FEEDING ECOLOGY OF SUBADULT GREEN SEA TURTLES IN SOUTH TEXAS WATERS. A Thesis by MICHAEL SCOTT COYNE

FEEDING ECOLOGY OF SUBADULT GREEN SEA TURTLES IN SOUTH TEXAS WATERS A Thesis by MICHAEL SCOTT COYNE Submitted to the Office of Graduate Studies of Texas A&M University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE May 1994 Major Subject: Wildlife and Fisheries Sciences

FEEDING ECOLOGY OF SUBADULT GREEN SEA TURTLES IN SOUTH TEXAS WATERS A Thesis by MICHAEL SCOTT COYNE Approved as to style and content by: Submitted to Texas A&M University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE André M. Landry, Jr. (Co-Chair of Committee) Robert B. Ditton (Co-Chair of Committee) David W. Owens (Member) Robert Brown (Head of Department) May 1994 Major Subject: Wildlife and Fisheries Sciences

iii ABSTRACT Feeding ecology of subadult green sea turtles in South Texas waters. (May 1994) Michael Scott Coyne, B.S., University of Florida Co-Chairs of Advisory Committee: Dr. André M. Landry, Jr. Dr. Robert B. Ditton Feeding ecology of green sea turtles (Chelonia mydas) at South Padre Island, Texas was characterized from April 1991 - March 1993. Ninety turtle captures in entanglement and cast-nets were comprised of 24 individuals from jetty habitat at Brazos Santiago Pass and 27 others over grassbed habitat of South Bay/Mexiquita Flats. Size range of captured turtles [22.2-81.5 cm straight carapace length (SCL)] indicated that green sea turtles use the two sites as developmental habitats. Size of turtles differed significantly between sites (jetty: 22.2-47.9 cm, mean - 31.3 cm SCL; grassbeds: 29.6-81.5 cm, mean - 44.6 cm SCL). Comparison of food items flushed from green turtle stomachs with available forage material suggested significant feeding selectivity. Turtles from jetty environs fed strictly on algae, showing a preference for Ulva fasciata, Rhodymenia pseudopalmata, Family Ceramiaceae, Bryocladia sp., and Hypnea musciformis. Grassbed turtles fed primarily on seagrasses and exhibited a preference for the least abundant taxon, Halodule wrightii. The von Bertalanffy growth interval equation was L t = 113.84(1-0.960e -0.0768t ). Age at maturity estimates ranged from 18 to 26 years. Highest growth rate was observed in spring and summer (0.62 and 0.64 cm/month, respectively) and lowest in winter (0.14 cm/month). Activity patterns varied seasonally, with increased movement during months when mean water temperature exceeded 25 C. Strong site fidelity also was observed during warmer months.

iv ACKNOWLEDGMENTS There are many to whom I am indebted for their assistance, advice, and moral support in the completion of my thesis. Lack of space and a poor memory makes it impossible to name everyone. However, I offer my sincerest gratitude to all who have made this effort possible. To the members of my committee, Dr. André M. Landry, Jr., Dr. Robert B. Ditton, and Dr. David W. Owens, I owe many thanks. Their patience and wisdom proved invaluable. Special thanks goes to Dr. André M. Landry, Jr., my principal advisor. His own drive made it possible for me to accomplish my goals and pushed me to always improve myself. My gratitude extends to the many agencies whose cooperation made it possible to complete this research. These include the National Marine Fisheries Service, U.S. Army Corps of Engineers, Texas A&M University Sea Grant Program, and U.S. Fish and Wildlife Service. Special mention goes to University of Texas Pan American Laboratory for allowing us free run of their facility. Extraordinary praise go to the turtle crew. The many brave souls that risked sharks, stingrays, and mind-numbing grass sorting in collecting and processing data: David Costa, Kelly Craig, Brett Williams, Russel O brien, Karen St. John, Vicky Poole, Leonard Kenyon, Stacie Arms, Travis Hanna, Sarah Werner, John Christensen, Randy Clark, all of the MARB 485 students, and others who lent assistance. I would like to acknowledge the Biology department in College Station and the Marine Biology department in Galveston for providing Teaching Assistantships and other sources of funding enabling the completion of this degree. Finally, I would like to recognize friends and family who said the right things at the right times, and who enabled me to enjoy every aspect of the experience: Dr. George W. and Laura J. Coyne, Christean Coyne, Matthew Polly, Nancy Mettee, Hava Berman, Denise Brown, Lynette Goodman, the Ultimate players that have touched my life, and all the graduate students with whom I ve worked.

v TABLE OF CONTENTS Page ABSTRACT...iii ACKNOWLEDGMENTS...iv TABLE OF CONTENTS...v LIST OF TABLES...vii LIST OF FIGURES...ix LIST OF APPENDIX TABLES...xi INTRODUCTION...1 STUDY AREA...5 METHODS...9 TURTLE CAPTURE AND RELATED ACTIVITIES...9 Turtle Capture...9 Visual Observations...9 Tagging Activities...10 Stomach Sample Analysis...10 HABITAT CHARACTERIZATION...10 Hydrological Monitoring...11 SCUBA Surveys...11 GROWTH MODELING...12 RESULTS...13 SEA TURTLE CAPTURE AND RELATED ACTIVITIES...13 Sea Turtle Sightings...19 STOMACH EVACUATION...23 HABITAT CHARACTERIZATION...24 Hydrological Monitoring...24 Jetty Characterization...26 Grassbed Characterization...29 HABITAT UTILIZATION...35 Recaptures...35 Tracking...35 GROWTH...36

vi Page DISCUSSION...43 POPULATION DYNAMICS...43 BEHAVIOR...46 FEEDING ECOLOGY...47 SUMMARY...50 LITERATURE CITED...51 APPENDICES...55 VITA...76

