site Preparation and Data Collection 14 Analyzing Vegetation Data 16 Calculating Species FTI Numbers 17 RESULTS AND DISCUSSION 20 Flood Analysis of Study Sites 20 Vegetation Data 22 Weighted Averaging 36 Statistical Analysis of the Vegetation Data 37 Cluster Analyses 38 Discriminant Function Analysis 47 Regional Variation in Species FTI Numbers 56 SUMMARY AND CONCLUSIONS 59 APPENDIX A SITE DESCRIPTIONS AND MAP LOCATIONS 62 APPENDIX B GUIDE FOR COMPUTER PROGRAM FOR ANALYZING HYDROLOGIC DATA 90 APPENDIX C HYDROGRAPH FOR STEELE BAYOU (SITE 3) 97 APPENDIX D IMPORTANCE VALUES FOR SPECIES BY ZONE AND VEGETATION LAYER 102 APPENDIX E FTI PLANT LIST AND COMPARISON WITH TWO OTHER WATER -TOLERANCE RATING SYSTEMS 186 REFERENCES 202 BIOGRAPHICAL SKETCH 207 VI LIST OF TABLES Table Page 1 Hydrologic Zones Occurring in Bottomland Forests of the Southeastern United States 12 2 Annual Flood Frequency (Percent of Years in Which Boundary Is Exceeded at Least Once during Growing Season for More than 7 days) for Zone Boundaries .... 21 3 Average Annual Duration of Flood Events (days) for Zone Boundaries 23 4 FTI Numbers of Commonly Occurring Species in this Study 25 5 Variations in Species Flood Tolerance Index Numbers According to Life Stage 30 6 Comparison of Three Water-Tolerance Ratings for Selected Bottomland Forest Tree Species 34 7 Relative Frequencies in Each Hydrologic Zone of Tree Species Used in the Statistical Analyses; Groupings of Species Correspond to Cluster Membership 40 8 Relative Frequency of Occurrence of Each Sapling Species in the Hydrologic Zones along with Their Cluster Memberships 45 9 Relative Frequencies in Each Hydrologic Zone of the Vine Species Used in Statistical Analyses 47 10 Mean Importance Values for Species in Each Cluster Used in the DFA, Arranged by Zone/Sample 50 11 Predicted Hydrologic Zones (Columns) and Actual Zones (Rows) Based on DFA Results Using Only Tree Importance Values 51 12 Predicted Hydrologic Zones (Columns) and Actual Zones (Rows) Based on DFA Results Using Average FTI Values for All Observed Tree Species at the Site . . 52 vn Table Page 13 Cross -Validation Results of Zone Membership Using Linear Discriminant Function Analysis 53 D-l Importance Values for Species Occurring at Site 1, Arranged by Zone and Vegetation Layer 103 D-2 Importance Values for Species Occurring at Site 2, Arranged by Zone and Vegetation Layer 108 D-3 Importance Values for Species Occurring at Site 3, Arranged by Zone and Vegetation Layer 113 D-4 Importance Values for Species Occurring at Site 4, Arranged by Zone and Vegetation Layer 121 D-5 Importance Values for Species Occurring at Site 5, Arranged by Zone and Vegetation Layer 124 D-6 Importance Values for Species Occurring at Site 6, Arranged by Zone and Vegetation Layer 127 D-7 Importance Values for Species Occurring at Site 7, Arranged by Zone and Vegetation Layer 130 D-8 Importance Values for Species Occurring at Site 8, Arranged by Zone and Vegetation Layer 135 D-9 Importance Values for Species Occurring at Site 9, Arranged by Zone and Vegetation Layer 137 D-10 Importance Values for Species Occurring at Site 10, Arranged by Zone and Vegetation Layer 141 D-ll Importance Values for Species Occurring at Site 11, Arranged by Zone and Vegetation Layer 147 D-12 Importance Values for Species Occurring at Site 12, Arranged by Zone and Vegetation Layer 152 D-13 Importance Values for Species Occurring at Site 13, Arranged by Zone and Vegetation Layer 158 D-14 Importance Values for Species Occurring at Site 14, Arranged by Zone and Vegetation Layer 163 D-15 Importance Values for Species Occurring at Site 15, Arranged by Zone and Vegetation Layer 168 V11L Table Page D-16 Importance Values for Species Occurring at Site 16, Arranged by Zone and Vegetation Layer 173 D-17 Importance Values for Species Occurring at Site 17, Arranged by Zone and Vegetation Layer 180 E-l FTI Plant List 187 IX LIST OF FIGURES Figure Page 1 Zonal classification of bottomland forest wetlands (adapted from Clark and Benforado 1981) 6 2 Study area and sites in the southeastern United States . 9 3 Representation of a typical research site 15 4 Ecological amplitude of some commonly occurring species; CAAQ: Carya aquatica; FOAC: Forestiera acuminata; FRPE: Fraxinus pennsylvanica; LIST: Liquidambar styraciflua; NYAQ: Nyssa aquatica; PITA: Pinus taeda; QUAL: Quercus alba; QULY: Quercus lyrata; SAAL: Sassafras albidum; QUNI : Quercus nigra; TADI : Taxodium distichum; ULAM: Ulmus americana 32 5 Cluster diagram for trees 39 6 Cluster diagram for saplings and shrubs 42 7 Cluster diagram for saplings alone 43 8 Cluster diagram for shrubs alone 44 9 Cluster diagram for vines 46 10 Cluster diagram for herbs 48 11 Mean tree FTI numbers plotted versus observed and predicted hydrologic zones for all 55 sites 57 A-l Neches River (sites 1 and 2) 64 A-2 Steele Bayou (site 3) 66 A-3 Ouachita River (sites 4 and 5) 68 A-4 Yazoo River (site 6) 70 A-5 Big Black River (site 7) 72 A-6 L'Anguille River (sites 8 and 9) 74 Page A-7 Pearl River (site 10) 76 A-8 Apalachicola River (site 11) 78 A-9 Apalachicola River (site 12) 79 A-10 Ocmulgee River (site 13) 81 A-ll Altamaha River (site 14) 83 A-12 Edisto River (site 15) 85 A-13 Lynches River (site 16) 87 A-14 Waccamaw River (site 17) 89 B-l Program logic for computation of days saturated 95 C-l Hydrograph for Steele Bayou (site 3); the shaded areas represent the nongrowing season, a: 1961-1965, b: 1966-1970, c: 1971-1975, d: 1976-1980 98 XI Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy FLOOD TOLERANCE OF PLANT SPECIES IN BOTTOMLAND FORESTS OF THE SOUTHEASTERN UNITED STATES By Russell Francis Theriot May 1992 Chairman: Jerome V. Shireman Major Department: School of Forest Resources and Conservation Vegetation data on species composition along a hydrologic gradient were collected at 17 bottomland forest sites throughout the southeastern United States. Weighted averages based on importance values calculated from 55 stands resulted in flood tolerance index (FTI) numbers, the optimum position for each species along the defined hydrologic gradient, for 312 identified species. Commonly occurring species were evaluated using cluster analyses and discriminant function analyses. Data on tree, sapling, and vine species clustered into distinct groups, with tree species being the most reliable; however, shrubs and herbaceous species did not cluster dis- tinctly. Discriminant function analysis using FTI numbers for tree species proved to be 82 percent reliable in predicting zones. The accuracy of the Flood Tolerance Index (FTI) numbers did not vary regionally in the southeastern United States. Therefore, a single XII FTI number calculated for each species can be used to predict hydrologic zones for the entire study area. xm INTRODUCTION Bottomland forests are found in the floodplains of rivers in the southeastern United States from eastern Texas to Virginia. They have distinct topographic features that are the result of historical hydro- logic characteristics of the rivers, including periodic fluctuations in water levels and changes in stream course. Recognizable floodplain topographic features include first bottoms, second bottoms or terraces, uplands, riverfront, swamp, poorly drained flats, well-drained flats, and sloughs (Putnam, Furnival , and McKnight 1960). These features are characterized by different hydrologic regimes and can be identified as a hydrologic gradient transitional between permanent water and terrestrial uplands . Many studies have previously described the forest communities associated with these floodplain features (Putnam, Furnival, and McKnight 1960; Broadfoot and Williston 1973; Chambless and Nixon 1975; Hodges and Switzer 1979; Mohler 1979; and Hupp and Osterkamp 1985). However, studies describing the relationship between plant species dis- tribution and specific inundation/saturation regimes in bottomland for- ests are rare (Bedinger 1971; Mohler 1979; Huffman 1980; and Leitman, Sohm, and Franklin 1984). Even so, these studies all demonstrated that frequency and duration of inundation/saturation exert a controlling influence on the composition, structure, and distribution of wetland plant communities. As an example, Bedinger (1971) found a definite 1 relationship between the distribution of plant species and the frequency and duration of flooding in the lower White River Valley, Arkansas. Using flood frequency and duration, he defined four species associations on the White River floodplain, each of which had a distinctly different tolerance to inundation. He concluded that based on plant species- flooding relationships, plant communities could be used as a basis to transfer flooding parameters to ungauged streams. Plant Community Organization The concept of community structure has been debated for decades. Clements (1916) first described communities as discrete, self -organizing entities that could be considered as discrete organisms. Gleason (1917) disagreed with Clements' organismal concept and proposed a hypothesis relating to the individualistic occurrence of plants. His hypothesis has developed into the continuum concept, which indicates that plant species distribution is determined by the species' response to its envi- ronment. Whittaker (1967) and Mcintosh (1980) later developed Gleason' s ideas, expanding on the continuum concept. They maintain that since plant species adapt differently, no two occupy the same zone. This results in a continuum of overlapping species associations, each responding to subtly different environmental factors (e.g., water, soil pH, nutrients, and solar radiation). A continuum can be described for each factor in various increments or zones. Zonation simply describes the different levels of an environmental gradient to which a species is responding. The reason zonation is so obvious in some ecosystems is that environmental gradients are "ecologically" steep and groups of species have fairly similar tolerance that tend to group them on these gradients (Mitsch and Gosselink 1986) . Gleason's individualistic hypothesis can be supported by several studies (Curtis and Mcintosh 1951; Brown and Curtis 1952; Bray 1956; Whittaker 1956; Curtis 1959; Whittaker and Niering 1965; and Mohler 1979) . These studies show that although species have different ecologi- cal amplitudes and, in fact, do not occupy the same niche, they organize as units based on similar ecological conditions. Moreover, intergrades caused by interspecific competition occur between defined types of plant associations. These intergrades can be attributed to continuous envi- ronmental variability in time or space or to environmental modification. Bottomland Forest Community Organization Van Der Valk (1981) developed a qualitative model of succession in freshwater wetlands based on the "individualistic" approach to vegeta- tion proposed by Gleason. He based his approach on three key life his- tory features of plant species: life-span, propagule longevity, and propagule establishment requirements. These features are all directly affected by the flooding on bottomland forests. Brinson (1990) , in discussing the "power line" designation for a wetland classification developed by Kangas (1990), considered the power and frequency of inundation as the way in which flood events organize the plant communities in riverine forests. He characterized the flood events as high, medium, and low power events with flood power and fre- quency of inundation being inversely proportional. High power flood events have a low frequency and determine patterns of the large flood- plains features (e.g., oxbow lakes, relict levees, and low ridges and swales) that persist for hundreds to thousands of years. Medium power flood events, which occur at an intermediate frequency, affect ecosystem structures that exist from decades to hundreds of years. He identified tree species associations as an ecosystem component likely to be influ- enced at this scale. The low power, high-frequency flood events occcur annually and affect short-term patterns such as seed germination and seedling survival. His characterization emphasized the dramatic impact flooding has on the regeneration of vegetation in bottomland forests. Grubb (1977) stated that scientists have failed to understand ade- quately how plant communities maintain themselves because of a failure to account for the phenomenon of regeneration in plant communities. Huenneke and Sharitz (1986) , in a study of microsite abundance and dis- tribution of woody seedlings in a South Carolina cypress- tupelo swamp, concluded that the availability and nature of microsites may affect the distribution and composition of the seedling and sapling strata, thus differentiating the "regeneration niche" described by Grubb. Although plant species association is determined by a number of interacting environmental factors, it is generally agreed that flooding is the dominant environmental factor at work in bottomland forests, affecting regeneration and life under saturated soil conditions. Flood- ing persisting for more than a few days will prevent the replenishment of soil oxygen once the soil microbes and plant roots consume the avail- able soil oxygen in the root zone during respiration. Only those plant species that have evolved a mechanism for living in reducing (anaerobic) soil conditions will survive such conditions. In most instances, recur- ring flooding provides a competitive advantage for plant species that are adapted to saturated and reduced soils. Chemistry of wet soils (Pearsall and Mortimer 1939; Patrick and Mikkelsen 1971; Ponnamperuma 1972; Patrick and Delaune 1976; and Faulkner et al . 1991), and the various physiological effects on vegeta- tion under reducing conditions are well documented (Cannon and Free 1920; Conway 1940; Dubinina 1961; Hosner and Boyce 1962; Hook and Brown 1973; Hook and Scholtens 1978; Vester 1972; and Hook and Crawford 1980). Zonation of Bottomland Forests The hydrologic gradient in bottomland forests ranges from zones of nearly continuous inundation/saturation in deep swamps to infrequent inundation/saturation events for brief periods on upland sites. Because different species respond to different timing and duration of inunda- tion, a strong correlation exists between the distribution of a species and its associated hydrologic and soil-moisture conditions (Hosner and Boyce 1962; Dickson, Hosner, and Hosley 1965; Bedinger 1971, 1978; Larson et al . 1981; Best, Segal, and Wolfe 1990; and Faulkner et al . 