The endangered green sea turtle (Chelonia mydas) is a tropical to subtropical circumglobal species that utilizes shallow coastal and inshore seagrass meadows as developmental and internesting foraging grounds (National Marine Fisheries Service and US Fish and Wildlife Service [NMFS and USFWS] 1991, 2007; Hirth 1997; Musick and Limpus 1997; International Union for the Conservation of Nature and Natural Resources [IUCN] 2004). Atlantic green turtles occur in US continental waters from Texas to Massachusetts, and Texas' Laguna Madre supported a commercial green turtle fishery harvesting over 230,000 kg of meat/hr (Doughty 1984) during the mid-1800s (Hildebrand 1982). By 1900, overexploitation of this resource eliminated the fishery and greatly diminished the number of green turtles in Texas waters (Hildebrand 1982). A lack of monitoring and management data left a void in our understanding of the population status of remaining constituents until recently.

Although most green turtle research and management has focused on nesting populations, sea turtles spend over 90% of their life at sea and, as such, in-water assessments are critical to a better understanding of their biology, ecology, and population status. In-water surveys must be conducted throughout this species' range to 1) provide data prerequisite to revising outdated recovery plans; 2) define distinct population segments for management purposes, similar to that proposed for the Hawaiian green turtle population; and 3) identify threats to survival and critical habitats (Chaloupka and Musick 1997; Epperly 2000; Bjorndal et al. 2005; NMFS and USFWS 2007). This is particularly true for the green sea turtle, which migrates hundreds to thousands of kilometers across international waters to reach nursery areas, foraging grounds, and/or nesting beaches (Hirth 1997; Musick and Limpus 1997). To this end, the Sea Turtle and Fisheries Ecology Research Lab (STFERL) at Texas A&M University at Galveston has conducted long-term entanglement netting surveys in coastal waters of the northwestern Gulf of Mexico to document green turtle occurrence and habitat use patterns.

In-water surveys by the STFERL and others (Coyne 1994; Renaud et al. 1995; Shaver 1994; Arms 1996; Landry and Costa 1999) have identified the lower Texas coast as developmental grounds for immature conspecifics. Postpelagic green turtles, approximately 20–40 cm straight carapace length (SCL), initially recruit to algae-laden jetty habitat at tidal passes then transition to adjacent seagrass beds at > 30 cm SCL (Coyne 1994). These individuals also form the only reported sea turtle assemblage to overwinter in nearshore waters of the northwestern Gulf (Arms 1996). Although important, these findings were limited in geographic coverage to only a small portion of Texas inshore waters, and were conducted over a decade ago. This necessitated the current study on green turtle dynamics to cover a greater range of the Texas coast and to update the overall status of green turtles in this region.

Research reported herein was designed to 1) characterize current recruitment, size structure, and relative abundance of postpelagic, neritic green turtles at algae-laden, jettied-pass, and seagrass habitats; 2) assess behavioral patterns of any ontogeny-related dispersal of older juveniles to adjacent inshore waters; and 3) use historical green turtle data from the northwestern Gulf to provide a long-term, comprehensive assessment of green turtle population status needed for proper management and continued recovery.

METHODS

Study Areas

In-water assessments were conducted in lower, middle, and upper Texas bay systems (Fig. 1). The southernmost study area was the lower Laguna Madre (LLM) and included netting sites adjacent to Port Isabel, Laguna Atascosa National Wildlife Refuge, and Port Mansfield, all exhibiting extensive meadows of shoal (Halodule wrightii), manatee (Syringodium filiforme), and turtle grasses (Thalassia testudinum). The LLM was sampled with various seasonal frequencies during 1991–1994, 2002–2003, and 2006–2010 to provide the longest-term dataset among respective Texas bay systems. Middle coast habitats of the Aransas Bay complex south of Corpus Christi sampled during 2006–2007 included turtle grass in Corpus Christi Bay and areas dominated by shoal grass in Redfish and Aransas bays. The Lavaca-Matagorda Bay system, particularly lower reaches adjacent to Port O'Connor exhibiting patchy shoal grass distribution, was the upper coast study area sampled during 1996, 2001–2002, and 2006–2007. Seagrass coverage (km²) in these study areas is taken from Pulich and Calnan (1999) and is displayed in Table 1.