vii LIST OF TABLES Table Page 1 Sea turtle capture effort in the South Padre Island study area during April 1991 - March 1993....13 2 Number of green sea turtles taken by capture method in the South Padre Island study area during April 1991 - March 1993...14 3 Percent biomass and frequency of occurrence of identifiable phyla found in stomach contents of green sea turtles ( n = 76 ) captured in the South Padre Island study area during April 1991 - March 1993...25 4 Dominant ( 5% of total identifiable biomass) food items found in stomach contents of green sea turtles ( n = 76 ) captured in the South Padre Island study area during April 1991 - March 1993...25 5 Dominant ( 5% of total identifiable biomass) food items in green sea turtles captured from Brazos Santiago Pass and South Bay/Mexiquita Flats sites during April 1991 - March 1993...26 6 Seasonal occurrence of dominant ( 5% of total identifiable biomass) food items in green sea turtles captured from Brazos Santiago Pass and South Bay/Mexiquita Flats sites during April 1991 - March 1993...27 7 Average monthly air and water temperature and salinity in the South Padre Island study area during April 1991 - March 1993. Detailed monthly hydrological data are presented in Appendix Table 3...28 8 Epiphytic and encrusting taxa found at Brazos Santiago Pass habitat characterization stations during summer 1991 and 1992...30 9 Date, location, and straight carapace length (SCL) of repeated green sea turtle captures from Brazos Santiago Pass during April 1991 - March 1993. See Appendix Table 1 for a complete description of each capture....37 10 Date, location, and straight carapace length (SCL) of repeated green sea turtle captures from South Bay/Mexiquita Flats during April 1991 - March 1993. See Appendix Table 1 for a complete description of each capture...38 11 ID number, capture and release date, straight carapace length (SCL), and date of last contact of green sea turtles tracked at Brazos Santiago Pass by National Marine Fisheries Service personnel during 1992...39 12 Average annual growth rate between each capture of 10-cm size classes (N = sample size) for green sea turtles from the South Padre Island study area...40

viii Table Page 13 Average annual growth rate across seasons ( N = sample size ) for green sea turtles in the South Padre Island study area...40 14 Estimated values for the parameters a, k, and b from non-linear regression of von Bertalanffy growth interval equations for green turtles...41 15 Estimated values of asymptotic length (a), intrinsic growth rate (k), and age at sexual maturity from non-linear regression of the von Bertalanffy growth interval equation for green turtles at various locations...45

ix LIST OF FIGURES Figure Page 1 Number of stranded green sea turtles within NMFS statistical zones 1 through 21 during April 1980 - May 1991. Data provided by NMFS Sea Turtle Stranding and Salvage Network...4 2 Sea turtle capture and habitat characterization study area...6 3 Brazos Santiago Pass sampling zones A - Z....7 4 South Bay and Mexiquita Flats sampling stations...8 5 Location and number of green sea turtles taken at Brazos Santiago Pass during April 1991 - March 1993...15 6 Location and number of green sea turtles taken at South Bay/Mexiquita Flats sampling stations...16 7 Length frequency at first capture of green sea turtles from Brazos Santiago Pass and South Bay/Mexiquita Flats during April 1991 - March 1993...17 8 Number of green sea turtle captures and stationary netting hours at South Bay/Mexiquita Flats during each month....18 9 Number of green sea turtle captures and days of effort at South Bay/Mexiquita Flats during each hour...18 10 Number of sea turtle sightings per hour of observation effort at Brazos Santiago Pass during April 1992 - March 1993. Total number of monthly sightings is shown on top of histogram bars...19 11 Average number of sea turtle sightings per hour of observation effort at Brazos Santiago Pass jetty-zone observation posts (A - Z) during April 1992 - March 1993...20 12 Average number of sea turtle sightings per observation hour at Brazos Santiago Pass from April 1992 - March 1993 (8 = 0800-0900 hrs, 9 = 0900-1000 hrs, etc...)...21 13 Number of sea turtle sightings at increasing distances from Brazos Santiago Pass jetties during April 1992 - March 1993...22 14 Number of individual sea turtles sighted per month at Brazos Santiago Pass during April 1992 - March 1993...23 15 Average monthly air and surface water temperature (C) and salinity (ppt) in the South Padre Island study area from April 1991 - March 1993....29

x Figure Page 16 Percent composition of sea grasses, epiphytic algae, and bryozoans at habitat characterization stations in South Bay and Mexiquita Flats...32 17 Percent composition of dominant ( 5%) seagrass and algae taxa at Mexiquita Flats station 15 during April 1992 - February 1993...33 18 Percent composition of dominant ( 5%) seagrass and algae taxa at South Bay station 21 during August 1991 - March 1993 (* June 1992 samples were taken from an adjacent grassbed separated from station 21 by a narrow channel).....34 19 Percent composition of dominant ( 5%) seagrass and algae taxa at South Bay station 28 during September 1991 - March 1993...34 20 Predicted growth curve for green sea turtles from the South Padre Island study area during April 1991 - March 1993. Dotted lines indicate size classes for which no data were available....42

xi LIST OF APPENDIX TABLES Table Page 1 Capture statistics for green sea turtles from the South Padre Island study area during April 1991 - March 1993...56 2 Stomach contents from green sea turtles captured in the South Padre Island study area during April 1991 - March 1993...58 3 Hydrological data from the South Padre Island study area during April 1991 - March 1993...65

1 INTRODUCTION All sea turtles in U.S. waters are listed as either endangered or threatened (Public Law 93-205), and as such, are covered under the Endangered Species Act of 1973 which provides for their conservation, protection, and propagation. Congress amended the Endangered Species Act in 1988 (Public Law 100-478) to authorize an independent review of scientific information on sea turtles by the National Academy of Sciences (NAS) in order to establish a sound technical basis for protecting these species (Magnuson et al., 1990). The subsequent review reported that human activities, including commercial fisheries, dredging, boat collisions, oil platform removal, and discard of plastics and debris, are major causes of sea turtle mortality. The committee s recommendations to optimize research on these stocks included identifying and characterizing the status, size, age structure, distribution, and concentration, of the green sea turtle (Chelonia mydas). The green turtle is the only herbivorous sea turtle, and, as such, its feeding habits have been the subject of much research. Descriptive information on feeding habits has been collected from the Bahamas (Bjorndal, 1980), U.S. Virgin Islands (Ogden et al., 1983; Williams, 1988), Nicaragua (Mortimer, 1981), and Mosquito Lagoon, Florida (Mendonca, 1983). A review of the literature (Bjorndal, 1985) suggests that green turtles feed on either sea grasses or algae, possibly as a result of their hindgut fermentation. Because the complex carbohydrates found in sea grasses and algae are different, a different microflora may be required to digest each efficiently. Cellulose, the major structural carbohydrate in sea grasses, is present in only small amounts in algae (Percival, 1964), most of which contain complex structural carbohydrates such as glucan, mannan, xylan, agar, carrageenan, alginic acid, and uronic acid (Chapman and Chapman, 1973). However, it has been suggested that gut microflora may change over time allowing for long term adjustment to both grass and algae diets (Bjorndal, 1980). The green turtle forages primarily on sea grasses in most of its range (Hirth, 1971), with algae making up the bulk of the diet where seagrasses are lacking, e.g., off the coast of Brazil, Tahiti, and Hawaii (Balazs, 1979a, b), Galapagos Islands (Pritchard, 1971), and South African coast (Hughes, 1974). In addition, some locales support green turtle colonies that feed mainly on sea grasses within a few kilometers of others This thesis follows style and format of the journal Copeia