1991) . The National Wetlands Technical Council (NWTC) proposed the zonal classification of floodplain forests (Clark and Benforado 1981) . The classification system defined six hydrologic zones based on fre- quency and duration of inundation and soil saturation (Figure 1) and provides the basis for testing in this study. Larson et al . (1981) summarized the works of others on the occur- rence of plant species in the Gulf Coastal Plain from 238 belt transects in Texas, Louisiana, Arkansas, Mississippi, Alabama, and Florida accord- ing to their maximum tolerance to soil -moisture or hydrologic regimes. Larson and his cohorts developed a list of 79 tree and shrub species associated with one or more of the NWTC hydrologic zones. However, the i-i % W a) O 4-1 ■d C tri ■-I e o ■U 4J O CO On o •o cfl S-l O 4-1 c a) CQ T3 C cd A! 03 r-l o a o 1-1 4-1 -a pi (0 0) u ■rH fa o •r-l 4-1 •H W tn o o < O >- z ±i bj z < o z,_ s Q 3 LiJ UJ Qi-yco O OO >- O < CJ CJ c/i o o: a: < O bj (K _i en llJ b. _) Z> bj li- bj b_ b. Q_ O L.Q1LOI/) list identifies only presence or absence of a species in a zone and does not identify the ecological amplitude or optimum position of each species along the hydrologic gradient. Purpose and Objectives The purpose of the study was to develop flood tolerance index (FTI) numbers that refect the optimum position for plant species occur- ring along the hydrologic gradient in bottomland forests of the south- eastern United States. The resulting FTI numbers can then be used to estimate the hydrologic regimes of similar ungauged areas using vegeta- tion. Specific objectives were to develop methods for translating recorded hydrologic data into hydrologic zone elevations for southeast- ern bottomland forests, calculate weighted averages of plant species based on dominance, and determine methods for applying FTI numbers to species occurring in bottomland forests of the southeastern United States. METHODS Study Area The study was conducted in portions of the subtropical ecoregion of the southeastern United States (Bailey 1980) , including portions of eastern Texas and the Gulf and South Atlantic states. Northern limits of the area extended across northern Arkansas, Mississippi, Alabama, Georgia, and South Carolina. The study area included the states of Louisiana, Arkansas, Mississippi, and Alabama. Georgia and South Caro- lina were included, except for the piedmont region. Only the extreme eastern portion of Texas was included, as was the northern portion of Florida (Figure 2) . The intent was to study natural undisturbed sites encompassing the largest possible area where the resulting FTI numbers would be applicable without including areas that would introduce too many additional species or different climatic variables. Specific sites were selected according to the following criteria: (1) No major disturbance (e.g., timber harvesting, ditching, or diking) had occurred during the past 20 years; (2) sufficient hydrologic data (10 to 20 years of daily stream gauge readings) accurately portraying water-level fluctuations on the site (considering ponding, tributary influence between site and gauge, etc.) were available; (3) no site changes (e.g., timber harvesting or ditching) were anticipated during the study period; (4) soil data (e.g., soil surveys, soil series, and/or 8 SCALE 200 200 400 KM Figure 2. Study area and sites in the southeastern United States 10 soil phases, texture, and permeability coefficients) were available; and (5) plant communities were characteristic (e.g., plant communities with few rarely occurring species) of the study area. Site Selection Several hundred potential sites were considered, but most were eliminated because of insufficient stream gauge data. More than 50 sites were visited, but only 17 (Figure 2) satisfied all site criteria and were used in the study. Although all 17 sites met the selection criteria, not all hydrologic zones in each site were suitable for study. Some zones were too narrow and others had been disturbed recently by agricultural or silvicultural practices. Sites 1 and 2 were located in the Neches River basin in southeast- ern Texas. The Steele Bayou, Yazoo River, and Big Black River basins in Mississippi, respectively, were designated sites 3, 6, and 7. Sites 4 and 5 were located in the Ouachita River and sites 8 and 9 in the L'Anguille River basins in Arkansas. Site 10 was located in the Pearl River basin in Louisiana, and sites 11 and 12 in the Apalachicola River basin in Florida. Sites 13 and 14 were located in Georgia in the Ocmulgee River and Altamaha River basins, respectively. Sites 15, 16, and 17 were located in South Carolina in the Edisto, Lynches, and Waccamaw River basins, respectively. All sites were characterized by a growing season of greater than 200 days and average annual rainfall ranging from 105 to 170 cm. The overstory typically ranged from cypress -tupelo or willow in depressions and low flats to white oak-hickory or pine on the high ridges. Inter- mediate areas included overcup oak-bitter pecan, green ash, willow oak, 11 and American elm overstory communities. The herbaceous understory was typically dense with diverse species of trees and shrubs, vines, and herbs. Appendix A includes a general description of each study site. Determining Hydrologic Zone Elevations Hydrologic data for each site were obtained either from the U.S. Geological Survey (flow data) or from the local Corps of Engineers District (stage or flow data) . Data were analyzed using a FORTRAN com- puter program developed for determining hydrologic zone elevations in study sites where flooding occurred. The program output is the duration of inundation plus soil saturation of each hydrologic zone boundary, expressed as flow rate or stage data. Table 1 presents inundation/ saturation frequency and duration for Zones 2 to 6 . Hydrologic zone elevations for each site were computed using the most recent 10 to 20 years of daily stream gauge data. When gauge data were provided as daily discharges (flow rate), a rating table (relation- ship between stage and discharge) was obtained to determine the corre- sponding stages (elevation) . Plant species show little or no adverse effects from flooding in the winter (dormant) season (Hall and Smith 1955; Bruckner, Bowersox, and Ward 1973). Therefore, hydrology during the dormant season was not used in this study to determine zones. The dates of the first and last day of the growing season for each site were provided as input to the computer program. Growing season for this study was defined as the period between the last average occurrence of 32° F in the spring and the first average occurrence of 32° F in the fall. The program eliminated all nongrowing season data and ranked the Table 1 Hydrologic Zones Occurring in Bottomland Forests of the Southeastern United States 12 Zone 2 Name Semipermanently to permanently inun- dated or saturated Regularly inun- dated or saturated Seasonally inun- dated or saturated Irregularly inun- dated or saturated 6 Intermittently inundated or saturated Typical Duration15 Inundation/Saturation Frequency3 (percent) Annual (1 year frequency) >75-100 90 to 100 years/100 years 51 to 90 years/100 years (>l-year to 2 -year frequency) >25-75 51 to 90 years/100 years (>l-year to 2-year frequency) >12.5-25 11 to 50 years/100 years (well >5-12.5 drained) (>10 years - 2-year frequency) 1 to 10 years/100 years (poorly drained) (100 years, 10-year frequency) 1 to 10 years/100 years <5 (100 years, 10-year frequency) Source : Adapted from Larson et al . (1981). a Although typical inundation/saturation frequencies are provided for each zone, almost any frequency could be associated with any duration of inundation/saturation. Therefore, only duration of inundation/soil saturation was used to determine hydrologic zones . b Duration based on the growing season. remaining daily readings during the period of record from highest to lowest flow (or stage). Elevations corresponding to the 75, 25, 12.5, and 5 percent durations of inundation were computed. Because the resulting elevations did not include the period during which the soils remain saturated after a period of inundation, saturation effects were integrated. A general description of the soil series occurring in each zone of the study site was obtained from Soil Conservation Service (SCS) 13 county soil surveys. An estimated range of permeabilities for the top 30 cm of the soil profile (i.e., defined for this study as the effective root zone) was determined. This range approximated the period required for the root zone to become saturated after inundation. The slowest value in the range of permeabilities was used to determine the minimum duration of inundation required to saturate the soil. A second range of soil permeabilities between the 30-cm and 90-cm depth was determined. The slowest permeability value of the soil profile between 30 and 90 cm was used to estimate the time required for draining of the root zone after dewatering. A mean daily transpiration factor for floodplain forests of 5.6 mm (Brown 1981) also was incorporated for computing desaturation. Permeability and transpiration coefficients were provided as pro- gram input, and new flow (or stage) values for hydrologic zone bound- aries were derived that reflected both inundation and soil saturation. This iterative process required a computer search. The computer program added the days of saturation to the days of inundation, and the output was flow (or stage) values that represented the estimated boundary of each hydrologic zone, based on inundation and saturation. The gauge elevation was added to the stage for each zone to obtain the mean sea level elevation at the gauge. When the site was not immediately adja- cent to the gauging station, the change in water surface elevation between the study area and the gauging station was determined using the best available water surface profile data. Appendix B explains how the computer program analyzes the hydrology data to produce zone boundaries. 14 Site Preparation and Data Collection A temporary benchmark was established at each of the 17 sites by surveying from a permanent benchmark. A reconnaissance of the area was conducted for suitable sites, and mean sea level elevations for each hydrologic zone boundary were surveyed along the topographic gradient. The contours of each hydrologic zone boundary within the site were marked with surveyor flags. Fifty-five hydrologic zones were estab- lished on the 17 study sites. Sampling methods were adapted from meth- ods described by Whittaker (1973), except where noted. Sample plots were established parallel to the hydrologic zone boundary (Figure 3) . Plots were positioned on the downslope side of the boundary with at least a 5-m buffer between the sample plots and the upper and lower boundary of the hydrologic zone. A belt transect (20 m wide by 40 m long) containing 10 sample subplots (8 m by 10 m) was established within each zone. Small soil pits in each sample plot were dug with a tile spade to a depth necessary to identify the soil series. In all cases a county soil survey was used to identify the mapped soil series, and information was obtained to verify the soil series on site. Assistance from the local SCS office was used to determine the correct soil series and soil permeability coefficients for each zone sampling site. Vegetation was sampled by vegetative layer. All trees in each sample plot were identified by species and the diameter at breast height (1.5 m) of individuals having a diameter of greater than or equal to 7.5 cm was measured and recorded to the nearest whole centimeter. ®BM 15 ZONE 6 LEGEND <2)BM BENCHMARK <2>TBIYI TEMPORARY BENCHMARK ^Tp HYDROLOGY GAUGE SURVEYED CONTOUR LINE SAMPLE PLOTS Figure 3. Representation of a typical research site 16 All saplings and shrubs (woody plants less than 7.5 cm in diam- eter, but greater than 1.0 m in height, excluding vines) in each sample plot were identified by species, and the height class of each individual was recorded. Saplings or shrubs with more than one stem clustered from a single root system were counted as individuals only when separation occurred at or below ground level. The following height classes were used: Class 1 = 1.0 to 2.0 m, Class 2 = 2.1 to 3.0 m, Class 3 = 3.1 to 4.0 m, Class 4 - 4.1 to 5.0 m, and Class 5 = >5 . 0 m. All climbing woody vines greater than 1.0 m in height in each sample plot were identified by species, the stems of each species coun- ted, and the height class of the highest individual on each tree or sapling/shrub recorded. The following height classes were used: Class 1 = 1.0 to 3.0 m, Class 2 = 3.1 to 6.0 m, Class 3 = 6.1 to 12.0 m, and Class 4 = >12.0 m. Vines were recorded when any portion of the plant occurred in, or overhung, the plot. Individual stems were recorded when separation from the root system occurred at or below ground level. Percent cover was estimated for each species of herb and woody seedling (greater than 1.0 m in height) rooted in the plot in two randomly located 1.0-m2 quadrats in each subplot using the Daubenmire (1968) cover class method. Analyzing Vegetation Data Importance values for species in all vegetation layers except the herbaceous layer were calculated by adding values for relative density, relative frequency, and relative dominance. Importance values for her- baceous species were calculated by summing relative frequency and 17 relative dominance. Importance values were used to determine the FTI number for each species. When species could not be positively identified in the field, voucher specimens were collected and later identified. Species nomen- clature was determined using the National List of Scientific Plant Names (U.S. Department of Agriculture 1982). Calculating Species FTI Numbers Changes in composition of biotic communities along environmental gradients can be addressed with several statistical techniques, the most notable being gradient analysis (Whittaker 1978) . Gradient analysis can take several different forms depending on the objective of the analysis. Inferring environmental values (e.g., hydrologic zones) from vegetative species composition is called a "calibration problem" by Ter Braak and Prentice (1988) and is the appropriate approach for this study. One method of calibration is to use weighted averaging (WA) to estimate environmental factors at sites based on species optima. If a species exhibits a unimodal distribution with respect to an environmen- tal variable, its occurrence is concentrated around the peak of this function (Ter Braak and Prentice 1988). Species with similar optima will naturally tend to occur together. Therefore, an intuitive estimate of the environmental factor of a site is the average of the optima for the species present. The FTI numbers represent weighted averages of species occurrance . Two additional statistical methods of calibrating an environment (hydrologic zone) with vegetation, recommended by Ter Braak and Prentice 18 (1988), are cluster analysis and discriminant function analysis. These two methods were applied to test the reliability of the FTI numbers. FTI numbers were calculated for each species occurring in each vegetation layer. A species could have three different FTI numbers at a given site, depending on its growth form. For example, Quercus nigra would have three different FTI numbers when present on a site as a tree, sapling, and seedling. Species FTI numbers for each site were computed by the following formula: S-i(j • IV,..) where i = the ith species ja = 2.5, 3. 