View Figure

• Download as Powerpoint
• Download Figure

Table 1. Estimated seagrass coverage (as of 1994) within the 3 Texas bay systems sampled during this study. Source: Pulich and Calnan 1999 (Texas Parks and Wildlife Department).

View Fullscreen

Turtle Capture and Visual Surveys

In-water capture was accomplished during daytime sets of 91.4-m-long entanglement nets that were 2.7 or 3.7 m deep with 12.7–17.7-cm bar mesh of #9 twisted nylon. Water depth and current dictated net type used at a particular station, but with all capture effort restricted to depths < 3 m. Netting effort at all stations consisted of 2–4 nets set in tandem or perpendicular to one another for 6–12 hrs per day. Net checks occurred every 20 min or more frequently as splashes or other signs of potential capture dictated. Pinger devices (manufactured by Fumunda Marine, Queensland, Australia) emitting high-frequency sounds were attached to nets to reduce incidental capture of bottlenose dolphins (Tursiops truncatus). Although 3 d of sampling effort was targeted for each study area during respective visits, weather conditions and equipment problems at times prevented netting in a particular area or given month. Additionally, lack of or insufficient funding during some periods prohibited sampling and resulted in gaps to the netting database. This, in turn, resulted in unequal and/or missing data that negated monthly and/or seasonal comparisons. Nevertheless, enough monthly sampling was conducted during historical netting surveys to allow for annual comparisons of catch data.

All green turtles were measured for SCL (cm) and examined for flipper tags/scars and passive integrated transponder (PIT) tags. Untagged individuals received an Inconel-style 681 tag (National Brand and Tag Co., Newport, KY), issued by the National Marine Fisheries Service Southeast Fisheries Science Center–Miami, affixed to the trailing edge of each front flipper and a PIT tag embedded under the dorsal surface of the right front flipper prior to release at their capture location.

Sea turtle use of Texas' jettied and tidal pass habitats was examined at 2 locations in 2009: 1) Brazos-Santiago Pass (BSP), at the southernmost end of South Padre Island; and 2) Mansfield Cut (MC), just east of Port Mansfield (Fig. 2). These passes provide major sources of ingress and egress by sea turtles into and out of the LLM. Jetties protecting these passes consist of granite boulders located north and south of each navigation channel, yet they differ in overall length (BSP jetties: ∼ 1500 m; MC jetties: ∼ 900 m). Jetties at each location were subdivided into observation zones, with the BSP jetties consisting of 25 100-m-long zones (labeled A–Y), and the MC jetties consisting of 27 50-m-long zones (labeled A–AA). Zones A–E and M–U at BSP and A–C and Q–X at the MC extended beyond the shoreline and allowed for visual surveys on either the beach or channel side of the north or south jetty. Visual surveys, following protocol utilized by Coyne (1994), were employed to assess sea turtle occurrence and behavior at the aforementioned jettied tidal passes during 2009.

View Figure

• Download as Powerpoint
• Download Figure

Visual surveys involved 1 observer monitoring sea turtle presence and behavior across 50 m of jetty length for 1 hr at random times of the day (2 observers per zone at BSP; 1 observer per zone at MC). Observations were conducted at 4 randomly selected jetty zones each sample day. Zone location and sea state dictated the side of the jetty on which observations were conducted, with the calmer side typically selected. The following information was recorded for each turtle sighting: 1) time of sighting; 2) species; 3) estimated size class (i.e., SCL  =  20–24.9 cm, 25–29.9 cm, 30–34.9 cm, etc.); 4) behavior; 5) distance from jetty (m); 6) estimated duration at surface; and 7) distinguishing markings or characteristics (i.e., barnacle patterns) to allow recognition of specific individuals sighted more than once. Turtle behaviors were categorized as “breathing”, “swimming”, “resting”, and “feeding”, with no behavior considered mutually exclusive (i.e., multiple behaviors could occur simultaneously, such as swimming and breathing or resting and breathing, etc.). Frequency of turtle sightings was expressed as number of turtles sighted · hr⁻¹ · 50 m⁻¹ of jetty length.