2 that forage primarily on algae. Such conditions exist along the west coast of Honduras (Carr, 1952), Gulf of California (Felger and Moser, 1973), Fiji (Hirth, 1971), Gulf of Aden (Hirth et al., 1973), and Torres Strait of Australia (Garnett and Murray, 1981). Green turtles feed predominantly on three seagrass species: Thalassia testudinum, Syringodium filiforme, and Halodule wrightii. Syringodium was found to be twice as abundant as Halodule from green turtles studied at Mosquito Lagoon, Florida (Mendonca, 1983). Thalassia was the primary dietary component for green turtles in Bahamian waters (Bjorndal, 1980, 1985) and in the U.S. Virgin Islands (Ogden et al., 1983; Williams, 1988). Green turtles also forage on green, brown, and red algae. Ferreira (1968) found red algae dominated the diet of Brazilian green turtles. Only red and green algae were found in stomachs of 26 green turtles from Torres Strait, Australia (Garnett and Murray, 1981). Green turtles in Tokelau, south central Pacific, fed largely on green algae, with some brown algae (Balazs, 1983). Two species of green and red algae were major dietary components of green turtles near islands of the Hawaiian Archipelago. Three green, one red and one brown alga comprised the bulk of green turtles diet in the northwestern islands of the same archipelago (Balazs, 1980). Brown algae has only been reported as a major dietary component in green turtles from the Ogasawara Islands, Japan (Kurata et al., 1978). It is uncertain what roles feeding selectivity and relative abundance of different forage species play in determining green turtle diet. Ferreira (1968) attributed the high incidence of red algae in green turtle stomachs solely to greater abundance of red algae in Brazilian feeding areas. Conversely, Balazs (1980) presented evidence that both factors influence feeding habits of Hawaiian green turtles. Neck (1978) described historical occurrence of green turtles in Texas waters during the late 1800 s in a review of the journals of geologist Robert Penrose. These journal entries indicated that green turtles may have nested in abundance on beaches adjacent to the Rio Grande during 1889. Doughty s (1984) review of data from turtle canneries at Indianola, Fulton, Corpus Christi, and Point (Port) Isabel presents strong evidence that green turtles were abundant in inshore Texas waters during the 1800 s. Landings estimates for Texas indicate that the fishery peaked in 1890 when 265,000 kg of turtle (approximately 2,159 turtles) were processed. Thereafter, a steady decline in turtle landings resulted in all canneries closing or moving to Mexico by 1897. The last year for which data were available indicated 680 kg (approximately 6 turtles) were

3 landed in 1927. Rabalais and Rabalais (1980) reported 10 green turtles stranded along the Texas coast from 1976 to 1979, five of which were juveniles heavily fouled with oil from the August 1979 Ixtoc I well blowout. More recently, Shaver (1990a) reported the capture of one green turtle in the Laguna Madre near Mansfield Pass and 40 captures of 25 turtles at the Mansfield Pass jetties during June 1989 - May 1990. Additionally, Shaver (1990b) reported 45 green turtles cold stunned in the Laguna Madre during February 1989. Information concerning feeding behavior is restricted to gut contents of a few stranded animals (Stanley, unpublished). The limited data available on green turtles in Texas waters, especially those from near shore, mandate research on this species natural history to improve development of management decisions (i.e. regulation of commercial fisheries, dredging, oil platform removal, and habitat protection). Recent sightings near jetties and in bays along the lower Texas coast (Williams and Manzella, 1990) indicate that green turtles are moderately abundant in South Texas waters. In addition, data collected by the National Marine Fisheries Service s (NMFS) Sea Turtle Stranding and Salvage Network (Fig. 1) indicate the lower Laguna Madre is second only to the Florida Keys in green turtle strandings. These data confirm that South Texas waters are ideal for the study of green sea turtles. This study was conducted to generate information on feeding ecology of green sea turtles from South Texas waters in order to: 1) identify food items consumed and feeding preferences; 2) delimit foraging areas; 3) assess seasonal variations in diet and possible causes for this variation; and 4) compare and contrast size composition and growth of turtles captured in different foraging areas.

4 0 1-9 Number of stranded turtles 10-25 26-49 50-99 100 + TX 19 20 21 18 17 MS LA 12 13 16 15 14 11 AL 10 9 8 5 4 7 6 FL Gulf of Mexico 3 2 1 Figure 1. Number of stranded green sea turtles within NMFS statistical zones 1 through 21 during April 1980 - May 1991. Data provided by NMFS Sea Turtle Stranding and Salvage Network.