5... 6. 5 (hydrologic zone) IVij = importance value for species i in the hydrologic zone j After species FTI numbers were computed for all species in all sites, the average FTI number (FTI) for each species across all sites was cal- culated using the following formula: E.ni. FTI,, J=i ij FTIi = n< Because vegetation was sampled between zone boundaries, midrange zone numbers (e.g., 2.5 for Zone 2, 3.5 for Zone 3, etc.) for zones were used in calculating FTI numbers. 19 where i = the ith species j = sites 1 to 17 FTIij = FTI number of species i at site j nL = number of sites at which species i occurred RESULTS AND DISCUSSION Flood Analysis of Study Sites Twenty years of hydrologic data were used for all sites except sites 4, 5, and 14. Sites 4 and 5 had a 19-year hydrologic record, and site 14 had a 12-year record. Calculations of change in water surface elevation between the gauging station and site was necessary for sites 1, 2, 3, 5, 9, 12, and 17. All other sites were adjacent to the gauging station and did not require adjustments. Hydrologic analyses of sites 11 and 12 at the Apalachicola River were verified using information from another study (Leitman, Sohm, and Franklin 1984) . The hydrologic records for all sites were analyzed by season for five-year increments. In all cases, variation in flow through time was determined to be within normal seasonal and annual fluctuations. There- fore, it was assumed that the hydrologic record reflected normal condi- tions (i.e., no major drainage projects during the period of record had significantly impacted the plant community structure) . The hydrologic data also were analyzed to determine annual flood frequency and duration for each site. The boundaries between Zones 2 and 3 and between Zones 3 and 4 are flooded virtually every year (Table 2). The boundary between Zones 4 and 5 is flooded at least every other year, and the boundary between Zones 5 and 6 is flooded from once 20 Table 2 Annual Flood Frequency (Percent of Years in Which Boundary Is Exceeded at Least Once during Growing Season for More than 7 days) for Zone Boundaries 21 Zone Boundaries Site 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 2-3 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 100 3-4 90 90 100 94 85 96 100 100 100 100 100 100 100 100 90 100 100 4-5 70 70 85 61 55 77 92 96 96 83 95 95 90 92 85 100 90 5-6 10 10 25 20 20 20 75 60 65 70 70 70 65 67 55 70 70 22 in 20 years to 3 out of every 4 years. Similar findings were reported by Clark and Benforado (1981) and Roelle et al . (1987). In general, the average duration of a flood event by site ranged from 3 months to longer than 5 months for the Zone 2-3 boundary, from 3 weeks to greater than 2 months for the Zone 3-4 boundary, from 1 week to 1 month for the Zone 4-5 boundary, and from less than 1 day to 10 days for the Zone 5-6 boundary (Table 3) . As an example, a hydrograph of site 3 at Steele Bayou (Appendix C) for 20 years of data shows that flooding conditions vary greatly from one season to another and from one year to another. Also, flooding during the growing season usually occurs during early spring and is usually continuous with the nongrowing season flooding. Flooding rarely occurs late in the growing season. Unusual events such as the 1973 flood can greatly affect the calculated value of average events. For example, if the data for 1973 were excluded, the average duration per flood event at site 3 is reduced to 37 days, 25 days, 9 days and less than 1 day, from 143, 30, 14, and 6 days, for Zone boundaries 2 through 5, respectively (Table 4). Such an event has an especially large effect on average duration even over a 20-year period, especially in the higher zones . Vegetation Data Vegetation data were collected at each of the 55 hydrologically defined zones for the 17 study sites (Appendix D) . Eleven stands were sampled in Zone 2, 15 stands in Zone 3, 14 stands in Zone 4, 4 stands in Zone 5, and 11 stands in Zone 6. The total possible number of stands 23 Table 3 Average Annual Duration of Flood Events (days) for Zone Boundaries Zone Boundaries Site 2-3 3-4 4-5 12 5-6 1 119 21 <1 2 119 21 12 <1 3 143 30 14 6 4 92 41 19 7 5 93 51 24 8 6 152 44 22 9 7 153 50 25 10 8 139 20 15 <1 9 139 20 15 <1 10 141 32 14 3 11 143 20 16 8 12 143 20 18 6 13 162 29 15 8 14 159 53 24 5 15 162 52 25 9 16 174 58 8 <1 17 198 66 31 10 Study Area Average 143 + 26 37 + 18 18 + 6 6 ± 3 24 that could have been sampled for each zone was 17. Some zones were unsuitable for analysis due to disturbance or because they were too narrow to support sampling areas (Appendix A) . Zone 5 was especially susceptible to disturbance. In some cases Zone 5 was cleared because it was dry enough to be farmed during the growing season. In other cases, Zone 5 was too narrow because it was located near the toe of a slope. Therefore, Zone 5 was sampled only at sites 3, 11, 12, and 14. FTI numbers were calculated for 74 tree species, 118 species of saplings and shrubs, 31 species of woody vines, and 268 species of herbs and woody seedlings, representing 312 different plant species. Because some species occurred in more than one vegetative layer, the total num- ber of species FTI numbers exceeded the total plant species identified. Appendix E contains a listing by stratum of the plant species identified in this study and the calculated FTI numbers with standard deviations provided for each species. FTI numbers were calculated for all plants identified by species in this study, regardless of how frequently they occurred in the study. The FTI numbers calculated for species with few occurrences may be sus- pect. FTI numbers for some of the more commonly occurring species in the study are listed in Table 4. These species can generally be found along the moisture gradient in any bottomland hardwood forest in the southeastern United States in roughly the order from wettest to driest community as presented. Some species, such as Nuttall's oak, are common in only part of the region, and not all species can be expected on the same site due to the species' response to other environmental gradients. 25 Table 4 FTI Numbers of Species Commonly OccurrinE in Bottomland Forests Scientific Name Nyssa aquatica Salix nigra Fraxinus caroliniana Taxodium distichwn Planera aquatica Acer drummondii Forestiera acuminata Gleditsia aquatica Carya aquatica Quercus lyrata Quercus laurifolia Betula nigra Acer rub rum Ilex decidua Fraxinus pennsylvanica Ulmus americana Quercus nuttallii Quercus phellos Acer negundo Celtis laevigata Carpinus caroliniana Common Name Trees Water tupelo Black willow Pop -ash Bald cypress Water elm Drummond red maple Swamp privet Water locust Water hickory Over cup oak Laurel oak River birch Red maple Possumhaw holly Green ash American elm Nuttall's oak Willow oak Box elder Sugarberry American hornbeam FTI Number 2.62 2.83 2.87 2.97 3.12 3.48 3.48 3.50 3.54 3.73 3.89 4.01 4.21 4.35 4.44 4.46 4.50 4.81 4.83 4.84 4.84 26 Table 4- -continued Scientific Common FT I Name Name Number Liquidambar styraciflua Sweetgum 5.03 Platanus occidentalis Sycamore 5.18 Nyssa sylvaCica Black gum 5.27 Carya illinoensis Pecan 5.57 Quercus nigra Water oak 5.73 Horus rubra Red mulberry 5.75 Ilex opaca American holly 5.79 Ulmus alata Winged elm 6.43 Pinus taeda Loblolly pine 6.41 Quercus alba White oak 6.50 Sassafras albidum Sassafras 6.50 Fagus grandifolia American Beech 6.50 Cornus florida Flowering dogwood 6.50 Ostrya virginiana American hophornbeam 6.50 Quercus stellata Post oak 6.50 Quercus falcata Southern red oak 6.50 Carya tomentosa Mockernut hickory 6.50 Sapl: Lngs and Shrubs Salix nigra Black willow 2.83 Itea virginica Virginia willow 2.83 Planera aquatica Water elm 3.01 Cephalanthus occidentalis Buttonbush 3.13 Sty rax americana Snowbell 3.41 27 Table 4- -continued Scientific Common Name Name Forestiera acuminata Swamp privet Cyrilla racemi flora Titi Celtis laevigata Sugar berry Crataegus viridis Green hawthorn Carya illinoensis Pecan Platanus occidentalis Sycamore Acer negundo Box elder Magnolia grand if lor a Southern magnolia Liquidambar styraciflua Sweetgum Cornus drummondii Rough leaf dogwood Vaccinium elliottii Elliott blueberry Quercus nigra Water oak Sambucus canadensis Elderberry Halesia diptera Silverbell Morus rubra Red mulberry Cercis canadensis Redbud Vaccinium arboreum Farkleberry Gleditsia triacanthos Honey locust Quercus alba White oak Cornus florida Flowering dogwood Ilex vomitoria Yaupon Woodv Vines Ipomoea wrightii Morning glory FT I Number 3.57 3.72 4.37 4.46 5.00 5.05 5.20 5.43 5.52 5.69 5.82 5.92 5.95 6.09 6.25 6.37 6.45 6.50 6.50 6.50 6.50 2.50 28 Table 4- -continued Common FTI Name Number Walter's greenbriar 3.05 Ladies' eardrops 3.58 Peppervine 3.94 Trumpet creeper 4.05 Cat grape 4.07 Star jasmine 4.18 Riverbank grape 4.27 Bamboo -vine 4.33 Carolina moonseed 4.37 Rattan vine 4.55 Saw greenbriar 4.75 Poison ivy 4.82 Common greenbriar 5.18 Muscadine grape 5.71 t Virginia creeper 5.93 Japanese honeysuckle 6.50 Carolina jessamine 6.50 Herbs Primrose -willow 2.50 Yellowcress 2.50 Creeping burhead 3.00 Lowland white aster 3.04 Small-spike falsenettle 3.34 Scientific Name Smilax waiter i Brunnichia cirrhosa Amplelopsis arborea Campsis radicans Vitis palmata Trachelospermum difforme Vitis riparia Smilax laurifolia Cocculus carolinus Berchemia scandens Similax bona-nox Toxicodendron radicans Smilax rotundifolia Vitis rotundifolia Parthenocissus quinquefolia Lonicera japonica Gelsimium sempervirens Ludwigia decurrens Rorippa islandica Echinodorus cordifolius Aster simplex Boehmeria cylindrica 29 Table 4- -continued Scientific Common FT I Name Name Number Spermacoce glabra Smooth buttonweed 3.50 Saururus cernuus Lizard- tail 3.65 Leersia lenticularis Catchfly grass 3.67 Justicia ovata Waterwillow 3.83 Urtica chamaedryoides Nettle 4.42 Clematis virginiana Virgins bower clematis 4.57 Eupatorium rugosum White snakeroot 4.67 Cocculus carolinus Snailseed 4.78 Viola missouriensis Missouri violet 4.83 Hypericum hypericoides St. Johnswort 5.25 Mitchella repens Partridge berry 5.32 Arundinaria gigantea Giant cane 5.34 Vitis rotundifolia Muscadine 5.89 Gelsemium sempervirens Carolina jessamine 6.31 Sanicula canadensis Black snakeroot 6.39 Galium aparine Catchweed bedstraw 6.50 Carex albolutescens Sedge 6.50 Cynanchum laeve Cynanchum 6.50 Lonicera sempervirens Trumpet honeysuckle 6.50 30 Examples of FTI numbers by life stage for woody species are pre- sented in Table 5. Although differences between FTI numbers for tree and sapling life stages are not great, FTI numbers for saplings tend to be slightly higher than for trees because saplings are generally more sensitive to flooding than trees. The sapling life stage also tends to occur in a broader range of zones than the tree life stage due to tem- poral variations in selective pressures (e.g., competition and response to flooding) . FTI numbers of seedlings have little value in determining hydrologic zones because they only reflect seed dispersal potential and germination wherever seedbed conditions are favorable. For example, seedlings of least tolerant species (e.g., Sassafras albidum) Table 5 Variations in Species Flood Tolerance Index Numbers According: to Life Stage Scientific Name Tree 2.97 Sapling 3.33 Seedling Taxodium distichum 3.09 GlediCsia aquatica 3.50 3.15 3.27 Carya aquatica 3.54 3.70 3.69 Quercus lyrata 3.73 3.99 3.80 Fraxinus pennsylvanica 4.44 4.27 4.00 Quercus nuttallii 4.50 4.50 4.50 Acer negundo 4.83 5.20 5.58 Celtis laevigata 4.84 4.37 4.77 Liquidambar styraci flua 5.03 5.52 4.87 Quercus nigra 5.73 5.92 5.85 Sassafras albidum 6.50 6.50 6.07 31 occasionally occur in lower zones (Table D-3, Zone 3), but the individuals do not survive to maturity unless the hydrologic regime is drastically altered. The average importance value was plotted for each species in every zone where it occurred in the study. Although a species can be expected to occur in a number of zones, many had a peak occurrence in a particu- lar zone (Figure 4). However, because this study did not analyze a continuous gradient, FTI numbers were calculated from mean importance values across the entire study and do not necessarily represent the maximum in ecological amplitude for a species. Three general species distribution patterns are shown in Figure 4. The first pattern includes species such as water tupelo (NYAQ) and bald cypress (TADI) , in which mean importance value is greatest in Zone 2 and the species no longer occurs after either Zone 3 or 4. This pattern is indicative of species having the strongest competitive advantage in areas of greatest duration of inundation/soil saturation. The second pattern is typified by species such as water oak (QUNI) , loblolly pine (PITA) , sassafras (SAAL) , and white oak (QUAL) , in which the greatest mean importance value occurs in Zone 6 (uplands) and decreases from Zone 5 to 2 . Hence they have a stronger competitive advantage in areas where inundation/soil saturation is less than 5 percent of the growing season. However, some of these species (e.g., water oak and loblolly pine) may occasionally occur as dominants in wetlands. The third pattern is typi- fied by species having the greatest mean importance values in Zones 3, 4, and 5. Species in this group sometimes occur as dominant species in either Zone 2 or 6 , but are best adapted for occurrence at some point in Zones 3, 4, and 5. Species having the greatest mean importance values 32 100 90 80 70 60 UJ H 50 g 40 a. 30 20 10 0 I I I I I I A ' 1 - V* X / - - V \ X \// %/ \ - J0Z \ / T ^0^/ - ^c I ^ V£Rpe^£ ^\ WET ZONE 2 3 4 5 HYDROLOGIC ZONES DRY Figure 4. Ecological amplitude of some commonly occurring species; CAAQ: Carya aquatica; F0AC: Forestiera acuminata; FRPE: Fraxinus pennsylvanica; LIST: Liquidambar styraciflua; NYAQ: Nyssa aquatica; PITA: Pinus taeda; QUAL: Quercus alba; QULY: Quercus lyrata; QUNI: Quercus