Attempts to capture jetty turtles with cast nets were made concurrent to or immediately following visual surveys. This capture method was most effective when turtles surfaced frequently, spent extended time (> 30 sec) at the surface, and were within 5 m of the jetty. Turtles captured at the jetties, besides being processed in the same manner described for those taken in entanglement nets, received a painted number (using nontoxic white window paint) on their carapace for possible resighting identification during subsequent visual surveys.

Data Analyses

Relative abundance of green turtles captured during entanglement netting surveys (expressed as catch per unit effort [CPUE]) was calculated as number of turtles per kilometer-hour of netting effort. These CPUE values were log-transformed to approximate a normal distribution for statistical analyses. Gaps in the monthly netting database from the LLM created disparate sample sizes among different years, and thus, prohibited the use of analysis of variance (ANOVA) to detect statistical differences in green turtle relative abundance across individual years. However, catch data from several years were grouped in 2 blocks (1991–1994 vs. 2002–2010) and a t-test was performed to evaluate differences in comparable monthly CPUE (April–August) between these time periods. Additionally, linear and nonlinear regression analyses were run to examine the overall trend in annual CPUE across all years sampled in the LLM. Differences in mean daily CPUE between the 3 estuarine study areas, which were all sampled only during 2006–2007, were examined via ANOVA to assess the relative abundance and geographic distribution of this species along the Texas coast. SCL within the LLM and across study areas (during 2006–2007 only) was also log-transformed to normalize the data and compared via ANOVA and regression analyses. Annual mean growth rate (cm/yr) for green turtles in the LLM was calculated via mark-and-recapture methods as (SCL at recapture − SCL at initial capture)/recapture interval (Seminoff et al. 2002). Statistical analyses were conducted with α  =  0.05, using SPSS software.

RESULTS

Relative Abundance and Distribution

A total of 273 green turtles were captured via entanglement netting across all locations during the 11-yr study. Relative abundance and distribution of green turtles across the 3 representative Texas estuaries sampled during comparable years (2006–2007) are shown in Fig. 3. Mean daily CPUE for each of these study areas during 2006–2007 ranged from a low of 0.07 turtles/km-hr in Lavaca-Matagorda Bay to a high of 1.5 turtles/km-hr in the LLM, which after log transformation (log[x + 1]), was significantly higher than that observed for the other 2 bay systems (F_2,30  =  6.63; p  =  0.004). This trend also coincides with seagrass areal coverage along the Texas coast (Table 1). A significantly higher CPUE was observed in the lower reaches of the LLM (Mexiquita Flats: 2.8 turtles/km-hr and South Bay: 2.9 turtles/km-hr) as compared with that of upper reaches adjacent to Port Mansfield (t₅  =  3.09; p  =  0.03).

View Figure

• Download as Powerpoint
• Download Figure

Historical Trends in CPUE

Yearly CPUE statistics from the LLM revealed a significant exponential increase from 1991 to 2010 (r²  =  0.72, p  =  0.001), despite a drop in CPUE during 2007 (Fig. 4). Log-transformed monthly CPUE during April to August of 1991–1994 vs. 2002–2010 also exhibited a significant increasing trend, with the overall mean for 2002–2010 approximately 10 times greater than that for 1991–1994 (t₂₂  =  −4.395; p < 0.001). There were no discernible abundance trends in Lavaca-Matagorda Bay across the 5 yrs sampled (1996, 2001–2002, and 2006–2007). Yearly CPUE in this bay system remained relatively low and did not exceed 0.12 turtles/km-hr.

View Figure

• Download as Powerpoint
• Download Figure

The 247 green turtles initially tagged by STFERL in the LLM during this study (1991–2010) yielded 24 recaptures and a recapture rate of 10%. One recapture, initially captured and tagged in 1993 (SCL  =  59.9 cm), was recorded nesting in Cuba in 2002 (no measurements were taken; Cooperative Marine Turtle Tagging Program, unpubl. data, 2003).