5 STUDY AREA The study was conducted in inshore waters adjacent to South Padre Island, Texas from April 1991 through March 1993 (Fig. 2). Specifically, jetty habitat at Brazos Santiago Pass (BSP) and grassbed habitats in South Bay and Mexiquita Flats were sampled. Brazos Santiago Pass is a narrow (92-m wide) channel between the south end of South Padre Island and the north end of Brazos Island with a maximum depth of 11.6 m. Two 1.5-km long granite mound structures known as the North and South Jetties border north and south sides of BSP at its entrance into the Gulf of Mexico. Shallow, relatively flat coves (Dolphin and Barracuda Coves along the north and south sides, respectively) with a hard sand bottom and average depth of 2.1 m are located inside the west end of each jetty. The Gulf side of each jetty exhibits a gentle-sloping, barren, hard-sand-bottom beach out to a maximum depth of 5.5 m. The BSP area was divided into 40 100-m long zones to identify sampling locations (Fig. 3). South Bay and Mexiquita Flats consisted of grassbed and channel habitats along the easternmost reaches of the Brownsville Ship Channel (BSC) between Channel Markers 28 and 16 (Fig. 4). Grassbed habitats in South Bay and Mexiquita Flats were less than 1.5-m deep and consisted of a silt and hard sand substrate. The BSC was 2.25-km long, 61.0-m wide with an average maximum depth of 11.0 m and was bordered on the south and north by seagrass habitats of South Bay and Mexiquita Flats, respectively. BSC channel substrate consisted of mud with varying proportions of sand, clay, and silt extending to grassbed habitat at a depth of 1.8 m. This area contained six sampling stations.

6 N TEXAS Galveston Bay LA Matagorda Bay Corpus Christi Bay GULF OF MEXICO Lower Laguna Madre MEXICO Port Isabel Mexiquita Flats South Bay Padre Island Brazos Island Brownsville Ship Channel Brazos Santiago Pass Figure 2. Sea turtle capture and habitat characterization study area.

Z Y South Padre Island N Dolphin Cove X W V U T S R Q P O N M North Jetty Brazos Santiago Pass J I H G Brazos Island F E D C B A South Jetty K Barracuda Cove Jetty Characterization Stations L Gulf of Mexico Figure 3. Brazos Santiago Pass sampling zones A - Z. 7

8 31 South Padre Island Brownsville Ship Channel Brazos Santiago Pass Long Island Mexiquita Flats 15 29 22 28 16 Brazos Island Gulf of Mexico 21 South Bay 30 Sampling Station # Channel Marker Figure 4. South Bay and Mexiquita Flats sampling stations.

9 METHODS TURTLE CAPTURE AND RELATED ACTIVITIES Turtle Capture: Turtle capture was accomplished with 91.4-m long entanglement nets of different depth and mesh size specifications deployed in two configurations. These nets were 3.7- or 7.4-m deep with 12.7-cm bar mesh of #9 twisted nylon or 4.9-m deep with 25.4-cm bar mesh of #9 twisted nylon. Jetty, grassbed and channel habitats of BSP and South Bay/Mexiquita Flats (SB/MF) were sampled during the day with one to four stationary entanglement nets set adjacent to one another for 8 to 12 hours. In addition, a more active capture method of encircling turtles with entanglement nets was deployed at jetty habitats when turtles displayed certain behavioral traits. These traits included: 1) exhibiting a predictable pattern of surfacing and sounding; and 2) surfacing occurs no more than 10 m from and along a limited length of jetty. Encirclement sets were deployed from a boat with the net secured to the jetty creating a semi-circle around a turtle s expected surfacing spot. Standard bait cast nets (2-m diameter) were utilized to capture turtles in areas unfavorable for stationary or encirclement netting. Ability to capture turtles with cast nets was dependent on a turtle s behavior and proximity to the jetty. Cast nets were utilized only when turtles exhibited frequent and extended surface intervals within 5 m of the jetty proper. Any day in which multiple attempts were made to capture sea turtles using a cast-net was considered to be a cast-netting day. Visual Observations: Observations were initially conducted throughout the BSP study area in April 1992 to identify sea turtle occurrence and potential capture sites. These observations continued at random through June 1992. Thereafter, a systematic observation protocol was deployed at all North and South Jetty zones to generate data on number of sightings and individual turtles seen per hour of observation as well as to characterize spatial distribution of turtles in relation to jetty and channel habitats. Visual sightings were conducted along the North and South Jetties to locate turtles, aid in their capture and define the habitats they frequented. Jetties were divided into 38 100- m lengths and each tip for a total of 40 sampling zones at BSP (Fig. 3). Observations were made by 2-5 individuals each surveying a randomly chosen jetty zone for one hour at random times of day from July 1992 through March 1993. Species, date, time, location, and any observed behaviors were recorded each time a turtle was sighted.

10 Tagging Activities: Immediately following capture, all turtles were transported to the University of Texas Pan American Coastal Studies Laboratory on South Padre Island where they were held in a 3.0-m diameter fiberglass holding tank for 24 to 72 hours before being tagged and released. All turtles were measured to the nearest 0.1 cm, photographed and tagged at the lab. Straight line carapace length and width were measured with a forester s caliper. Over-the-curve carapace length and width were measured with a soft measuring tape. Turtles were tagged with an inconel flipper tag, provided by NMFS, on the trailing edge of each fore flipper and released at the same location as captured. Tagging data were submitted to NMFS (Miami) on two data forms entitled NMFS/SEFC Marine Turtle Tagging Data (Rehabilitated, Netted or other Release) and NMFS/SEFC Marine Turtle Tagged/Recapture Data. Turtles meeting certain size and weight criteria were provided to NMFS (Galveston) and Texas A&M University (TAMU) personnel for attachment of sonic and radio transmitters and subsequent tracking. Stomach Sample Analysis: A stomach evacuation technique developed by Balazs (1980) was used to flush the foregut of each turtle captured to obtain material for food habit analysis. In short, sea water was flushed into the foregut through a plastic tube introduced into the esophagus. Food material was backwashed out through the mouth and collected in a plastic tub. All food material obtained was preserved in 10% formalin, labeled and held for laboratory analysis. Each sample was sorted in the laboratory and observed under a compound microscope to identify vegetation and animal matter to the lowest possible taxon. Resulting data were used to analyze frequency of occurrence of food items. A one-way analysis of variance (ANOVA) was used to determine seasonal variability. HABITAT CHARACTERIZATION Turtles captured during netting activities across the study area, meeting size and weight criteria, were released in the same area in which they were captured, and were tracked by the aforementioned personnel. Tracking data were used to pinpoint habitats occupied by turtles. These habitats were then characterized by on-site efforts which consisted of: 1) hydrological monitoring; and 2) SCUBA surveys. A description of these characterization techniques follows.