nigra; SAAL: Sassafras albidum; TADI : Taxodium distichum; and ULAM: Ulmus americana 33 in Zone 3 are overcup oak (QULY) and bitter pecan (CAAQ) , while American elm (ULAM) and sweetgum (LIST) develop the greatest mean importance values in Zones 4 and 5, respectively. Other systems have been developed to identify the degree of wet- ness for which a species is best adapted (e.g., Hook 1984, Reed 1988). These systems used qualitative descriptions, such as "most tolerant" or "obligate hydrophyte," and are based primarily on the literature or "expert evaluations" and not on a single coordinated study. Species evaluations were also made by life forms in this study, and no such dis- tinction was made in the other two systems. However, all three systems, including the FTI numbers, were developed to identify a degree of wet- ness for which a species is best adapted. They all have five categories that vary from wettest to driest. Therefore, an obvious comparison would be to compare species having an FTI integer of 2, with the most tolerant and obligate designation, an FTI integer of 3, with the highly tolerant and facultative wet description, etc. Species identified in this study are listed along with correspond- ing FTI numbers and ratings of those species from the systems developed by Hook and Reed in Appendix E. Among selected tree species shown in Table 6, all species having an FTI number from 2 to 4 are obligate plants (OBL) in the National List of Plant Species that Occur in Wet- lands: Southeast (Region 2) (Reed 1988) and have a water-logging- tolerance rating of most or highly tolerant (Hook 1984) . All listed species except Pinus taeda that have an FTI number of 6 to 6.5 are fac- ultative upland (FACU) species (Reed 1988) and are rated by Hook (1984) as the least- tolerant species. P. taeda (FTI-6.41) has an indicator status of facultative (FAC) and is rated by Hook as moderately tolerant 34 Table 6 Comparison of Three Water-Tolerance Ratings for Selected Bottomland Forest Tree Species Species Nyssa aquatica Salix nigra Fraxinus caroliniana Taxodium distichum Forestiera acuminata Gleditsia aquatica Carya aquatica Quercus lyrata Betula nigra Diospyros virginiana Acer rubrum Fraxinus pennsylvanica Ulmus americana Quercus phellos Acer negundo Carpinus caroliniana Celtis laevigata Liquidambar styraciflua Carya illinoensis Pinus taeda Cornus florida Fagus grandifolia Quercus alba Sassafras albidum NWIb Waterlogging FTIa ± Standard Status Tolerance Deviation Region 2 Rating Group0 2.62 + 0.20 OBL Most tolerant 2.83 + 0.58 OBL Most tolerant 2.87 + 0.41 OBL Most tolerant 2.97 + 0.61 OBL Most tolerant 3.48 + 0.50 OBL Most tolerant 3.50 + 0.00 OBL Highly tolerant 3.54 + 0.34 OBL Highly tolerant 3.73 + 0.68 OBL Highly tolerant 4.01 + 1.73 OBL Moderately tolerant 4.13 + 0.82 FAC Moderately tolerant 4.21 + 0.68 FAC Moderately tolerant 4.44 + 0.67 FACW Moderately tolerant 4.46 + 0.62 FACW Moderately tolerant 4.81 + 1.07 FACW Moderately tolerant 4.83 + 0.47 FACW Moderately tolerant 4.84 + 0.61 FAC Weakly tolerant 4.84 + 0.56 FACW Weakly tolerant 5.03 + 0.65 FAC+ Moderately tolerant 5.57 + 1.01 FAC+ Weakly tolerant 6.41 + 0.14 FAC Moderately tolerant 6.50 + 0.00 FACU Least tolerant 6.50 + 0.00 FACU Least tolerant 6.50 + 0.00 FACU Least tolerant 6.50 + 0.00 FACU Least tolerant Mean for all study sites. Taken from Reed (1988) (see Appendix A) . Taken from Hook (1984) (see Appendix B) . 35 of waterlogged soils. All species except Betula nigra that have an FTI number of 4 to 6 are facultative wet (FACW) or FAC and are rated as moderately or weakly tolerant. B. nigra, an obligate species, occurs on well-drained soils, often on natural berms . FTI numbers were computed only for bottomland forests and do not reflect occurrence in other wet- land types (e.g., pocosins and Carolina bays); thus, slight deviations from the above pattern should be expected for some species. Also some species may have genetic variants that possess varying degrees of flood tolerance. Hook et al. (1988) reported that interspecific variation in tolerant to waterlogging exists in loblolly pine (Pinus taeda) . This may also be true of other species. Although the system using the National Wetlands Inventory (NWI) indicator status (Reed 1988) does not allow comparison of hydrologic definitions, the actual average duration of flooding in this study com- pared with Hook's (1984) waterlogging tolerance rating definitions yields strong agreement. For example, assuming an average 225 -day grow- ing season, the most tolerant rating can be defined as approximately 200 days (Appendix E) . The boundary between Zones 2 and 3 (which theo- retically would be slightly drier because the designation most tolerant is best compared to Zone 2, not the boundary between Zones 2 and 3) has a duration of inundation/soil saturation that ranges from 92 to 198 days. Highly tolerant ranges from 30 to 90 days versus 20 to 66 days for the Zone 3-4 boundary. Weakly tolerant ranges from 1 to 4 weeks, while the Zone 4-5 boundary ranges from 8 to 31 days. "Least tolerant" is defined as waterlogging for a few days, but usually less than 2 percent of the growing season. Using the 225-day growing season, 36 this could be assumed to be 3 or 4 days. The boundary between Zones 5 and 6 varied from less than 1 to 10 days. Weighted Averaging Weighted average estimates of species optima in this study were calculated as FTI values as previously described, with importance values used as the indicator of species abundance at each site. The use of the weighted averaging approach requires that a number of conditions be met, including: (1) species exhibit unimodal abundance distributions, (2) species optima are equally spaced along the environ- mental variable, (3) species have equal tolerances of the environmental variable, and (4) species have equal maximum values for the environmen- tal variable. Strict adherence to some of these conditions is not always possible. Additional considerations should also be noted. Species-rich samples should not occur at one end of the gradient. Environmental tolerances of species should not vary substantially. The standard deviation of FTI numbers is an estimate of tolerance in this analysis. Species with narrow tolerances have low FTI numbers and standard deviations and those with wide tolerance have high standard deviations . Some of the aforementioned conditions are not strictly met in this study. Although species richness was fairly even across the hydrologic gradient, few species had a peak abundance in Zone 5, perhaps because Zone 5 was undersampled (n = 4) relative to the other zones. The condi- tion of a unimodal abundance distribution is upheld for tree species with peaks in hydrologic Zones 2, 3, and 6, but several species (e.g., Ilex opaca, Quercus nigra, Nyssa sylvatica , and Liquidambar styraciflua) 37 exhibit bimodal distributions with peaks in Zones 4 and 6. Again, this pattern may have been influenced by fewer samples being taken in Zone 5. The condition of equal tolerances is also violated somewhat. For instance, those species with abundance peaks in Zones 2 and 6 have nar- row tolerances; whereas, those that commonly occur in Zone 4 are also fairly common in other hydrologic zones as well. However, these devia- tions from the conditions for weighted averaging analysis do not neces- sarily render the FTI method invalid. Additional analyses were applied to help determine its validity. Statistical Analysis of the Vegetation Data To evaluate a method of identifying hydrologic zones based on vegetative associations, the analysis must be based on common species occurring within the region of interest. Species that occur infre- quently may be excellent indicators of hydrologic conditions when pres- ent, but their limited abundance makes evaluating their usefulness in determining a hydrologic zone difficult. For this reason, species occurring relatively infrequently in this study were not used in testing the validity of the weighted averages (FTI numbers) . More than half of the 74 tree species (n = 44) recorded had 20 or fewer individuals throughout the study region and, therefore, were not used in the analy- ses. The remaining 30 tree species accounted for 90.4 percent of the individual trees. One hundred and eighteen species of saplings and shrubs were recorded. Only those 29 species which represented at least one percent or more of the total and together accounted for 78 percent of the total saplings and shrubs data set were used in the statistical analysis. A total of 31 vine species were recorded; the 20 more common 38 species accounted for 96.5 percent of all individuals and were the spe- cies analyzed statistically. The herbaceous ground-cover data set con- tained 268 species. Most occurred only rarely in this study, and only 30 species whose abundances equaled or exceeded 1 percent of the total herb individuals were included in the analysis. These species accounted for 48.9 percent of all individuals in this vegetative category. Cluster Analyses A method of calibration suggested by Ter Braak and Prentice (1988) is cluster analysis in which an environmental value (hydrologic zone) is predicted through use of species abundance indicators (relative fre- quency) . In addition to relative frequency, cluster membership was used as a predictor of hydrologic zone. Species were clustered according to similar abundance distributions across the hydrologic gradient. Cluster analysis was used to group the 30 tree species into five clusters based on the five hydrologic zones. Results (Figure 5) show that, with the exception of chinaberry (MEAZ) and deciduous holly (ILDE), five distinct groups can be discerned. Table 7 gives the rela- tive frequencies of occurrence of each species in each hydrologic zone (species are grouped by cluster) . Inspection of the data reveals why chinaberry and deciduous holly did not group readily. Chinaberry is the only species that occurs almost exclusively in Zone 5. Its occurrence is also restricted to a single site. The distribution of deciduous holly peaks in Zones 2 and 3 , but not to the extent of other common species in these zones. Because chinaberry and deciduous holly did not group readily, they were eliminated from further analysis. 39 DISTANCES MEAZ OSVI QUST CATO PITA ILOP QUNI NYSY LIST CACA QUPH CELA ULAM FRPE ACNE ACRU QULA CAAQ QULY DIVI NYBI ACDR FOAC NYAQ TADI FRCA NYOG SANI PLAQ ILDE 0.000 1 1 1.000 0.843 0.000 0.000 0.003 0.112 0.018 0.108 0.019 0.124 0.027 0.025 0.050 0.131 0.067 0.047 0.152 0.007 0.003 0.003 0.025 0.008 0.001 0.427 0.000 0.004 0.006 0.000 0.021 0.121 Figure 5. Cluster diagram for trees. Distance metric is 1-Pearson correlation coefficient, single linkage method (nearest neighbor) 40 Table 7 Relative Frequencies in Each Hydrologic Zone of Tree Species Used in the Statistical Analyses: Groupings of Species Correspond to Cluster Membership Hvdrc ilogic Zone Species 2 85.0 3 15.0 4 0 5 0 6 Cluster NYAQa 0 TAD I 82.7 15.6 1.8 0 0 FRCA 79.1 20.9 0 0 0 1 NYOG 93.9 6.1 0 0 0 SANI 96.5 3.5 0 0 0 PLAQ 67.7 31.6 0.7 0 0 QULA 0.9 70.1 26.2 0 2.8 CAAQ 0.8 80.0 19.2 0 0 QULY 6.5 72.9 20.0 0.6 0 DIVI 0 70.6 20.6 2.9 5.9 2 NYBI 15.3 84.7 0 0 0 ACDR 7.1 88.1 4.8 0 0 FOAC 3.6 94.0 2.4 0 0 CACA 0.6 2.8 63.7 17.0 15.9 QUPH 0 8.6 74.1 10.3 6.9 CELA 2.0 9.8 62.8 21.6 3.9 ULAM 1.9 21.2 50.0 25.0 1.9 3 FRPE 0 45.2 42.9 11.9 0 ACNE 0 31.1 53.3 4.4 11.1 ACRU 6.8 38.4 52.0 1.4 1.4 ILOP 0 0 41.7 0 58.3 QUNI 0 1.6 32.8 7.4 58.2 4 NYSY 0 0 55.9 5.9 38.2 LIST 0.3 6.0 45.4 14.9 33.3 OSVI 0 0 0 0 100 QUST 0 0 0 0 100 5 CATO 0 0 0 0 100 PITA 0 0 7.3 0 92.7 Represents the first two letters of the plant genus and the first two letters of the species name, i.e., NYAQ stands for Nyssa aquatica. Species codes for all species are identified in Appendix E. 41 A closer inspection of Table 7 indicates that the five clusters have modal peaks that correspond, with varying amplitude, to the five hydrologic zones. Trees in cluster 1 are found almost exclusively in Zones 2 and 3. Likewise, trees in cluster 2 occur most commonly in Zone 3. Cluster 5 had the most restrictive distribution, with three of the four species occurring exclusively in Zone 6. The remaining clus- ters, 3 and 4, had less distinctive modes, but exhibited greater distri- butions in Zones 4 and 6. Cluster analysis on the entire sapling and shrub data set did not produce distinct groupings (Figure 6). Silver bell (HADI) was not eval- uated further because it occurred at a single site. The data set was split into the two sapling and shrub (bush) components, and cluster analysis was recalculated on each component. The saplings alone (n - 17) grouped more distinctly (Figure 7) than the shrubs (n = 11) (Figure 8). Shrubs, therefore, were not used in any further analyses. Saplings alone grouped into five distinct groups but not as strongly as trees. The relative frequency of occurrence and cluster membership of the sapling species are given in Table 8. The vine data produced five clusters, morning glory (IPWR) being a single species cluster corresponding to Zone 2 (Figure 9) . Two vine species poison ivy (TORA) and trumpet creeper (CARA) were not included in larger clusters, largely because of their apparent wide tolerances of hydrologic conditions. Table 9 depicts the relative abundance distribu- tions of each vine species across the hydrologic gradient within a clus- ter and illustrates a pattern similar to but not as strong as that of the tree species. 