Size Composition and Growth Rate of Netted Turtles

The log-transformed SCL of green turtles captured during 2006–2007 was statistically similar across the 3 study areas (F_2,51  =  1.89; p  =  0.162). Turtles netted in the LLM ranged in size from 31.4 to 68.6 cm and averaged 42.2 cm (n  =  38), both statistics similar to those for conspecifics in the Aransas Bay complex (range  =  29.4–71.2 cm; mean  =  45.1 cm; n  =  13). The higher mean SCL (51.2 cm) recorded for green turtles in Lavaca-Matagorda Bay was based on measurements of 3 individuals and may not be representative of constituent size composition.

Historical trends in green turtle size composition in the LLM were assessed via length-frequency histograms (Fig. 5a–c), ANOVA, and a plot of annual mean SCL over all years sampled in lower reaches (Fig. 6). Percentage of contribution of individuals within the 30–39.9-, 40–49.9-, and 50–59.9-cm size classes during 1991–1994 was similar, with 40–49.9-cm individuals exhibiting a slightly higher proportion (Fig. 5a). Contribution by 40–49.9-cm individuals became more pronounced during 2002–2003 (Fig. 5b). A shift toward larger contributions from smaller individuals was observed during 2006–2010, such that a majority of constituents belonged to the 30–39.9- and 40–49.9-cm size classes (Fig. 5c). This shift also produced a decreasing trend in annual mean SCL (Fig. 6) over the long-term study period (r²  =  0.65; p  =  0.003). Influx of smaller individuals, specifically during 2008, resulted in log-transformed SCL values in the LLM differing significantly among years (F_10,231  =  4.243; p < 0.001).

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Mean growth rate for 24 green turtles recaptured in the LLM across sampling years was 5.7 ± 0.8 cm/yr SE (initial SCL range  =  30.8–61.1 cm). When grouped according to 10-cm initial size class intervals, estimated growth rates were not significantly different from one another (30–39.9 cm: 4.5 ± 0.7 cm/yr SE; 40–49.9 cm: 5.7 ± 1.0 cm/yr SE; 50–59.9 cm: 6.5 ± 2.2 cm/yr SE; F_2,22  =  0.521; p  =  0.60).

Jetty Surveys

A total of 704 turtle sightings were documented during 64 hrs of observation across LLM jetty locations and months in 2009. Eighteen of the 25 jetty zones established at BSP and 14 of 26 zones on the MC jetties were monitored at least once. The channel side of jetties was observed almost 5 times more often than that of the beach side (53 vs. 11 hrs, respectively) due to rough sea state conditions on the beach sides. Green turtles constituted 97% of all sightings, followed by unidentified sightings (2%) and those of loggerheads (Caretta caretta; 1%). No statistical difference was detected in frequency of sea turtle sightings (including all species) between months and/or jetty locations (F_2,63  =  1.323; p  =  0.274); however, Port Mansfield yielded a slightly higher mean frequency of sightings whereas a decreasing trend in sightings was observed from April to October/November (Fig. 7). The majority (35%) of turtles sighted were estimated to be 20–24.9 cm SCL, followed by 25–29.9 cm SCL (24%). Frequently observed behaviors of jetty turtles were breathing (65% of all observed behaviors) and swimming at the surface (26% of observed behaviors), whereas resting and feeding were both less commonly reported (5% and 4%, respectively). On 2 occasions, once at each jetty location, a green turtle was captured by a cast net. Capture of a 28.8-cm-SCL green turtle at BSP occurred on the channel side of the South Jetty, while a 29.7-cm-SCL individual was cast-netted adjacent to the MC South Jetty.