11 Hydrological Monitoring: Surface and bottom measurements (depth permitting) were taken whenever a netting site was occupied to characterize hydrological conditions. Temperature, salinity, and conductivity were measured (to the nearest 0.01-C, -ppt, and - mω/s, respectively) with a Beckman Instrument Company SCT meter; visibility was measured to the nearest 0.1 m with a Secchi disk; and tidal flow was estimated by field personnel. SCUBA Surveys: Turtle capture locations and telemetric tracking by NMFS and TAMU personnel resulted in a total of six sampling stations. Two stations were in South Bay, one at the mouth of South Bay, and three in Mexiquita Flats (Fig. 4). Each station was characterized initially by a SCUBA survey to describe habitat and potential food sources available to sea turtles. This survey consisted of several tasks: 1) subsurface visual observations; 2) quantitative transects; and 3) sample quadrates (flora and fauna). Subsurface visual observations were conducted by two to four SCUBA divers to describe the habitat (i.e. channel, grassbed, barren-bottom) and assess prevailing water conditions (i.e. visibility, current). This assessment dictated which characterization task(s) could be conducted and the number of samples taken for each task. Quantitative transects were deployed to characterize available forage species. Three 25-m transect lines laid parallel to one another approximately 20-m apart were visually surveyed by two SCUBA divers swimming down either side of the line. Each diver recorded all organisms and habitat characteristics along a 1-m corridor. Quadrates were utilized to characterize prey availability, vegetative cover and sediment texture. After swimming transects were completed, 0.25-m 2 quadrates were placed randomly within each site and contents recorded and collected. Algae and sea grasses within the 0.25-m 2 quadrate were cut 2.54 cm above the sediment, bagged, labeled and frozen or preserved in 10% formalin and returned to the laboratory for analysis. Six quadrates were deployed at each SB/MF station every two months to determine seasonal variation. Each vegetation sample was rinsed with sea water and separated to the lowest possible taxon. Separated taxa were placed in paper bags, labeled and stapled shut. Each paper bag was dried in an oven for approximately 40 hr at 82 C. After drying, each bag and its contents were weighed and the mass of an empty, stapled bag subtracted. The resulting weight was used as dry mass for each species. Percent biomass was determined by dividing the dry mass of each species in a quadrate by the dry mass of all taxa in the quadrate. An ANOVA of the dry mass of dominant species ( 5% biomass) at

12 stations 15, 21 and 28, was used to determine whether changes observed within quadrates were due to seasonal variation. Vegetation also was collected at BSP jetty stations to develop a species list. Nine of 40 zones at the BSP jetties (Fig. 3) were designated as characterization stations. Five stations were sampled during summer 1991 and 1992 (P outside, P inside, A outside, E, and W). The other four stations (K, G-H, A-B inside, and Y) were sampled only during summer 1992. GROWTH MODELING Preliminary attempts to develop a growth curve for green turtles from the study area follow methodology of Frazer and Ehrhart (1985). Straight line carapace measurements at capture and recapture and time intervals between capture and recapture were fit to the von Bertalanffy growth interval equation with JMP v. 2 non-linear least squares regression procedure (SAS Institute Inc., 1989). The general von Bertalanffy equation used was: L = a(1 - be -kt ), (1) where L is carapace length, a is asymptotic length, b is a parameter related to length at hatching, e is the base of the natural logarithm, k is the intrinsic growth rate, and t is age in years. The von Bertalanffy growth interval equation (Fabens, 1965) for recapture data was: L r = a - (a - L c )e -kd, (2) where L r is carapace length at recapture, L c is length at first capture, and d is time in years between capture and recapture. Once an estimate was available for a, equation 1 was rearranged to calculate a value for b if the size of the organism at birth (L 0 ) was known (Fabens, 1965). At hatching (t = 0), equation 1 simplifies to: L 0 = a(1 - b). (3) Rearranging equation 3 yielded: b = 1 - L 0 /a. (4)

13 RESULTS SEA TURTLE CAPTURE AND RELATED ACTIVITIES Capture activities consisted of 1,649 hours of stationary entanglement netting effort at 7 locations, 45 encirclement-net attempts, and 7 days of cast-netting effort (Table 1). All stationary netting effort during 1991 at BSP took place in zones W and K, in Dolphin and Barracuda Coves, respectively. Most capture efforts at BSP during 1992 and 1993 were deliberate attempts to catch observed turtles through encirclement or cast-netting. Entanglement netting efforts at SB/MF were divided between stations 15 (743 hours), 28 (74 hours), and 29 (74 hours), with an additional 5 hours of effort in South Bay proper. Encirclement netting was first attempted at BSP in July 1991 and became the standard sampling technique for that site. Four encirclement attempts at SB/MF during 1993 were concerted efforts to capture turtles previously equipped with radio and sonic telemetric gear. The cumulative netting effort yielded 90 green sea turtle captures, 48 from BSP and 42 from SB/MF (Table 2 and Appendix Table 1). Thirty-nine recaptures involved 15 individuals and produced a 43.3% recapture rate. Thirty-six individuals were captured once, four twice, five on three occasions, two on four occasions, three on five occasions, and one individual was captured eight times. Table 1. Sea turtle capture effort in the South Padre Island study area during April 1991 - March 1993. Brazos Santiago Pass South Bay/Mexiquita Flats 1991 1992 1993 1991 1992 1993 Stationary Netting (hours) 454 14 2 360 605 214 Encirclement Netting (attempts) 7 31 3 0 0 4 Cast-Netting (days) 0 7 0 0 0 0