42 DISTANCES 0.000 ULAM CELA ILDE ULAL HADI SYTI VAAR CATO MYCE QUNI NYSY LIST ILOP CACA ACNE VAEL OSVI CODR CRVI QULA FRPE FOAC SANI FRCA PLAQ CRAE CEOC ACRU STAM 0.500 0.126 0.142 0.152 0.013 0.002 0.000 0.000 0.000 0.016 0.029 0.027 0.115 0.054 0.164 0.126 0.016 0.017 0.301 0.309 0.145 0.126 0.166 0.064 0.074 0.083 0.162 0.183 0.318 Figure 6. Cluster diagram for saplings and shrubs. Distance metric is 1-Pearson correlation coefficient, single linkage method (nearest neighbor) 43 DISTANCES 0.000 SAMI FRCA OSVI CODR ULAM CELA ULAL QUNI CATO NYSY LIST LOP CACA ACNE ACRU QULA FRPE 2.000 0.064 1.112 0.017 0.191 0.125 0.152 0.023 0.021 0.029 0.027 0.115 0.054 0.164 0.396 0.182 0.147 Figure 7. Cluster diagram for saplings alone. Distance metric is 1-Pearson correlation coefficient, single linkage method (nearest neighbor) 44 DISTANCES 0.000 ILDE CRVI VAEL SYTI VAAR MYCE ST AM PLAQ FOAC CREA CEOC 1.000 0.300 0.613 0.736 0.000 0.000 0.747 0.368 0.169 0.167 0.161 Figure 8. Cluster diagram for shrubs alone. Distance metric is 1 Pearson correlation coefficient, single linkage method (nearest neighbor) 45 Table 8 Relative Frequency of Occurrence i of Each Sanlins Species in the Hvdrologic Zones along with ' rheir Cluster Memberships Species 2 99.4 3 0.6 4 0.0 5 0.0 6 Cluster SANI 0.0 1 FRCA 61.0 21.2 2.7 0.0 15.1 ACRU 1.4 35.8 35.1 4.6 23.0 FRPE 0.0 50.0 21.9 21.9 6.3 2 QULA 4.7 53.3 36.7 3.6 1.8 ULAM 0.0 2.0 42.1 32.2 23.7 3 CELA 0.0 13.1 47.5 19.5 19.9 OSVI 0.0 0.0 0.0 71.2 28.8 4 CODR 0.0 0.0 1.5 80.0 18.5 ULAL 0.0 0.0 0.8 17.7 81.5 QUNI 0.0 0.4 14.2 11.1 74.2 CATO 0.0 0.0 0.0 0.0 100.0 NYSY 0.0 0.0 24.7 22.7 52.6 5 LIST 0.0 5.4 24.2 14.6 55.8 ILOP 0.0 0.0 55.5 5.0 39.5 CACA 0.2 1.4 40.6 15.7 42.0 The herb data also did not cluster distinctly (Figure 10) . Inspection of the abundance distributions of the herbaceous species reveals an absence of strong unimodal peaks that correspond to hydro - logic zones; therefore, this vegetative type is less useful in classifi- cation. The lack of significant clustering of the shrub and herbaceous data sets was consistent with the results of ordination attempts of the ground layer at Neches River sites by Mohler (1979). Mohler concluded that the herbs and low shrubs were relatively unimportant components of the forest when compared to the trees. Clearly, the low shrubs and herbaceous species are more prone to bias due to successional processes and vegetative change caused by local disturbances which may be unrelated to the environmental gradient. 46 0.000 DISTANCES GESE CLLI ARTO BICA VIRO TORA CARA VIPA TRDI BROV VIRI SMLA COCA SMWA AMAR PAQU SMRO BESC SMBO IPWR 2.000 0.000 0.000 0.140 0.144 0.418 0.176 0.133 0.009 0.029 0.097 0.021 0.017 0.026 0.009 0.184 0.108 0.058 0.074 1.140 Figure 9. Cluster diagram for vines. Distance metric is 1-Pearson correlation coefficient, single linkage method (nearest neighbor) 47 Table 9 Relative Fre quencies in Each Hydrc iloeic Zone of the Vine Sp ecies Used in Statistical Analyses Hydroloeic Zone Species' 2 100 3 0 4 0 5 0 6 Cluster IPWR 0 1 VI PA 0 64.1 30.5 5.5 0 TRDI 2.1 70.2 26.9 0 0.8 2 BROV 13.4 65.2 17.8 3.6 0 VIRI 1.0 33.3 43.8 7.6 14.3 SMLA 1.8 47.4 45.6 0 5.3 COCA 0 41.9 52.9 0 5.2 3 SMWA 9.1 36.4 53.8 0 0.7 AMAR 13.2 37.2 46.3 0 3.3 PAQU 1.5 55.9 2.2 40.4 SMRO 0 2.4 71.4 7.1 19.1 4 BESC 4.6 13.7 61.1 14.5 6.1 SMBO 6.7 13.3 43.3 26.7 10.0 GESE 0 0 0 0 100 CLLI 0 0 0 0 100 ARTO 0 0 0 0 100 5 BICA 0 19.8 3.1 25.0 52.1 VIRO 1.6 6.0 20.0 24.4 48.0 a Groupings of species correspond to cluster membership. Results of cluster analyses suggest that cluster membership for tree data provide better indicators of hydrologic zones than saplings or vines. Shrubs and herbaceous species would be the least useful of all strata types. Discriminant Function Analysis Discriminant function analysis (DFA) is used to determine func- tions which allow one to classify an individual into one of the 48 DISTANCES 0.000 SMRO LIST QUPH MIRE VIPA LEVI BROV AMAR QULA JUOV TRDI COCA CARA ILDE BESC ULAM ACRU QULY CAAQ TORA TADI PLAQ BOCY LYJA CELA EURA VIRO QUNI BICA CARE 3 0.200 0.004 0.031 0.035 0.046 0.033 0.017 0.016 0.013 0.002 0.019 0.042 0.010 0.019 0.011 0.032 0.064 0.077 0.009 0.106 0.106 0.086 0.043 0.152 0.118 0.170 0.018 0.003 0.014 0.095 Figure 10. Cluster diagram for herbs. Distance metric is 1-Pearson correlation coefficient, single linkage method (nearest neighbor) 49 hydrologic zones using a number of independent variables. Tbe indepen- dent variables in this analysis are the average importance values for each tree cluster and vine cluster in each zone. In cases where only one species in a cluster was present in a zone, that importance value alone was used. Table 10 lists the data set upon which this analysis was based. Three analyses were performed using DFA. The first analysis developed a classification model using only the importance value from tree clusters. This model correctly classified 47 of 55 sites for an overall misclassif ication rate of 14.6 percent (Table 11). In all cases, errors involved an assignment to a neighboring zone. For instance, a single Zone 3 site was assigned to Zone 2, and three Zone 6 sites were assigned to Zone 5. The greatest error in classification occurred for Zone 4 sites, which is expected because of the greater tolerance of species for this zone. As previously shown (Figure 4), Zone 4 is that portion of the hydrologic gradient that has species com- mon to all three distribution patterns identified. Three Zone 6 sites were assigned to Zone 5 primarily due to the lack of Zone 6 species being identified as commonly occurring. A better agreement would be expected if all species were used. The second model examined the ability of vine cluster importance values to predict zone. Although vines did not cluster as well as trees, a DFA was examined because their cluster patterns were similar. However, this model correctly classified only 26 of 55 zones for a 52 percent misclassif ication. This model was not considered adequate, and the results of its classification are not presented. 50 Table 10 Mean Importance ' Values for Species ; in Each i Cluster Used in the DFA. Arranged bv Zone/Sample Zone TCla TC2 TC3 TC4 TC5 VC2 VC3 VC4 VC5 2 78.77 0 0 0 0 0 0 0 0 2 70.26 0 0 0 0 0 0 0 0 2 65.12 0 0 0 0 0 0 0 0 2 91.7 8.26 8.32 0 0 64.06 0 203.9 0 2 62.38 0 0 0 0 0 0 0 0 2 83.49 21.16 9.48 5.66 0 0 103.24 0 162.5 2 73.45 6.19 0 0 0 141.17 38.76 53.5 0 2 68.41 12.58 0 0 0 112.22 32.13 0 11.33 2 58.62 32.76 0 0 0 0 41.32 0 0 2 58.12 4.19 0 0 0 0 300 0 0 2 74.99 33.22 3.34 0 0 300 0 0 0 3 29.57 39.47 0 8.4 0 101.56 63.02 33.87 0 3 30.3 47.5 0 19.12 0 68.2 47.04 0 0 3 0 50.08 26.13 0 0 12.98 65.86 8.65 0 3 0 71.81 0 0 0 100 0 0 0 3 24.37 66.92 7.92 0 0 93.14 0 4.89 0 3 22.84 84.85 22.59 0 0 51.13 16.28 5.57 0 3 99.86 25.52 6.41 0 0 82.8 45.12 0 0 3 30 47.29 11.72 0 0 53.87 32.81 18.92 0 3 0 85.22 13.67 11.78 0 19.27 28.8 22.65 58.54 3 32.72 54.74 10.42 0 0 89.33 78.78 0 53.11 3 7.39 48.01 16.18 37.94 0 4.2 34.88 0 12.78 3 39.94 56.21 14.8 15.66 0 0 233.26 0 0 3 10.06 99.1 47.87 13.23 0 0 144.56 47.53 5.97 3 44.79 43.95 17.52 0 0 130.77 0 0 0 3 11.06 48.2 36.72 0 0 0 0 0 0 4 0 15.04 65.41 31.36 0 135.02 0 28.46 23.4 4 0 6 52.66 40.82 0 76.47 5.96 7.9 59.12 4 0 72.7 41.12 20.82 0 13.16 24.22 8.55 8.83 4 5.94 40.19 66.29 48.27 0 30.2 68.16 25.77 0 4 16.64 31.88 155.17 34.4 0 73.4 79.8 0 0 4 0 8.35 52.08 72.63 0 5.46 13.51 42.44 0 4 0 0 40.72 126.62 0 30.46 28.14 28.88 0 4 0 17.68 21.24 81.27 0 5.74 25.98 5.97 76.54 4 10.3 12.76 18.17 187.41 0 40.49 13.95 9.71 81.55 4 0 55.18 33.31 35.37 0 5.21 44.58 60.2 25.9 4 0 7.69 19.56 67.09 0 0 0 16.41 84.39 4 0 68.09 27.28 27.85 14.53 9.97 58.12 6.32 0 a TCI indicates tree cluster 1; VC2 indicates vine cluster 2, 51 Table 10- -continued Zone TCI TC2 TC3 TC4 TC5 VC2 VC3 VC4 VC5 4 4 0 0 0 28.79 42.94 56.29 31 0 18.46 0 0 44.81 125.37 22.4 0 88.65 150 0 5 5 5 5 0 0 0 0 0 0 5.81 6.02 22.97 26.84 13.86 39.94 76.73 64.91 64.23 28.87 0 0 0 0 4.34 15.28 6.84 0 17.4 0 0 39.86 16.49 10.36 14.87 9.59 9.3 51.22 73.04 155.82 35.05 29.44 26.42 0 0 0 150 10.82 21.19 53.32 0 0 27.3 136.35 35.32 33.28 0 0 24.69 26.87 52.85 6 0 5.66 5.8 71.55 0 0 3.75 14.2 54.37 6 0 4.61 26.58 50.91 0 5.3 0 17.08 49.18 6 0 0 11.79 15.8 34.58 0 11.21 14.04 55.3 6 0 0 6.06 12.3 36.91 0 52.25 10.44 49.67 6 0 16.54 0 20.22 69.98 0 0 0 96.96 6 0 0 0 29.17 154.4 4.35 6.12 4.35 114.03 59.44 89.43 0 0 0 0 71.67 120.16 0 0 90.37 5.84 Table 11 Predicted Hydrologic Zones (Columns) and Actual Zones (Rows) Based on DFA Results Using Only Tree Importance Values Actual Hydrologic Predicted Hvdrc ilogic Zones3 Zones 2 3 4 5 6 Total 2 11 0 0 0 0 11 3 1 14 0 0 0 15 4 0 2 10 2 0 14 5 0 0 0 4 0 4 6 _0 12 0 16 0 10 _3 10 _8 8 11 55 Misclassif ication rate = 14.6 percent, 52 A final model was generated to examine the effectiveness of using the average site FTI numbers of all observed tree species at a site for each of the 55 sites. This model correctly classified 45 of 55 sites for an overall misclassif ication rate of 18.2 percent (Table 12). All misclassif ications were in an adjacent zone. Zones 2, 3, and 5 were mis- classified twice; Zone 4 was misclassif ied 3 times; and Zone 6 was mis- classified only once. Using FTI numbers for all observed tree species improved the results obtained in the first DFA using the importance values of only the commonly occurring species. In examining the data (Table 13) , it appears that three of the mis- classifications occur at site 10 (Pearl River, observations 27 to 30). The hydrology at this site was obtained from two gauges, and it is Table 12 Predicted Hydrologic Zones (Columns) and Actual Zones (Rows) Based on DFA Results Using Average FTI Values for All Observed Tree Species at the Site Actual Hydrologic Predicted Hydro logic Zone sa Zones _2 9 2 _4 0 _5 0 _6 0 Total 2 11 3 0 13 2 0 0 15 4 0 2 11 1 0 14 5 0 0 2 2 0 4 6 _0 _0_ _0 _1 10 11 9 17 15 4 10 55 Misclassif ication rate =18.2 percent. 53 Table 13 Cross -Validation Results of Zone Membership Using Linear Discriminant Function Analysis Posterior Probability of From To Membership in Zone : 2 3 Observation Zone 2 Zone 2 5 6 4 1 0.9876 0.0124 0.0000 0.0000 0.0000 2 3 3 0.1759 0.0000 0.8212 0.0000 0.0029 3 4 4 0.0000 0.1019 0.0062 0.0001 0.8919 4 6 6 0.0000 0.0075 0.0000 0.9924 0.0000 5 2 2 0.9876 0.0000 0.0124 0.0000 0.0000 6 3 3 0.2436 0.0000 0.7548 0.0000 0.0016 7 4 4 0.0000 0.0361 0.0376 0.0000 0.9263 8 6 6 0.0000 0.0007 0.0000 0.9993 0.0000 9 2 2 0.9801 0.0000 0.0199 0.0000 0.0000 10 3 3 0.0137 0.0001 0.8865 0.0000 0.0997 11 4 4 0.0000 0.0526 0.0202 0.0000 0.9273 12 5 5 0.0000 0.6341 0.0000 0.0993 0.2666 13 6 6 0.0000 0.0958 0.0000 0.9016 0.0026 14 3 3 0.1244 0.0000 0.8703 0.0000 0.0052 15 4 3a 0.0012 0.0025 0.5429 0.0000 0.4534 16 3 3 0.0597 0.0000 0.9245 0.0000 0.0158 17 4 4 0.0003 0.0075 0.2747 0.0000 0.7175 18 3 3 0.0806 0.0000 0.9092 0.0000 0.0103 19 4 4 0.0000 0.1909 0.0017 0.0005 0.8069 20 2 3a 0.1909 0.0000 0.8079 0.0000 0.0013 21 3 3 0.0748 0.0000 0.9138 0.0000 0.0114 54 Table 13- -continued Posterior Probability of From To Membership in Zone 2 3 Observation Zone 4 Zone 4 5 6 4 22 0.0000 0.0163 0.9242 0.0594 0.0000 23 3 3 0.2436 0.0000 0.7548 0.0000 0.0016 24 4 4 0.0008 0.0036 0.4558 0.0000 0.5398 25 3 3 0.0150 0.0001 0.8945 0.0000 0.0904 26 6 6 0.0000 0.0002 0.0000 0.9998 0.0000 27 2 3a 0.0205 0.0000 0.9672 0.0000 0.0124 28 3 4a 0.0011 0.0027 0.4321 0.0000 0.5641 29 4 4 0.0000 0.1286 0.0039 0.0002 0.8673 30 6 5a 0.0000 0.6307 0.0000 0.0044 0.3649 31 2 2 0.8970 0.0000 0.1030 0.0000 0.0000 32 3 3 0.3105 0.0000 0.6886 0.0000 0.0009 33 4 4 0.0000 0.0526 0.0202 0.0000 0.9273 34 5 4a 0.0000 0.3664 0.0003 0.0060 0.6274 35 4 3a 0.0014 0.0021 0.5719 0.0000 0.4246 36 5 4a 0.0000 0.4502 0.0001 0.0119 0.5378 37 6 6 0.0000 0.0785 0.0000 0.9196 0.0019 38 2 2 0.9503 0.0000 0.0497 0.0000 0.0000 39 3 3 0.2436 0.0000 0.7548 0.0000 0.0016 40 4 4 0.0000 0.4057 0.0002 0.0047 0.5894 41 6 6 0.0000 0.0013 0.0000 0.9987 0.0000 42 2 2 0.9332 0.0000 0.0668 0.0000 0.0000 43 5 5 0.0000 0.6216 0.0000 0.0699 0.3085 44 6 6 0.0000 0.0013 0.0000 0.9987 0.0000 55 Table 13- -continued Posterior Probability of From To Membership in Zone 2 3 Observation Zone 2 Zone 2 5 6 4 45 0.9503 0.0497 0.0000 0.0000 0.0000 46 3 4a 0.0001 0.0093 0.1163 0.0000 0.8743 47 6 6 0.0000 0.0015 0.0000 0.9985 0.0000 48 2 2 0.8729 0.0000 0.1271 0.0000 0.0000 49 3 3 0.2144 0.0000 0.7835 0.0000 0.0020 50 4 4 0.0000 0.1806 0.0020 0.0005 0.8170 51 6 6 0.0000 0.0011 0.0000 0.9989 0.0000 52 2 2 0.7969 0.0000 0.2031 0.0000 0.0000 53 3 3 0.1004 0.0000 0.8923 0.0000 0.0074 54 4 5a 0.0000 0.7509 0.0000 0.0739 0.1753 55 6 6 0.0000 0.0020 0.0000 0.9980 0.0000 Miclassif ied observation. 56 possible that an error was made in combining the hydrologic data from the two gauges. It seems ironic that the only zone correctly predicted at this site was Zone 4. Table 13 also shows the percentage probability of each of the 55 sites occurring in a zone. Figure 11 shows the mean site FTI plotted against the observed and predicted hydrologic zones. As expected, mean FTI numbers were greater than the observed hydrologic zones at the low end (Zone 2) and less than the zones at the upper end (Zone 6) of the hydrologic gradiant due to the lack of outlying zones (e.g., Zones 1 and 7) to pull these averages toward either extreme. Therefore, using DFA classification decision points shows that average site FTI numbers as high as 3.45 would still be in Zone 2, and average site FTI numbers as low as 5.33 would still be in Zone 6. The lower end of the predicted Zone 4 (4.16) compares favor- ably with the observed (4.0). Zone 5 predicted and observed zone values are the same (5.0). Regional Variation in Species FTI Numbers Because the 17 sites in this study occur over a broad geographic area, the possibility of regional differences in species FTI numbers was a concern. To test for possible differences, the sites were grouped into three regions: Gulf Coast (sites 1, 2, 10, 11, 12), Mississippi Valley (sites 3 through 9), and Atlantic Coast (sites 13 through 17). A two-factor analysis of variance (ANOVA) was used to test for differences in importance values between regions and clusters for trees. There was no significant interaction between region and cluster trees (F = 0.71, p = 0.68); therefore, importance values of species within a cluster do not differ among regions. There was a significant difference ZONE 6 I I 4- ZONE 5 I I I _L J_ 3 4 5 OBSERVED HYDROLOGIC ZONE ZONE 4 ZONE 3 ZONE 2 57 o N 5.33 y 5. 00 g _i o Q >- x 4. 16 S u D UJ 45 a Figure 11. Mean tree FTI numbers plotted versus observed and predicted hydrologlc zones for all 55 sites 58 between regions (F = 4.02, p = 0.019), reflecting the fact that impor- tance values of trees were greater in the Mississippi Valley, averaging 49.1 + 47.6, than the Gulf Coast (37.5 + 37.2) and Atlantic Coast (39.3 + 34.4) regions. A number of factors may contribute to this phe- nomenon, including stand maturity and localized disturbances. Another two-factor ANOVA also was used to determine whether the predicted zone values generated were more accurate in one region or another. The absolute value of the difference between the predicted and actual zones was used as the dependent variable. There was neither an interaction (F = 1.44, p - 0.18), nor a main effect (F = 1.54, p = 0.22) involving region, which indicates that the hydrologic zones can be pre- dicted with equal accuracy among the specified regions. SUMMARY AND CONCLUSIONS Bottomland hardwood forests are dynamic and complex systems. Frequent flooding from adjacent streams provides the forcing function that characterizes the affected plant communities. Frequency and dura- tion of floodwater determine the extent of anaerobic soil conditions that directly affect plant populations. Plant species adapted for life in anerobic soil conditions are located in the topographically lowest areas subject to long duration flooding. Species composition changes as the elevational and associated moisture gradient changes from wettest to driest, and reflects species adaptations to the prevailing hydrologic regimes . Determination of a hydrologic gradient often requires extensive data gathering over a long period. However, many studies have shown that a definite relationship exists between plant species and the tim- ing, frequency, and duration of inundation and soil saturation (Larson et al . 