View Figure

• Download as Powerpoint
• Download Figure

DISCUSSION

Green turtles utilizing seagrass habitats in Texas inshore waters are primarily juveniles to subadults and are distributed in lower coast estuaries, with the greatest abundance found in the southernmost portion of the LLM. Size composition and distribution of this assemblage mirror those of previous assessments in the northwestern Gulf (Coyne 1994; Shaver 1994; Arms 1996; Landry and Costa 1999) and also are reinforced by data from multiple cold-stunning events along the Texas coast in 2007, 2010, and 2011. The January 2007 cold-stunning event involved 141 green turtles (116 alive, 25 dead; D. Shaver, unpubl. data, 2007), most of which stranded in the LLM (n  =  134; D. Shaver, unpubl. data, 2007) and exhibited a size composition (range  =  22.4–74.6 cm; mean  = 45.1 cm) similar to that of conspecifics captured during entanglement operations reported herein. A comparison of these data to published growth curves for Atlantic green turtles in the southeastern United States suggests the majority of green turtles in the Mexiquita Flats area of the LLM are 6–12 yrs old (Zug and Glor 1998; Goshe et al. 2010). Growth rates of green turtles recaptured in the LLM also coincide with estimates for similarly sized conspecifics reported in other studies from Texas (Coyne 1994: mean rate  =  5.25 cm/yr; Shaver 1994: mean rates  =  6.6–9.0 cm/yr), Florida (McMichael et al. 2008: mean rate  =  4.5 cm/yr), and the Caribbean (Bjorndal and Bolton 1988: mean rates  =  4.0–9.0 cm/yr; Boulon and Frazer 1990: 3.5–6.9 cm/yr). The predominance of juvenile-stage green turtles and absence of adult conspecifics in Texas' lower coast estuaries suggest this region serves as developmental benthic foraging habitat for C. mydas, comparable to that in other western Atlantic sites, including Bermuda; Secretary and Zapatilla, Panama; Cedar Key and Indian River Lagoon, Florida; the US Virgin Islands; and the Bahamas (Meylan et al. 2011). This concept of an “immature-dominated, benthic developmental stage” located in habitats often geographically distinct from adult foraging areas, and a regular part of the life cycle of most cheloniid sea turtles, was examined and supported by Meylan et al. (2011). A developmental benthic foraging area can be characterized by multiple traits including 1) benthic feeding on seagrasses, algae, or macroinvertebrates; 2) occupation exclusively or almost exclusively by postpelagic immature turtles; 3) residency and site fidelity; 4) maturation occuring elsewhere; 5) relatively high genetic diversity; and 6) long developmental migrations.

The distribution pattern of Texas' green turtle assemblage suggested by this study and cold-stunning events closely follows that of seagrass coverage and climate patterns along the Texas coast. Texas' most extensive seagrass beds, composed of 5 species, are found in the upper and lower Laguna Madre (751 km²) where they provide vast foraging opportunities to herbivorous green turtles. This habitat and subtropical temperature regime enhance the likelihood of year-round occurrence of green turtles in the LLM and facilitate the overwintering behavior reported by Arms (1996).

Although seagrass distribution and density appear important to green turtle occurrence along the Texas coast, impact of localized differences and changes in seagrass composition and coverage on foraging quality and use of constituent habitats by conspecifics is unknown. Declines in seagrass coverage in the LLM during the mid-1960s to 1998, due mostly to maintenance dredging activities of the Gulf Intracoastal Waterway, have coincided with an approximately 250% increase in bare, unvegetated areas (Quammen and Onuf 1993; Onuf 2007). In addition, species composition has changed from a shoal grass assemblage (89% of seagrass meadows in 1965 to only 48% in 1998) to one dominated by manatee and turtle grasses (both increasing over 200%; Onuf 2007). Although green turtles in Florida and the Caribbean forage primarily on manatee grass and turtle grass, respectively (Bjorndal 1997), Coyne (1994) concluded that LLM green turtles foraging in the Mexiquita Flats/South Bay area had a higher occurrence of shoal grass in stomach samples, despite this being the least abundant seagrass in the area. Coyne's (1994) findings suggest a preference for shoal seagrass in the LLM and raise concern as to quality of future green turtle foraging in the region if shoal grass coverage continues to decline. Continued characterization of green turtle forging ecology in the LLM is needed, via additional analyses of stomach contents and/or stable isotopes, to better understand possible feeding preferences among conspecifics in this area, nutritional content/benefit of constituent seagrass species, and impact of changes in seagrass composition on the green turtle population health.