14 Table 2. Number of green sea turtles taken by capture method in the South Padre Island study area during April 1991 - March 1993. Brazos Santiago Pass South Bay/Mexiquita Flats 1991 1992 1993 1991 1992 1993 Stationary Netting 3 5 1 12 22 4 Encirclement Netting 1 26 4 0 0 3 Cast-Netting 0 5 0 0 0 0 Incidental 2 1 0 0 1 0 Over 83% of all turtle captures (40) at BSP occurred along the South Jetty (Fig. 5). Of these, 31 captures were made along a 500-m length of jetty between zones G and L. This area was characterized by submerged granite blocks on a silt bottom extending to the channel from zones G to K changing to small boulders on a sand/silt bottom out to the channel toward the cove. One capture was made between zones A and B (channel side) consisting of submerged granite blocks dropping quickly into the channel, and eight captures were along the outermost 200 m (Gulf side) made up of submerged and partially submerged granite blocks on a sand bottom extending along the beachfront. Captures along the North Jetty were limited to seven green turtles taken from the channel side - four at zone P (3 captures were of one turtle) and three along a 200 m span between zones R through T. Gulf and channel sides of the North Jetty exhibited physical attributes similar to those along corresponding sides of the South Jetty. Incidental captures included one green turtle foul-hooked by a fisherman, another found entrapped between granite blocks near the end of North Jetty, and two hand captured while feeding on algae along the jetty. Mexiquita Flats stations 15 and 29 along the BSC and opposite the mouth of South Bay yielded 37 turtle captures from grassbed locales (Fig. 6). Two captures were made on the opposite side of the Brownsville Ship Channel (BSC) at station 28 while one turtle was caught in a bay shrimpers trawl in the BSC between channel markers 16 and 22. Two additional turtles were taken in encirclement nets in the northwest corner of Mexiquita Flats (station 31).

Z Y South Padre Island N Dolphin Cove X W V U T S R Q P O N M North Jetty 1 1 1 4 Brazos Santiago Pass Barracuda Cove 5 11 6 L 1 K 2 J 1 1 I 3 H 1 G Brazos Island F E D C B 1 4 A 4 Gulf of Mexico South Jetty # Capture Location and Number Taken Figure 5. Location and number of green sea turtles taken at Brazos Santiago Pass during April 1991 - March 1993. 15

16 Long Island 2 31 Mexiquita Flats 16 29 S h i CM 22 1 21 e 15 n s v i l l 2 28 B r o w p CM 16 C h a n n e 21 Brazos Island l # Sampling Station North # Capture Location South Bay Figure 6. Location and number of green sea turtles taken at South Bay/Mexiquita Flats sampling stations.

17 Straight carapace length at initial capture for green sea turtles taken at BSP ranged from 22.2 to 47.9 cm and averaged 31.3 cm with a S.D. of 5.25. Turtles captured at SB/MF ranged from 29.6 to 81.5 cm and averaged 44.6 cm with a S.D. of 11.58. A pooled t-test of these data indicate a statistical difference ( t = -5.159 ) in the SCL of turtles captured from the two sites (Fig. 7). Monthly stationary netting efforts at SB/MF ranged from 34 hours during April to 236 hours during October (Fig. 8). Days of effort made in stationary netting at SB/ MF during each daytime hour ranged from 17 at 0700 hrs, peaked at 124 between 1100 and 1500 hrs, down to 8 at 1900 hrs (Fig. 9). Over half of all turtle captures (48) occurred between 0900 and 1300 hrs. Number of captures was closely related to effort except for low captures rates between 1000 and 1100 hrs, from 1300 to 1700 hrs, and no captures between 1800 and 1900 hrs. Number of Individuals 16 14 12 10 8 6 4 2 0 Brazos Santiago Pass South Bay/Mexiquita Flats 20.0-29.9 30.0-39.9 40.0-49.9 50.0-59.9 60.0-69.9 70.0-79.9 80.0 + Straight Carapace Length (cm) Figure 7. Length frequency at first capture of green sea turtles from Brazos Santiago Pass and South Bay/Mexiquita Flats during April 1991 - March 1993.

18 8 6 Captures Hours 250 200 Captures 4 150 100 Hours 2 50 0 J F M A M J J A S O N D Month 0 Figure 8. Number of green sea turtle captures and stationary netting hours at South Bay/Mexiquita Flats during each month. Captures Days 8 125 6 100 Captures 4 75 50 Days 2 25 0 7 8 9 10 11 12 13 14 15 16 17 18 19 Time of Day 0 Figure 9. Number of green sea turtle captures and days of effort at South Bay/ Mexiquita Flats during each hour.

19 Sea Turtle Sightings: Observations at BSP during April 1992 - March 1993 resulted in 448 hours of effort producing 447 sea turtle sightings. Green sea turtles constituted nearly all of these sightings. Monthly number of sightings ranged from 9 in December to 85 in August (Fig. 10). The rate at which turtles were sighted varied from less than 1.0 sighting/observation-hour during the cooler months of April - May and November - March to 2 sightings/observation-hour in June and August (Fig. 10). Frequency of turtle sightings and overall sighting rates were highest near South Jetty environs during April 1992 - March 1993 (Fig. 11). Overall, highest sea turtle sightings/observation-hour statistics were reported from the easternmost section of South Jetty, specifically zones A - D. Rates exceeding 5.0 sightings/observationhour characterized gulf waters adjacent to zone A while rates at the jetty tip, along the inside of zones A, B, and D, and outside zones C and B ranged from 1.2 to 3.2 sightings/observation-hour. 2.5 2 83 85 78 Sightings/Hour 1.5 1 21 47 38 0.5 15 20 9 26 15 10 0 APR MAY JUN JUL AUG SEP OCT NOV DEC JAN FEB MAR Month Figure 10. Number of sea turtle sightings per hour of observation effort at Brazos Santiago Pass during April 1992 - March 1993. Total number of monthly sightings is shown on top of histogram bars.

Z Y South Padre Island N Dolphin Cove X W V U T S R Q P O N M North Jetty Brazos Santiago Pass # of Sightings/Hour 0.9 1.0-1.9 2.0-2.9 3.0-4.9 5.0 Barracuda Cove L K J I H G Brazos Island F E D C B A Gulf of Mexico South Jetty Figure 11. Average number of sea turtle sightings per hour of observation effort at Brazos Santiago Pass jetty-zone observation posts (A - Z) during April 1992 - March 1993. 20

21 Another area yielding a considerable number of sightings was the westernmost end of South Jetty, where zones H to L exhibited 1.3 to 2.3 sightings/observation-hour. Turtle sightings along the North Jetty were patchy at best, with observation rates over 1.0 sighting/observation-hour only along zone U, inside zones S and P, and outside zones O and R. An analysis of observation rates in relation to time of day indicated a high of 1.8 sightings/observation-hour occurred between 1800 and 1900 hrs (Fig. 12). Observation rates started with a low of 0.3 sightings/observation-hour at 0800 hrs and increased to 1.4 at 1000 hrs. Activity peaked again at 1300 and 1600 hrs with rates > 1.3 sightings/observation-hour. 2 1.5 Sightings/Hour 1 0.5 0 8 9 10 11 12 13 14 15 16 17 18 Time of Day Figure 12. Average number of sea turtle sightings per observation hour at Brazos Santiago Pass from April 1992 - March 1993. (8 = 0800-0900 hrs, 9 = 0900-1000 hrs, etc...)