1981). This study was undertaken to express quantitatively the optimum position of various plant species along a hydrologic gradient. Previous studies have estimated the location of plant species and communities along a hydrologic gradient. Various systems have been proposed that use vegetation to predict the duration and/or frequency of flooding. However, previous studies were limited to a small geographic area, the developed systems are qualitative, and vegetation data used to 59 60 predict the degree of flooding for the entire southeastern United States previously have been literature -based involving many studies with vary- ing research designs. Vegetation data resulting from this study related four vegetation strata and three life forms occurring in 55 stands at 17 sites through- out the southeastern United States. Hydrologic regimes were calculated for a 10- to 20-year period of record for each stand. A flood tolerance index (FTI) system of weighted averages based on importance values was developed, and FTI numbers were calculated for various life stages of each species identified in the study. Three hundred and twelve species were identified for each of 4 strata in the study including 74 tree species, 188 species of saplings and shrubs, 31 species of woody vines, and 268 species of herbs and woody seedlings. Comparison of the FTI numbers with two other systems (Hook 1984; Reed 1988) using vegetation to estimate wetness showed gen- eral agreement among the systems, especially for mature trees. Cluster analysis and discriminant function analysis were used to evaluate the weighted averaging technique and explore the best method for using the FTI numbers in predicting hydrologic regimes. Tree, sapling, and vine data clustered into distinct groups. Her- baceous and shrub data did not group distinctively. Tree and vine importance values for each cluster in a zone/sample (data taken in a single zone at a site) and FTI numbers for tree data were used as inde- pendent variables for the discriminant function analyses. Tree species were found to be more useful than saplings, shrubs, vines, or herbaceous species in predicting hydrologic zones. The tree data alone using importance values provided 85 percent accuracy. Tree data alone using 61 FTI numbers was only slightly less accurate at 82 percent. All misclas- sifications assigned membership to a neighboring zone. Misclassif ica- tions are understandable for two important reasons. Zone 4 contains the more facultative species because as wetness decreases, other environmen- tal conditions begin to exert greater influence. Also, since Zone 5 is so narrow compared to other zones and most species occurring in Zone 5 occur in greater abundance in either Zone 4 or 6 , very little difference in the community structure exists between the top of Zone 4 and the bottom of Zone 6 . The accuracy of these predictions may be somewhat inflated, because hydrologic zone was a parameter used to derive species FTI numbers . There were no regional (Gulf Coast, Lower Mississippi Valley, and Atlantic Coast) differences in the accuracy of the weighted averaging and predicted values. Therefore, a single FTI number calculated for each species can be used to predict zones for the entire study area. The implication of this study is that the calculated FTI numbers can be used to estimate hydrologic regimes in bottomland forest systems of the southeastern United States. Trees were determined to be the most reliable vegetative growth form for determining hydrologic zones. How- ever, this study was conducted in relatively undisturbed areas. Because trees can remain for decades following hydrologic disturbance, a modifi- cation of the method using saplings and seedlings may prove to be more reliable . Techniques used in this study to develop FTI numbers in the south- eastern United States may be applicable to regions of the country that have similar types of riverine forest systems. APPENDIX A SITE DESCRIPTIONS AND MAP LOCATIONS Neches River (Sites 1 and 2) Location (Neches River Basin) - These sites are located in the National Big Thicket Preserve, Jack Gore Baygall unit, 6.4 km north of Evadale in Jasper County, Texas. Reference U.S. Geological Survey (USGS) map, Silsbee, Texas, N3015-W9400/15 , 1955. Hydrology data - Twenty years of hydrology data were obtained for a staff gauge on U.S. Highway 96 bridge at Evadale. Slope correction from gauge datum to study site was determined by using a water surface profile . General vegetation - Plant communities range from Taxodium distichum- Nyssa aquatica in deep sloughs to Quercus alba-Pinus taeda and Fagus grandifolia on the nearby ridges. Intermediate communities consist of Quercus lyrata , Carya aquatica , Quercus michauxii , Liquidambar styraci- flua, Ulmus americana , and Carpinus carol iniana. Soils and climate - Soils vary from the very poorly drained Angelina series in sloughs to the moderately well drained Spruger series on ridges. Other soil series encountered were Bleakwood, Urbo , and Attoyac . Average annual rainfall in the area is 170 cm, and the average growing season is 234 days. Delineated zones - Zones 2, 3, 4, and 6 were delineated for both sites. Zone 5 was too narrow to reliably delineate due to its position on the slope of the floodplain terrace. 63 64 •t) C w 0) o 0) z u 3, 65 Steele Bayou (Site 3) Location (Yazoo River Basin) - This site is located in the Yazoo National Wildlife Refuge, 6.4 km northeast of Glen Allen in Washington County, Mississippi. Reference USGS map, Percy, Mississippi, N3300- W9052.5/7.5, 1967. Hydrology data - Twenty years of hydrology data were obtained for a gauge on the bridge over Steele Bayou 6.4 km south of Grace, Missis- sippi. Slope correction from gauge datum to study site was determined by a water surface profile. General vegetation - Plant communities ranged from a Salix nigra- Taxodium distichum community at lower elevations to a Sassafrass albidum-Liquidambar styraciflua-Quercus nigra community at the highest elevation. Intermediate communities are dominated by Planera aquatica , Forestiera acuminata, Quercus lyrata , Carya aquatica , Fraxinus pennsyl- vanica, Celt is laevigata , Acer negundo , and Ulmus americana. Soils and climate - Soil series range from Sharkey at the lowest eleva- tions to Dundee at the higher elevations. Average annual rainfall in the area is 132 cm and the average growing season is 213 days. Delineated zones - Zones 2 through 6 were delineated for study at this site . 66 0) 4J 3 o a) OJ a) 67 Ouachita River (Sites 4 and 5) Location (Ouachita River Basin) - These sites are located in the Felsen- thal National Wildlife Refuge, 8 km west of Crossett and .8 km east of Felsenthal, respectively, in Union County, Arkansas. Reference USGS map, Felsenthal, Arkansas-Louisiana, N3300-W9200/15 , 1981. Hydrology data - Nineteen years of hydrology data were obtained for a gauge on a U.S. Highway 81 bridge, 8 km west of Crossett, Arkansas. A slope correction for site 5 was determined using a water surface pro- file. Site 4 was adjacent to the gauge, so no correction was necessary. General vegetation - Plant communities range from a Taxodium distichum- Cephalanthus occidental Is dominated community in lower areas to a nearly monotypic stand of Pinus taeda in higher areas. Intermediate communi- ties are dominated Carya aquatica , Quercus lyrata , Diospyros virginiana , Quercus phellos , Quercus nuttallii , and Liquidambar sCyraciflua. Soils and climate - All encountered soils are in the Una series. Aver- age annual rainfall is 140 cm and the average growing season is 211 days. Delineated zones - Only Zones 3 and 4 were delineated for both sites 4 and 5. Zone 2 was not used because the hydrology was not reflected by the gauge data. Zones 5 and 6 were not used due to major disturbances from recent silvicultural and agricultural practices. 68 Figure A-3. Ouachita River (sites 4 and 5) 69 Yazoo River (Site 6) Location (Yazoo River Basin) - This site is located on the north side of the Yazoo River, 8 km west of the U.S. Highway 61 bridge and 12.0 miles north of Vicksburg, Mississippi in Issaquena County. Reference USGS map, Long Lake, Mississippi-Louisiana, N3222 . 5-W9052 . 5/7 . 5 , 1962. Hydrology data - Twenty years of hydrologic zone elevations were com- puted by analyzing flow data from gauges on the Mississippi River at Vicksburg, Mississippi, on the Yazoo River 2.4 km east of the site, and at the Steele Bayou control structure immediately adjacent to the study site. General vegetation - Plant communities range from Quercus lyrata-Carya aquatica at lowest elevations to a Liquidambar styraciflua-Ulmus ameri- cana-Celtis laevigata association at the highest elevations. Other commonly occurring species include Ilex decidua, Carya illinoensis , and CercLs canadensis . Soils and climate - Soils were determined to be in the Sharkey series. Average annual rainfall is 127 cm. The average growing season is 221 days. Delineated zones for study - Only Zones 3 and 4 were delineated for this study. Zones 2, 5, and 6 were either too narrow or too disturbed to provide reliable data. 70 o o N 0J I < u 71 Big Black River (Site 7) Location (Big Black River Basin) - The site is located on the south bank of the Big Black River adjacent to the Fisher Ferry bridge on Fisher Ferry Road, 24 km south- southeast of Vicksburg, Mississippi. The site is in Claiborne County, Mississippi. References USGS map, name N3207.5- W9045/7.5, 1963. Hydrology data - Twenty years of hydrologic data were analyzed for a flow gauge on the U.S. Highway 80 bridge, 3.7 km east of Bovina, Missis- sippi. A slope correction from gauge location to site was determined by a water surface profile. General vegetation - Plant communities range from Taxodium disCichum- Nyssa aquatica at lower elevations to Liquidambar styraciflua-Celtis laevigata-Ulmus americana at higher elevations. Intermediate communi- ties are dominated by Planera aquatica, Carya aquatica, Quercus lyrata, and Fraxinus pennsylvanica. Soils and climate - Soils range from the Waverly series ( depress ional phase) in lowest elevations to the Faylaya series at higher elevations. Average annual rainfall for this area is 132 cm and the average growing season is 226 days. Delineated zones - Zones 2, 3, and 4 were delineated for study. Zones 5 and 6 were not delineated due to major vegetation disturbance induced by silvicultural and agricultural practices. 72 SITE 7 PLOT 2 ZONE 2 , >s TBM2 II CYPRESS // BOTTOM // PLOT 3 ZONE 3 Figure A-5. Big Black River (site 7) 73 L'Anguille River (Sites 8 and 9) Location (L'Anguille River Basin) - Sites 8 and 9 are located on the west bank of L'Anguille River, 0.8 km east and 7.2 km southeast, respec- tively, of Palestine in St. Francis County, Arkansas. Reference USGS map, Marianna, Arkansas, N3445-W9045/15 , 1957. Hydrology data - Twenty years of hydrologic data were analyzed for a gauge located on the U.S. Highway 70 bridge, 0 . 8 km east of Palestine, Arkansas. A slope correction was computed for site 9 using a water surface profile. No slope correction was necessary for site 8 because it was adjacent to the gauge. General vegetation - Plant communities range from Taxodium distichum- Nyssa aquatica dominated communities at the lowest elevations to a Carya tomentosa-Quercus alba-Liquidambar styraciflua dominated association on adjacent ridges. Intermediate communities are dominated by Quercus lyrata , Carya aquatica , Diospyros virginiana , Fraxinus pennsylvanica , and Ulmus americana. Soils and climate - Soil series range from Zachary at lower elevations to Loring on adjacent ridges. Average annual rainfall is 132 cm, and the average growing season is 219 days. Delineated zones - Zones 3 and 4 were delineated for sites 8. Zones 3 and 6 were delineated for site 9. All other zones were unacceptable. 74 _OLD_ _ _ROAD /PLOT3©n ~ TBM a;. M-1 3-4-81 FIRST NATIONAL BANK OF EASTERN ARKANSAS PLOT 4 | ZONE 4 \ . :SITE8 (\ ' TBM~T\ BED ZONE 3 WOODEN BRIDGE x<^. ELCANNON AME CHURCH FIELDS HOUSES It TBM 1 A s"" JJl PLOT 3 plot 6 v;.; y1-' ffi«>SITE II t ELCANNON M CEMETERY,1;; * POWER ! POLES = 4 I i SOYBEAN FIELD It |+^ TBM M-1 3-1 5-81 CO LU 8 NOT TO SCALE Figure A-6. L'Anguille River (sites 8 and 9) 75 Pearl River (Site 10) Location (Pearl River Basin) - This site is located in the Pearl River State Wildlife Management Area, 8 km north of Slidell in St. Tammany Parish, Louisiana. Reference USGS map Nicholson, Mississippi-Louisiana, N3022.5-W8937.5/7.5, 1955. Hydrology data - Twenty years of hydrology data were analyzed for two gauges. First 10 years data was extrapolated to present gauge on the Southern Railway bridge at Pearl River, Louisiana. No slope correction was necessary because the site is adjacent to the gauge. General vegetation - Plant communities range from Taxodium distichum- Nyssa aquatica dominated communities in sloughs to Liquidambar styraciflua-Quercus nigra dominated communities on low ridges. Interme- diate communities are dominated by Quercus laurifolia , Acer drummondii , Fraxinus pennsylvanica , Carpinus caroliniana , and Ilex opaca. Soils and climate - Soil series range from Rosebloom (depressional phase) in sloughs to Prentiss on ridges with Arkabutla at intermediate elevations. Average annual rainfall is 152 cm, and the average growing season is 237 days. Delineated zones - Zones 2, 3, 4, and 6 were delineated. Zone 5 could not be reliably delineated due to topography. 76 / SHOOTING I RANGE / ~~\\ ZONE 6 y^. IT MISSISSIPPI BRIDGE DESTROYED^-kBM DESTROYED SWAMP CLEARED /' PLOT 6 ^ O* / ^ NEW CONC BRIDGES- HONEY ISLAND BA YOUS z -ENGLISH-— 8**°° ^% LOUISIANA C,^ >lV± p.° ^/grTvelRO Ttbm ZONE 4 _< t£IRT RD CAUSEWAY » ^ SECTIONS NOT TO SCALE Figure A-7. Pearl River (site 10) 77 Apalachicola River (Sites 11 and 12) Location - Site 11 is located on the west bank of the Apalachicola River, immediately south of the Florida Highway 20 bridge, 1 . 6 km east of Blountstown in Calhoun County, Florida. Site 12 is located on the east bank of the river, 4.8 km north of Bristol in Liberty County, Florida. Reference USGS maps Blountstown, Florida, N3022 . 