Green turtle abundance in the LLM has increased exponentially since 1991 despite declining seagrass coverage over the past 30 yrs. CPUE statistics generated during 2002–2010 were approximately 10 times greater than those for 1991–1994. Recent record-breaking cold-stunning events in 2010 (∼ 600 green turtles stranded) and 2011 (∼ 1200 green turtles stranded) also suggest increases in green turtle population size along the Texas coast (D. Shaver, unpubl. data, 2012). Additionally, sighting frequency of postpelagic green turtles at the BSP jetties during 2009 was roughly 9 times (mean over all months  =  9.5 turtles/hr) greater than that reported by Coyne (1994) for comparable months during 1992–1993 (∼ 0.5–2.5 turtles/hr). The exponential increase in green turtle CPUE in the LLM could be due to shifts in habitat use resulting from habitat degradation elsewhere. However, increased frequency of smaller individuals (30–40 cm SCL) in studywide catches from the LLM since 2006 suggests these abundance trends are due to enhanced recruitment resulting from elevated nesting productivity at beaches in Mexico, Florida, the Caribbean, and the western Atlantic (Troëng and Rankin 2005; Meylan et al. 2006; NMFS and USFWS 2007). Although 1 green turtle initially tagged in the LLM in 1993 was found nesting in Cuba in 2002, the natal origin (or origins) of green turtles in the northwestern Gulf of Mexico is largely unknown and mandates genetic analyses to identify nesting rookeries contributing recruits to this region (Bass et al. 2006). Green turtle nests are found along the Texas coast and have increased in number in recent years, but the extremely low numbers (8 nests in 2012; D. Shaver, unpubl. data, 2013) most likely do not supply a significant number of juveniles to benthic foraging habitats in the northwestern Gulf. However, based on nesting abundances and geographic proximity, rookeries in Florida and Mexico are the most likely contributors to the juvenile green turtle assemblage found in Texas waters. Both of these areas have exhibited increases in nesting activity, with an average of 5055 green turtle nests/yr reported in Florida (based on observations from 2001–2005) and an average of 1500 nests/yr along the Yucatan Peninsula, Mexico (NMFS and USFWS 2007).

Although relatively few other studies have examined in-water green turtle abundance trends in the Atlantic basin, some comparisons to trends at other developmental foraging areas can be made. The increasing population trends in LLM's green turtle assemblage mirror those in Florida's Indian River Lagoon system, where increasing abundance of juvenile to subadult green turtles has coincided with growth in the Florida nesting populations from which these conspecifics are derived (Ehrhart et al. 2007). Epperly et al. (2007) examined trends in sea turtle catch rates from pound nets in North Carolina waters during 1995–1997 and 2001–2003 and found that green turtles were the second most abundant species captured, but there were no discernible trends in green turtle CPUE over the years, despite a 10% increase in related nesting assemblages. However, this result may have been due to insufficient statistical power to detect a trend. Bjorndal et al. (2005) used mark-and-recapture methods to assess trends in green turtle abundance at juvenile foraging habitats in the Bahamas over a 24-yr period and also found no significant trend overall, despite increases in related Costa Rican nesting assemblages. However, some specific foraging areas within the Bahamas had shown periods of significant increase, decrease, and stability.

Catch statistics generated by STFERL's entanglement netting in seagrass habitats have also been variable in recent years. The decline in CPUE observed in 2007 may have been related to effects from the cold-stunning event occurring earlier that year involving 134 green turtles in the LLM (D. Shaver, unpubl. data, 2007). However, there was no similar decline in LLM CPUE following a more severe January 2010 cold-stunning event resulting in approximately 600 green turtles stranded statewide (D. Shaver, unpubl. data, 2010). Conversely, a studywide CPUE peak in August 2008 was dominated by relatively small green turtles (mean  =  36.6 ± 0.65 cm SCL; range  =  29.1–44.3 cm SCL), thus suggesting an influx of postpelagic recruits into seagrass habitats. This prevalence of small recruits may have been a result of Hurricane Dolly, whose landfall on South Padre Island 2 wk prior to the August 2008 sampling possibly pushed turtles normally resident at the jetties into the seagrass beds adjacent to BSP. With seemingly more recruits settling in and utilizing jettied passes along the lower Texas coast, intraspecific competition at this habitat may force some individuals to recruit to seagrass beds at a smaller size.