22 A distributional analysis of sightings in relation to distance from the jetties provided a well defined pattern of sea turtle affinity for jetty habitats (Fig. 13). Fourhundred-thirty sightings could be assigned to distance-from-jetty zones including 0-5, 5-15, 16-30 and > 30 m. Sightings reported to cross into more than one distancefrom jetty zone or in which the observer was unable to judge distance from jetty were excluded. Over 61% of all turtle sightings occurred within 5 m of jetty habitat. The frequency of sightings beyond 5 m declined to approximately 30% for distances out to 15 m, and was less than 10% for all greater distances. A total of 170 turtles was individually recognized among the 447 sightings made during April 1992 - March 1993 (some individually recognized turtles were the same turtles observed from month to month or at different zones during one month). Area sea turtle abundance, as estimated from these sightings, varied with season (Fig. 14). Monthly abundance estimates during cooler periods of April - May and November - March were 15 turtles while those for summer - early fall ranged from 17 to 26 turtles. Peak abundance was recorded in June, with 26 individual turtles recognized. 300 Number of Turtle Sightings 250 200 150 100 50 0 0-5 6-15 16-30 >30 Distance from Jetty (m) Figure 13. Number of sea turtle sightings at increasing distances from Brazos Santiago Pass jetties during April 1992 - March 1993.

23 30 Number of Turtles 20 10 0 Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Month Figure 14. Number of individual sea turtles sighted per month at Brazos Santiago Pass during April 1992 - March 1993. STOMACH EVACUATION Seventy-six stomach evacuation samples were analyzed from 47 green turtles captured in the study area (Appendix Table 2). All stomach samples were small due to inherent limitations of the evacuation protocol and procedures taken to reduce stress on turtles. Consequently, the total mass of individual samples ranged from 0.01 to 1.55 g and averaged 0.13 g. Although 42 of 76 samples each contributed less than 0.1 g of material and a majority of food items (68.19% by mass) were unidentifiable because of their advanced state of digestion, remnants from a vast array of these food items could be distinguished in green turtle stomachs. These food items belonged to 24 taxonomic groups comprising seven different divisions/phyla and various debris (plastic, metal, etc.).

24 The most commonly encountered food group in green turtle stomachs was the red algae Division Rhodophycophyta (Table 3 and Appendix Table 2). These red algae exhibited the greatest species diversity (ten taxa) and frequency of occurrence (67.53%) recorded among food items and collectively comprised 54.63% of the identifiable biomass in stomach samples. Highest frequency of occurrence among identifiable food items (Table 4) was reported for the seagrass Halodule wrightii (31.58%) and the red alga Bryocladia sp. (28.95%). Food items that made up the greatest proportion of the total biomass included miscellaneous (not identifiable at least to family) red algae (20.13%) and Halodule wrightii (15.97%). There were notable differences between the stomach contents of green sea turtles captured from BSP and SB/MF sites (Table 5). Dominant food items in BSP turtles consisted of four red algae and one green alga. Miscellaneous red algae made up the bulk of dry biomass (20.87%) and occurred in 30 (75.00%) samples. Other dominant taxa included Hypnea musciformis (16.60% in 8 samples), Bryocladia sp. (11.88% in 22 samples), family Ceramiaceae (9.04% in 17 samples), Rodymenia pseudopalmata (8.67% in 4 samples), and Ulva fasciata (8.19% in 21 samples). Dominant food items in SB/MF turtles consisted of two sea grasses and miscellaneous red algae. Halodule wrightii made up the greatest proportion (35.20%) of biomass and occurred in 20 (55.56%) samples. Other dominant taxa included miscellaneous sea grasses (22.84% in 6 samples), miscellaneous red algae (18.40% in 11 samples), and Syringodium filiforme (11.85% in 11 samples). An ANOVA indicated no statistically significant variation in food items across seasons (Table 6) from turtles at BSP (F = 1.28, p = 0.3166) and slight variation in turtles at SB/MF (F = 3.42, p = 0.0526). Halodule wrightii and miscellaneous grasses were the dominant species during each season sea turtles were taken at SB/MF. Turtles at BSP selected miscellaneous red algae, Hypnea musciformis, or Bryocladia sp.. HABITAT CHARACTERIZATION Hydrological Monitoring: The study period was characterized by homogenous average monthly surface and bottom (when available) water temperatures and salinities (Table 7). Surface water temperature during April 1991 - March 1993 ranged from 13.65 C on 12 January 1993 to 32.17 C on 5 July 1991. Highest average monthly water temperature (Figure 15) occurred during September of both 1991 and 1992 (29.1 and 28.7 C,

25 Table 3. Percent biomass and frequency of occurrence of identifiable phyla found in stomach contents of green sea turtles (n = 76) captured in the South Padre Island study area during April 1991 - March 1993. Percent Frequency of Percent Food Items Biomass Occurrence Frequency Anthophyta 34.19 47 61.84 Chlorophycophyta 5.11 24 31.58 Phaeophycophyta 1.60 7 9.21 Rhodophycophyta 54.63 52 68.42 All other items (9) 4.47 Table 4. Dominant ( 5% of total identifiable biomass) food items found in stomach contents of green sea turtles (n = 76) captured in the South Padre Island study area during April 1991 - March 1993. Percent Frequency of Percent Food Items Biomass Occurrence Frequency Anthophyta Syringodium filiforme 7.35 17 22.37 Halodule wrightii 15.97 24 31.58 Miscellaneous grass 9.90 14 18.42 Rhodophycophyta Hypnea musciformis 9.90 8 10.53 Rhodymenia pseudopalmata 5.11 4 5.26 Family Ceramiaceae 5.43 19 25.00 Bryocladia sp. 6.71 22 28.95 Miscellaneous red algae 20.13 41 53.95 All other taxa (23) 19.49 respectively). Lowest average monthly water temperature (15.66 C) occurred during January 1993. Surface salinity ranged from 25.28 ppt on 11 June 1992 to 40.00 ppt on 21 November 1991 (Appendix Table 3). Highest average monthly salinities occurred during July 1991 (36.00 ppt) and November 1992 (38.33 ppt). Lowest average monthly salinity occurred during May 1991 and April 1992 (29.56 and 29.58 ppt, respectively).