5-W8500/7 . 5 , 1945, and Bristol, Florida, N3022 . 5-W8452 . 5/7 . 5 , 1945, respectively. Hydrology data - Twenty years of hydrology data for Site 11 were ana- lyzed for a gauge located 0 . 8 km south of Highway 20 bridge at the Neal Lumber Company Landing. Hydrology data for Site 12 were analyzed from data from a previous study (Leitman et al . 1984). General vegetation - Vegetation for both sites ranges from Nyssa aquatica- -dominated swamps at lower elevations to Nyssa sylvatica- Juglans nigra-Sassafras albidum dominated associations at higher eleva- tions. Intermediate plant communities consist of Fraxinus caroliniana, Gleditsia aquatica , Quercus lyrata , Carya aquatica , Ulmus americana , Helia azederach, Celtis laevigata , and Quercus nigra. Soils and climate - Soil series range from Bibb at lower elevation to Ochlochonee at higher elevations. Soils series occurring at intermedi- ate elevations were Chastain, Enoree, Jena, and Chewacla. Average annual rainfall is 137 cm, and the average growing season is 267 days. Delineated zones - Zones delineated for site 11 were 2, 3, 4, and 5. Essentially all of Zone 6 has been developed for agriculture. Zones delineated for site 12 were 4, 5, and 6. The hydrology of Zones 2 and 3 had been altered by an extensive network of beaver dams and was not reliable. 78 APALACHICOLA FLOW RIVER BERM / PLOT 5 ZONE 5 \ \ >e3 \ PLOT4 m ZONE4 STAFF GAUGE ^ SLOUGH SITE 1 1 / / NOT TO SCALE ca r-l o o 1-1 si u cd t-t ttf a, < 0) to 80 OcmulEee River (Site 13) Location (Ocmulgee River Basin) - This site is located across the river from Lumber City, and adjacent to the east side of Southern Railway and U.S. Highway 23 and 341 in Jeff Davis County, Georgia. Reference USGS map, Lumber City, Georgia, N3152 . 5-W8237 . 5/7 . 5 , 1971. Hydrology data - Twenty years of hydrology data were analyzed for a gauge on the U.S. Highway 23 and 341 bridge adjacent to the site. No slope correction was necessary. General vegetation - The plant communities range from Taxodium distichum-Nyssa aquatica communities at the lowest elevations to a Carya tomentosa-Quercus alba-Pinus glabra dominated association at the higher elevations. Intermediate communities consist of Planera aquatica , Quer- cus lyrata, Carya aquatica , Ulmus americana , Liquidambar styraciflua , Quercus phellos , Carpinus caroliniana , and Quercus nigra. Soils and climate - The soil series range from Bibb in the lowest areas to Riverview at the highest elevations. The Chastain series occurs at intermediate elevations. Average annual rainfall in this area is 117 cm, and the average growing season is 232 days. Designated zones - Zones 2, 3, 4, and 6 were delineated for study. Zone 5 was too narrow to provide reliable data. 81 Figure A- 10. Ocmulgee River (site 13) 82 Altamaha River (Site 14) Location (Altamaha River basin) - This site is located in the northeast quadrant at the intersection of U.S. Highway 1 and the Altamaha River, 50 km north of Baxley in Toombs County, Georgia. Reference USGS map, Baxley NE, Georgia, N3152 . 5-W8215/7 . 5 , 1970. Hydrology data - Twelve years of hydrology data were analyzed for a gauge on the U.S. Highway 1 bridge adjacent to the site. No slope cor- rection was necessary. General vegetation - Plant communities range from Taxodium distichum- Nyssa aquatica at lowest elevations to a Juniperus virginiana-Quercus stellata-Carya tomentosa-Pinus taeda community at highest elevations. Intermediate communities are dominated by Fraxinus pennsylvanica , Quer- cus michauxii , Quercus phellos , and Carpinus caroliniana. Soils and climate - Soils range from the Osier series in lowest eleva- tions to the Riverview series at highest elevations. The Chewacla series occurs at intermediate elevations. Average annual rainfall is 117 cm, and the average growing season is 232 days. Designated zones - Zones delineated for this study site were 2, 5, and 6. Zones 3 and 4 had ridge and swale topography which prevented sepa- rating them reliably. 83 USGS A BM N 201-Z r 1 1 BOA 7" RAMP' > 5 i 3 W )\ / / / SLOUGH--. T £1 21 SITE 14 s J /'* / V / / / / / RIDGE & SWALE ZONE 5 PLOT 5 SLOUGH li L. _ .. , ZONE 6 EBB PLOT 6 -BB&PZL°0^ ■SLOUGH— « RIVER GAUGE TBM A BERM FLOW ALT AM AH A RIVER NUCLEAR PLANT NOT TO SCALE Figure A-ll. Altamaha River (site 14) 84 Edisto River (Site 15) Location (Edisto River Basin) - This site is in Givhans Ferry State Park, north and west of South Carolina Highway 61 bridge, 4.8 km west of Givhans in Colleton County, South Carolina. Reference USGS map, Maple Cane Swamp, South Carolina, N3300-W8022 . 5/7 . 5 , 1979. Hydrology data - Twenty years of hydrology data were analyzed for a gauge on the South Carolina Highway 61 bridge adjacent to the site. No slope correction was necessary at this site. General vegetation - Plant communities range from Taxodium distichum- Nyssa aquatica-Fraxinus carol in i ana at lowest elevations to a Pinus taeda-Quercus virginiana-Quercus nigra dominated community at highest elevation. Species in communities at intermediate elevations include Quercus lyrata , Quercus laurifolia , Planera aquatica , Liquidambar styra- ciflua, Quercus nigra, and Carpinus caroliniana. Soils and climate - Soil series range from Osier at lowest elevations to Chipley at highest elevations. The Torhunta soil series occurs at intermediate elevations. Average annual rainfall is 132 cm, and the average growing season is 213 days. Delineated zones - Zones 2, 3, 4, and 6 were delineated, but Zone 5 was too narrow due to its topographic position. 85 c (U > i-l OS o W t-l w (V u 3, 1-1 86 Lynches River (Site 16) Location (Lynches River Basin) - This site is located in Lynches River State Park, 1.6 km south of Effingham in Florence County, South Caro- lina. Reference USGS map, Florence West, South Carolina, N3400- W7945/15, 1940. Hydrology data - Twenty years of hydrology data were analyzed for a gauge on the bridge on U.S. Highway 52, 1 . 6 km south of Effingham. A slope correction from gauge to site was determined by a water surface profile . General vegetation - Plant communities range from Taxodiwn distichum- Nyssa aquatica at the lowest elevations to Quercus falcata-Quercus stellata-Carya tomentosa at higher elevations. Species occurring in communities at intermediate elevations include Quercus lyrata , Quercus laurifolia , Liquidambar styraciflua , and Quercus phellos . Soils and climate - Soil series range from Chastain at lowest elevations to Chipley at highest elevations, with the Wehadkee series at intermedi- ate elevations. Average annual rainfall is 107 cm, and the average growing season is 237 days. Delineated zones - Zones 2, 3, and 6 were delineated for study. Zones 4 and 5 were too narrow to delineate due to their topographic positions. 87 SITE 16 NOT TO SCALE Figure A-13. Lynches River (site 16) Waccamaw River (Site 17) Location (Waccamaw River Basin) - This site is located 3 . 2 km southeast of Longs in Horry County, South Carolina. Reference USGS map, Longs, South Carolina-North Carolina, N3352 . 5-W7837 . 5/7 . 5 , 1947. Hydrology data - Twenty years of hydrology data were analyzed for a gauge on the bridge on State Highway 9, 3.4 km southeast of Longs, South Carolina. No slope correction was necessary. General vegetation - Plant communities range from a Fraxinus carol iniana dominated swamp at the lowest elevation to a Pinus taeda-Quercus nigra- Liquidambar styraciflua-dominated community at the highest elevation. Vegetation dominated by Quercus laurifolia , Magnolia virginiana , Acer rubrum, loblolly bay, Ilex opaca , Quercus phellos , and Quercus michauxii occurs at intermediate elevations. Soils and climate - Soils range from the very poorly drained Rutledge series at the lowest elevations to the moderately well drained Chipley series at highest elevation. Soils at intermediate elevations include the Rembert and Leon series. Average annual rainfall is 107 cm, and the average growing season is 248 days . Delineated zones - Zones 2, 3, 4, and 6 were delineated for this site. Zone 5 was too narrow to delineate due to its topographic position. 89 Si F1 Si SITE 17 TIN ^ gi ROOF ^--^.o:/ fic, BUILDING/VTT ""-TF^ai^My v>tV tt ^ « £// // ZONE4-^ GA«Ji/rL"RT ZONE 6 PLOT 6 ■*. turn _j\ inn} RD ./ ZONE 4 PLOT 4 -BLUE FLAGGING MARKS ENTRANCE NOT TO SCALE ; . ; RIVe/f Figure A- 14. Waccamaw River (site 17) APPENDIX B GUIDE FOR COMPUTER PROGRAM FOR ANALYZING HYDROLOGIC DATA Introduction Program Development The FORTRAN computer program was developed to streamline hydrologic data manipulation for hydrologic zone boundary determinations . Overview of Program Capabilities The program defines the lower limits of four hydrologic zone boundaries, enabling determination of five zones. Recognizing the inherent variability of some input parameters, the program allows input of a range of parameters values, thus yielding a range of boundary values. The program reads either standard formatted U.S. Geological Survey (USGS) flow rate or Corps of Engineers stage data. Basis for Hydrologic Zone Delineation Definitions Hydrologic zones . The program assumes the following definitions for hydrologic zone boundaries: Zone Definition 2 Soil inundated or saturated on average greater than 75 percent of the growing season. 3 Soil inundated or saturated on average between 75 and 25 percent of the growing season. 4 Soil inundated or saturated on average between 25 and 12.5 percent of the growing season. 5 Soil inundated or saturated on average between 12.5 and 5 percent of the growing season. 6 Soil inundated or saturated on average less than 5 percent of the growing season. Inundation. Inundation is defined as the physical overtopping of the soil by the adjacent stream water surface. Saturation. A soil is saturated when it will no longer absorb water without losing an equal amount. Growing season. The growing season is defined as the average frostfree period. It is the period between the last average occurrence of 32° F in the spring and the first average occurrence of 32° F in the fall. 91 92 Separation of Data into Growing Seasons Required inputs. The program used the starting and ending dates of the growing season and the daily flow rates or daily stage data at each site for a period of 10 to 20 years as input. Henceforth, the terms "flow rate" and "stage" are interchangeable, depending on the form of input data. The program read flow rate data into a matrix FL(I,J), where I is an index designating year, and J designated the Julian date. For exam- ple, FL(3,3) designated the flow rate of January 3 of the third year of data and FL(3,33) designated February 2 of the third year. Two matrices, K5GS(I) and KEGS (I), were created that contained Julian dates of the beginning and end of the growing season in each year I. These dates varied due to leap years. The program then created two more matrices: FLGS(I,J) and NDGS(I). FLGS(I,J) contained flow rates in year I, on date J from the beginning of the growing season. For example, FLGS(2,3) designated the flow rate on the third day of the growing season in the second year of data. NDGS(I) is the number of days in the growing season of the Ith year. Again, this varied due to leap years. Computation of Days Inundated Computation of days inundated at a given flow rate proceeded as follows : 1. Growing season flow rates for the entire entered record were ranked from highest to lowest. This was accomplished by transferring all data in the FLGS(I.J) matrix into a single subscripted matrix RANK(K) , and then ranked the data in RANK using a bubble sorting routine. RANK(l) represented the greatest flow rate for the entire record. RANK(NQ) was the lowest flow rate, where NQ was the total number of growing season flow rates in the entire record. 2. The number of days a given flow rate was exceeded (i.e., effecting inundation above that flow rate) within the growing season record was equal to the flow rate's ranking. A flow rate in RANK(IO) was equalled or exceeded 10 times within the historic record and corresponded to 10 days of inundation. Computation of Days Saturated General . The model for the program was a simplified water bal- ance. The soil root zone for wetland plant species was considered to be a water bucket, where: 1. Depth of the bucket was the critical depth of the saturated zone (e.g., 25 cm) . 2. When water overtopped the bucket (i.e., during inundation), the bucket filled at a rate PKW (inches/hour). 93 3. When inundation ceased, the bucket drained at a rate PKD (inches/hour) . 4. DEVAPR inches of water were lost daily out the top of the bucket by evapotranspiration. 5. When the bucket was partially full (i.e., not empty or over- topped by inundation) , a day of saturation was counted. Inputs . The following inputs were required for computing days of saturation: 1. PKD = rate of percolation of water through the soil column. This was estimated from soil conservation service county soil surveys . 2. PKW = rate at which inundation restores the soil zone to full saturation (inches/hour). (Note: PKW is different from PKD when the underlying soil layer was less porous than the upper soil layer. The soil would wet at the permeability rate of the upper layer, but would drain at the permeability rate of the underlying soil layer) . 3. CDSZ = critical depth of the saturated zone, which was the depth of wetland plant root zone (10 in.). 4. DEVAPR = average daily evapotranspiration rate for the site (inches/day) . Discussion. This model has no pretense of absolute accuracy; it does not include rainfall, contributions of water from upslope drainage, residual soil moisture held by soil after gravity drainage, and possibly other sources of water. Program algorithm, for a "fixed" flow rate. The following describes the program algorithm for a "fixed" flow rate (used at each zone boundary) for computation of days saturated: 1. For each year, the program started at CDSZ/24*PKW days before the growing season, with an empty "bucket," and used the gen- eral scheme described above to find the depth of water in the root zone at the start of the growing season. A flow rate in the record greater than the "fixed" flow rate filled the "bucket," while a lesser flow rate drained the "bucket." 2. For each year, the program then checked each growing season flow rate versus the "fixed" flow rate using this starting water depth, and either filled or drained the bucket as appropriate. Each day when the bucket was partially full counts as a day saturated. 