Recruitment dynamics, in addition to explaining population increases in LLM's green turtle assemblage, shed light on the role the lagoon's jettied pass and seagrass meadows play in this species' life history. The aforementioned habitats support at least 2 ontogenetic shifts characteristic of this species. The smaller size of green turtles observed at jettied habitats, as compared to that of conspecifics captured in entanglement nets adjacent to seagrass beds, suggests that youngest constituents exhibit an ontogenetic shift wherein they initially recruit from the pelagic zone to jetty habitats at approximately 20–25 cm SCL. As part of this shift, young green turtles transition from an omnivorous diet characteristic of all pelagic life stages to herbivory, where they depend on algal communities of hard structures such as the BSP jetties (Coyne 1994; Howell 2012). Despite no significant difference in the frequency of turtles sighted at the 2 jetty locations, the greater abundance of green turtles captured at netting sites in the southernmost portion of the LLM imply that BSP may be a more important source of ingress for these postpelagic recruits than is MC. This “first stop” in their developmental migration may last for up to 6 yrs (Musick and Limpus 1997; Zug and Glor 1998). A second ontogenetic shift commences when these jetty-dependent turtles reach 30–35 cm SCL, at which size they transition to seagrass beds of the LLM. While in coastal habitats, green turtles exhibit site fidelity to specific areas or home ranges (Bresette et al. 1998; Makowski et al. 2006) such as specific jetties or seagrass beds. Renaud et al. (1995) suggested abundance of food (i.e., algae) accounted for small green turtles exhibiting high site fidelity and small core areas at the BSP jetties. Arms (1996) reported juvenile green turtles (mean SCL  =  48.6 cm) overwintered in and displayed strong site fidelity to Mexiquita Flats seagrass habitats that were used as foraging and development grounds.

Localized environmental conditions, such as water temperature and salinity, may also influence the relative abundance, distribution, and habitat use patterns of green turtles in seagrass habitats along the Texas coast. Yet, these specific relationships were not examined in this study because monthly netting surveys were inconsistent over the years and primarily restricted to warmer periods (April–August), with little fluctuation in environmental parameters within or between years. For example, water temperatures during recent years of netting (2006–2010) were, on average, between 24°C and 31°C. These temperatures did not fall below the 20°C threshold known to induce changes in sea turtle activity patterns that include decreased feeding and hibernation as well as movements to a more favorable temperature (Milton and Lutz 2003). However, satellite telemetry of 5 individuals (40–70 cm SCL) captured during this study and tracked over winter months indicate green turtles utilizing the LLM are capable of seasonal migrations south into Mexican waters to escape unfavorable conditions brought on by relatively rapid and prolonged declines in water temperature, with a subsequent return to Texas waters and seagrass habitats as temperatures rise during spring (Landry and Metz 2009).

The lower Texas coast (particularly the LLM) serves as an important developmental foraging habitat for green turtles in the northwestern Gulf of Mexico and recent increases in relative abundance are a positive result of conservation measures for this species. However, additional netting sites and year-round sampling are needed to fully assess distribution, habitat use (i.e., are green turtles primarily settling in seagrass beds adjacent to tidal passes or are they utilizing more remote areas of the LLM?), and seasonal trends in Texas' green turtle assemblage. Although the increasing CPUE recorded for green turtles from the Mexiquita Flats area of the LLM since 1991 is a promising sign for this species' recovery, increased numbers of individuals will likely lead to increased or continued encounters with various anthropogenic threats to survival, such as dredging operations, boat strikes, and incidental capture in fisheries. Currently, green turtles are the species most commonly relocated and/or taken during US Army Corps of Engineers dredging operations along the lower Texas coast (R. Hauch., pers. comm., November 2007). In addition to these impacts, seagrass habitat loss and degradation may lead to decreased foraging opportunities for juvenile green turtles in the LLM, thus affecting growth and body condition, age at maturity/recruitment to nesting populations, and overall survival. Furthermore, the recent discovery of fibropapilloma virus in green turtles from the LLM (Tristan et al. 2010) merits further examination to ascertain factors contributing to the onset of this disease and its impact on continued population growth and recovery of green turtles in the northwestern Gulf. These issues must be addressed and Texas' green turtles taken into consideration when evaluating conservation measures and revising recovery plans for the green turtle in US waters. Continued in-water monitoring of green turtles in the northwestern Gulf as well as identification and protection of critical habitat are key elements to this species' on-going recovery.