26 Table 5. Dominant ( 5% of total identifiable biomass) food items in green sea turtles captured from Brazos Santiago Pass and South Bay/Mexiquita Flats sites during April 1991 - March 1993. Percent Frequency of Percent Food Items Biomass Occurrence Frequency Brazos Santiago Pass n = 40 Chlorophycophyta Ulva fasciata 8.19 21 52.50 Rhodophycophyta Hypnea musciformis 16.69 8 20.00 Rhodymenia pseudopalmata 8.67 4 10.00 Family Ceramiaceae 9.04 17 42.50 Bryocladia sp. 11.88 22 55.00 Misc. Red Algae 20.87 30 75.00 All other taxa (22) 24.66 South Bay/Mexiquita Flats n = 36 Anthophyta Syringodium filiforme 11.85 11 30.56 Halodule wrightii 35.20 20 55.56 Misc. grass 22.84 6 16.67 Misc. red algae 18.40 11 30.56 All other taxa (14) 11.71 Average monthly air temperature ranged from 16.67 C in December 1992 to 31.00 C in September 1992. Jetty Characterization: Jetty habitat at Brazos Santiago Pass consisted of three distinct biological zones. The first zone, extending 2 m out from the jetties to a depth of 1 m, was characterized by lush algae growth covering large granite boulders of uniform size. Underlying this nearsurface zone was a second stratum extending 10 m from the jetty to a depth of 3 m. This second biological zone contained various-sized boulders exhibiting only scattered algae growth. An exception to this pattern existed from zones W to Z along the North Jetty where the intermediate zone consisted only of a flat sandy bottom extending to the channel proper. The final zone encompassed deeper waters

27 Table 6. Seasonal occurrence of dominant ( 5% of total identifiable biomass) food items in green sea turtles captured from Brazos Santiago Pass and South Bay/ Mexiquita Flats sites during April 1991 - March 1993. Food Items Percent Biomass Winter Spring Summer Fall Brazos Santiago Pass n = 6 7 16 11 Ulva fasciata 14 4 13 9 Soliera tenera 0 4 1 16 Hypnea musciformis 18 32 0 0 Family Ceramiaceae 4 4 23 11 Bryocladia sp. 14 12 5 16 miscellaneous red algae 18 16 43 13 South Bay/Mexiquita Flats n = 8 9 4 15 Syringodium filiforme 14 8 0 12 Halodule wrightii 8 46 73 65 miscellaneous grass 48 0 2 3 miscellaneous red algae 18 25 0 17 beyond 10 m of the jetties where primary habitat was either: 1) small boulders typically devoid of algae or 2) a barren, silty-mud bottom. Transects completed at jetty stations yielded 34 species of algae belonging to three divisions (Table 8), Chlorophycophyta (6 species), Phaeophycophyta (7 species), and Rhodophycophyta (21 species). Species exhibiting the greatest frequency of occurrence ( 5 stations) were Ulva fasciata, Ceramium byssoideum, Padina vickersiae, Bryocladia cuspidata, Bryocladia thyrsigera, Sargassum fluitans, Soliera tenera, Rhodymenia pseudopalmata, and Spyridia aculeata. Layering of algae species was fairly uniform at all zones, with green alga in the intertidal zone and brown and red algae interspersed below that to a depth of up to 1 m. Of special note was the presence of large stands of the benthically attached Sargassum filipendula along the South Jetty in zones K and L and the North Jetty from zones S to V. No seasonal data are available for algae growth at the jetties, but there was an apparent increase in the width of the green algae layer during winter. An enlarged intertidal zone due to increased wave action may account for the increase in these hardier species.

28 Table 7. Average monthly air and water temperature and salinity in the South Padre Island study area during April 1991 - March 1993. Detailed monthly hydrological data are presented in Appendix Table 3. Month Air Temp (C) Water Temp (C) Salinity (ppt) surface bottom surface bottom Apr-91 23.97 22.00 n/a 31.00 n/a May-91 26.74 26.22 n/a 29.56 n/a Jun-91 28.69 26.74 28.40 35.43 34.96 Jul-91 27.74 27.19 27.55 36.01 35.12 Aug-91 28.18 28.47 28.14 35.64 35.59 Sep-91 28.43 29.12 n/a 35.32 n/a Oct-91 26.68 26.81 n/a 33.69 n/a Nov-91 20.91 20.67 n/a 34.60 n/a Dec-91 26.39 26.45 27.84 35.05 35.35 Apr-92 25.83 23.81 23.64 29.58 29.59 May-92 26.93 24.66 24.63 33.61 33.30 Jun-92 30.50 27.58 27.27 31.62 31.89 Jul-92 30.65 26.58 26.43 32.92 33.10 Aug-92 30.08 27.76 27.86 33.43 33.00 Sep-92 31.00 28.68 28.57 33.06 32.99 Oct-92 26.94 25.94 25.79 36.13 35.86 Nov-92 23.28 22.60 22.43 38.33 38.29 Dec-92 16.67 17.37 17.06 34.41 34.46 Jan-93 17.63 15.66 15.63 32.46 32.62 Feb-93 19.24 18.13 17.94 32.28 31.74 Mar-93 20.89 19.30 19.23 32.20 32.13 n/a - not available Other common epiphytes included sponges of the Class Demospongeae which occurred at 3 of 9 characterization stations (Table 8). The gorgonian Leptogorgia virgulata and two bryozoan species were common in the second and third biological zones at several stations. Small gastropods, copepods, and the crab Acanthonyx petiverii often were found amongst algae mats in the first biological zone. Xanthid crabs were observed at two jetty zones. Epiphytic diversity, ranging from 11 to 15 taxa, was fairly uniform across all stations with the exception of those at zones A outside and E during 1991 (3 and 6 species, respectively) and P outside during 1992 (6 species).