3. Days saturated for each year were added into a single total for that "fixed" flow rate. 94 4. A flowchart showing program logic for computation of days saturated is provided in Figure C-l. Computation of Frequency of Inundation General . Frequency of inundation was not used in defining hydro - logic zone boundaries. However, it was used for discussion purposes. Frequency of inundation for a given flow rate was determined by: 1. Counting the number of growing seasons in the historic record during which the flow rate for a particular hydrologic zone was exceeded at least once for a minimum of 7 days. 2. Dividing the total number of years in the record into "a," and multiplying by 100. For example: A flow rate of 20,000 cubic feet per second is exceeded in only two growing seasons of a 20-year record. This flow rate has a fre- quency of inundation of 2/20 x 100 = 10 percent. Computation of Parameters in Tabular Output Tabular output appeared as follows : (1) (2) (3) (4) (5) (6) DAYS (7) DURATION DAYS FREQ SATURATED SATURATED BOUNDARY INUN- DURATION INUN- DAYS PLUS DAYS AND FLOWRATE DATED INUNDATED DATED SATURATED INUNDATED INUNDATED 38852. 33. 4.3 100.0 38 5.0 All of the above have already been explained except columns (3) , (6) and (7). Column 3 - Duration inundated days inundated x 100. total growing season 2. Column 6 - Days saturated plus days inundated = Columns 2+5. 3 . Column 7 - Duration saturated and inundated days inundated + days saturated total growing season days x 100. 95 COMPUTE WATER TABLE AT START OF GROWING SEASON INPUT (ESTIMATED BOUNDARY FLOWRATE) SAVE #DAYS SATURATED INPUT FLOWRATE OF NEXT DAY Figure B-l. Program logic for computation of days saturated 96 EXAMPLE SUMMARY TABLE OF BOUNDARY FLOWRATES ZONE V 155. ZONE IV 144. ZONE III 128. ZONE II 63. NSMGS= 3 NSDGS= 3 NEMGS=11 NEDGS=11 PKW=. 90000 PKD=. 90000 CDSZ=10.0 DEVAP=.200 DAYS DURATION DAYS FREQ SATURATED SATURATED BOUNDARY INUN- DURATION INUN- DAYS PLUS DAYS AND FLOWRATE DATED INUNDATED DATED SATURATED INUNDATED INUNDATED 155. 22. 4.3 100.0 2. 24. 4.7 144. 48. 9.4 100.0 7. 55. 10.8 128. 116. 22.8 100.0 11. 127. 25.0 63. 375. 73.8 100.0 11. 386. 76.0 APPENDIX C HYDROGRAPH FOR STEELE BAYOU (SITE 3) 98 ID LU Z O N IT) T CO C\J UJ UJ UJ UJ z Z z Z o O O O N N N N SdO '3±VU MO"ld to s: u o 00 U <3\ c <-> 0) ■ W VO ■ •0 r-l cd r~ 43 Ov W .-1 0) X •• •u o ^s O m r~ in o\ VO 0) rH o\ 4-> i r-l •H VO 1 W VO t-l w a\ vO 1-1 Ov 3 .-1 o >-> •• cfl J3 • • cQ 03 - a) in i-l VO a) ct\ a) .h 4J i CO r-l VO 1-1 a\ O i-l 4-1 X. ■■ a, -, 99 percent) under natural conditions in wetlands. Facultative Wetland (FACW) . Usually occur in wetlands (estimated proba- bility 67 to 99 percent), but occasionally found in nonwetlands. Facultative (FAC) . Equally likely to occur in wetlands or nonwetlands (estimated probability 34 to 66 percent) . Facultative Upland (FACU) . Usually occur in nonwetlands (estimated pro- bability 67 to 99 percent), but occasionally found in wetlands (esti- mated probability 1 to 33 percent) . Obligate Upland (UPL) . Occur in wetlands in another region, but occur almost always (estimated probability >99 percent) under natural condi- tions in nonwetlands in the region specified. If a species does not occur in wetlands in any region, it is not on the National List. A positive (+) or negative (-) symbol was used with the Faculta- tive Indicator categories to define more specifically the regional 201 frequency of occurrence in wetlands. The positive sign indicates a fre- quency toward the higher end of the category (more frequently found in wetlands), and a negative sign indicates a frequency toward the lower end of the category (less frequently found in wetlands). REFERENCES Bailey, R. G. Description of the ecoregions of the United States. A/" 1980; Washington, DC: U.S. Department of Agriculture, Miscellaneous Publication No. 1391, 77 p. Bedinger, M. S. Forest species as indicators of flooding in the Lower White River Valley, Arkansas. Geological Survey Research, U.S. Geologi- cal Survey Professional Paper 750-C; 1971; C248-C253. Bedinger, M. S. Relation between forest species and flooding. Greeson, P. E. ; Clark, J. R. ; Clark, J. E., eds . Wetland functions and values: the state of our understanding; proceedings of the national symposium on wetlands. Minneapolis: American Water Resources Association; 1978; 427-435. Best, R. G.; Segal, D. S.; and Wolfe, C. Soil-vegetation correlations in selected wetlands and uplands of north-central Florida. 1990; 48 p. Biological Report 90(9). Bray, J. R. A study of natural occurrence of plant species. Ecology 37:21-28; 1956. r- Brinson, M. M. Riverine forests. Lugo, A. E.; Brinson, M. ; Brown, S., eds. Ecosystems of the world; Vol 15, Forested wetlands. New York: Elsevier; 1990; 87-141. Broadfoot, W. M. ; Williston, H. L. Flooding effects on southern for- ests. J. For. 71:584-587; 1973. Brown, R. T. ; Curtis, J. T. The upland conifer-hardwood forests of northern Wisconsin. Ecol. Monogr. 22:217-234; 1952. Brown, S. A comparison of the structure, primary productivity, and transpiration of cypress ecosystems in Florida. Ecol. Monogr. 51(4):403-427; 1981. Bruckner, W. ; Bowersox, T.W. ; Ward, W. W. Response of trees to flooding at the Curwensille Reservoir. University Park, PA: Pennsylvania State University; 1973; 4 pp. Cannon, W. A.; Free, E. E. Root adaptation to deficient soil aeration. Yearbook Vol 19. Washington, DC: Carnegie Institution of Washington; 1920; 62. 202 203 Chambless, L. F. ; Nixon, E. S. Woody vegetation-soil relations in a bottomland forest of east Texas. Texas Journal of Science XXVI (3, 4): 407-416; 1975. Clark, J. R. ; Benforado, J., eds . Wetlands of bottomland hardwood for- ests. New York: Elsevier; 1981. Clements, F. E. Plant succession. Washington, DC: Carnegie Institution of Washington; 1916. Publication 242. Conway, V. M. Aeration and plant growth in wet soils. Bot. Rev. 6:149-163; 1940. Curtis, J. T. The vegetation of Wisconsin: An ordination of plant communities. Madison, WI : University of Wisconsin Press; 1959. Curtis, J. T. ; Mcintosh, R. P. An upland forest continuum in the prairie-forest border region of Wisconsin. Ecology 32:476-496; 1951. Daubenmire, R. Plant communities; a textbook of plant synecology. New York: Harper & Row; 1968. Dickson, R. E. ; Hosner, J. F. ; Hosley, N. W. The effects of four water regimes upon the growth of four bottomland tree species. For. Sci. 11:299-305; 1965. Dubinina, J. M. Metabolism of roots under various conditions of aeration. Soviet Plant Physiology 8:314-322; 1961. Faulkner, S. P.; Patrick, W. H. , Jr.; Parker, W. B.; Gambrell, R. P.; Good, B. J. Characterization of soil processes in bottomland hardwood wetland-nonwetland transition zones in the lower Mississippi River Val- ley. Vicksburg, MS: U.S. Army Engineer Waterways Experiment Station; 1991. Contract report WRP-91-1. Gleason, H. A. The structure and development of the plant association. Torrey Bot. Club Bull. 44:463-481; 1917. Grubb, P. J. The maintenance of species -richness in plant communities: the importance of the regeneration niche. Bibl. Rev. 52:107-145; 1977. Hall, T. F. ; Smith, G. E. Effects of flooding on woody plants, West Sandy Dewatering Project, Kentucky Reservoir. J. For. 53:281-285; 1955 Hodges, J. D.; Switzer, G. L. Some aspects of the ecology of southern bottomland hardwoods. Stoneville, MS: Mississippi Agricultural and For- estry Experiment Station; 1979. Journal paper 4087; 360-365. Hook, D. D.; Brown, C. L. Root adaptations and relative flood tolerance of five hardwood species. For. Sci. 19:225-229; 1973. Hook, D. D. ; Crawford, R. M. M. , eds. Plant life in anaerobic environ- ments. Ann Arbor, Ml: Ann Arbor Science Publishers; 1980. 204 Hook, D. D. ; McKee , W. H. , Jr.; Smith, H. K. ; Gregory, J.; Burrell , V. G., Jr.; DeVoe, M. R. ; Sojka, R. E. ; Gilbert, S.; Banks, R. ; Stolzy, L. H.; Brooks, C.; Matthews, T. D. ; Shear, T. H. , eds . The ecology and management of wetlands. Portland, OR: Timber Press; 1988. Hook, D. D.; Scholtens, J. R. Adaptations and flood tolerance of tree species. Hook, D. D. ; Crawford, R. M. M. , eds. Plant life in anaerobic environments. Ann Arbor, MI: Ann Arbor Science Series Inc.; 1978: 299-311. Hook, D. H. Waterlogging tolerance of lowland tree species of the south. South. J. Appl. Forest. 8(3):136-49; 1984. Hosner, J. F. ; Boyce , S. G. Tolerance to water saturated soil of vari- ous bottomland hardwoods. For. Sci. 8:180-186; 1962. Huenneke , L. F. ; Sharitz, R. R. Microsite abundance and distribution of woody seedlings in a South Carolina cypress -tupelo swamp. American Midland Naturalist 115(2) : 328- 335 ; 1986. Huffman, R. T. The relationship of flood timing and duration to varia- tion in selected bottomland hardwood communities of southern Arkansas. Vicksburg, MS: U.S. Army Engineer Waterways Experiment Station; 1980. Miscellaneous Paper EL-80-4. Hupp, C. R. ; Osterkamp, W. R. Bottomland vegetation distribution along Passage Creek, Virginia, in relation to fluvial landf orms . Ecological Society of America 66(3) : 670-681 ; 1985. Kangas , P. C. An energy theory of landscape for classifying wetlands. Lugo, A. E. ; Brinson, M. ; Brown, S., eds., Ecosystems of the world; Vol 15, Forested Wetlands. New York: Elsevier; 1990; 15-23. Larson, J. S.; Bedinger, M. S.; Bryan, C. F. ; Brown, S.; Huffman, R. T. ; Miller, E. L. ; Rhodes, D. G. ; Touchet, B. A. Transition from wetlands to uplands in southeastern bottomland hardwood forests. Clark, J. R. ; Benf orado , J., eds. Wetlands of bottomland hardwood forests. New York: Elsevier; 1981:225-273. Leitman, H. M. ; Sohm, J. E. ; Franklin, M. A. Wetland hydrology and tree distribution of the Apalachicola River flood plain, Florida. Washing- ton, D.C.: U.S. Geological Survey; 1984. Water-supply paper 2196-A. Mcintosh, R. P. The background and some current problems of the theo- retical ecology, Synthese 43:195-255; 1980. Mitsch, W. J.; Gosselink, J. G. Wetlands. New York: Van Nostrand Rein- hold; 1986. Mohler, C. L. An analysis of floodplain vegetation of the lower Neches drainage, southwest Texas, with some considerations on the use of regression and correlation in plant synecology. Ithaca, NY: Cornell University; 1979. 658 p. Dissertation, Dissertation Abstracts Inter- national No. 792633. 205 Patrick, W. H. , Jr.; Delaune, R. D. Nitrogen and phosphorus utilization by Spartina alterniflora in a salt marsh in Barataria Bay, Louisiana: estuarine, Coast. Mar. Sci. 4:59-64; 1976. Patrick, W. H. , Jr.; Mikkelsen, D. S. Plant nutrient behavior in flooded soil. Fertilizer technology and use, 2d ed. Madison, WI : Soil Science Society of America; 1971. Pearsall, W. H.; Mortimer, C. H. Oxidation and reduction potentials in waterlogged soils, natural waters and muds. J. Ecol . 27:483-501; 1939. Ponnamperuma, F. N. The chemistry of submerged soils. Adv. Agron. 24:29-96; 1972. Putnam, J. A.; Furnival , G. M. ; McKnight, J. S. Management and inven- tory of southern hardwoods. Washington, DC: U.S. Department of Agri- culture, Forest Service Agricultural Handbook 181:3-7; 1960. Reed, P. B. , Jr. Wetland plants of the United States of America 1988, St. Petersburg, FL: U.S. Fish and Wildlife Service; 1988. WELUT- 88/W17.01. Roelle, J. E.; Auble, G. T. ; Hamilton, D. B.; Johnson, R. L. ; Segelquist, C. A., eds . Results of a workshop concerning ecological zonation in bottomland hardwoods. Ft. Collins, CO: U.S. Fish and Wildlife Service, National Ecology Center; 1987; NEC-87/14. Ter Braak, C. J. F. ; Prentice, I. C. A theory of gradient analysis. Advances in Ecological Research 18:271-313; 1988. U.S. Department of Agriculture. National list of scientific plant names, vols I and II. Washington, DC: Soil Conservation Service; 1982. SCS-TP-159. Van Der Valk, A. G. Succession in wetlands; a Gleasonian approach. Ecology 62(3):688-696; 1981. Vester, G. The physiology of flooding tolerances in trees. Bot. Sci. Edinb. Trans. 41(4) : 556-557 ; 1972. Whittaker, R. H. Vegetation of the Great Smoky Mountains. Ecol. Monogr. 26:1-69; 1956. Whittaker, R. H. Gradient analysis of vegetation. Biol. Rev. 49:207-264; 1967. Whittaker, R. H. Direct gradient analysis: techniques. Whittaker, R. H. , ed. Ordination and classification of communities. The Hague: Junk; Handbook of vegegation science 5:7-31; 1973. Whittaker, R. H. Ordination of plant communities, the Hague; 1978; 388 p. 206 Whittaker, R. H. ; Nierlng, W. A. Vegetation of the Santa Catalina Moun- tains, Arizona: a gradient analysis of the south slope. Ecology 46:429-452; 1965. BIOGRAPHICAL SKETCH Russell Francis Theriot was born January 15, 1947, in Crowley, Louisiana. He is married to the former Patricia Anne Bales; they have two children, Sheri and Craig. As an undergraduate, he attended Louisiana State University and the University of Southwestern Louisiana. He served three years in the United States Army and then returned to his undergraduate studies in 1969 at Northwestern State University in Louisiana. He received his Bachelor of Science degree in Wildlife Management and his Master of Science degree in Botany at Northwestern State University in 1972 and 1974, respectively. Upon graduating with his Master's degree, he worked for two years in the Florida Department of Natural Resources as an environmental spe- cialist. Since 1976, he has been employed at the U.S. Army Engineer Waterways Experiment Station in Vicksburg, Mississippi, where he pres- ently serves as Director of the Wetlands Research and Technology Center. He is a member of Beta Beta Beta, the Ecological Society of Amer- ica, and the Society of Wetland Scientists. He is also a member of several national committees, including the National Technical Committee for Hydric Soils, the National Review Panel of Plant Species that Occur in Wetlands, the Interagency Committee Developing the National Wetlands 207 208 Delineation Manual, and the Wetlands Research Subcommittee of the Com- mittee on Earth and Environmental Sciences. I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Jerome V. Shifeman, Chair Professor of Forest Resources and Conservation I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Ka£herine C? Ewel Professor of Forest Resources and Conservation I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Donald A. Graetz Professor of Soil Science ,(:L*.;*srtLl I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. William T. Haller Professor of Agronomy I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. ^>l Kenneth M. Portier Associate Professor of Statistics This dissertation was submitted to the Graduate Faculty of the School of Forest Resources and Conservation in the College of Agricul- ture and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. WP,;/ May 1992 Director, Forest Resources and Conservation Dean, Graduate School UNIVERSITY OF FLORIDA 3 1262 08556 8771