Habitat Use and Basking Behavior of a Freshwater Turtle Community Along an Urban Gradient
Abstract
Urbanization of riparian corridors may alter or eliminate suitable freshwater turtle basking habitat due to fragmentation of shoreline vegetation, reduction of basking sites, or frequent human disturbance. We used 3 indices of shoreline urbanization at 2 spatial scales to assess the relationship between shoreline urbanization and basking turtle behavior. Indices included local-scale Shoreline Modification and Disturbance Frequency and broad-scale Building Density. The community of basking turtles included the Texas river cooter, Pseudemys texana (Baur); red-eared slider, Trachemys scripta elegans (Wied); Mississippi map turtle, Graptemys pseudogeographica kohnii (Baur); midland smooth softshell, Apalone mutica mutica (LeSueur); pallid spiny softshell, Apalone spinifera pallida (Webb); and a turtle from the family Kinosternidae that could not be identified to species during basking surveys. At the local scale, abundances of basking turtles were greatest in areas of high Shoreline Modification, characterized by a substantial reduction in woody shoreline vegetation. Disturbance Frequency of human intrusion limited turtle basking in areas with daily disturbance. At the broad scale, most turtles basked adjacent to shorelines with nearby buildings. All species of the turtle community basked in urban environments, but their tolerance of urbanization varied.
Urbanization of riparian corridors and associated anthropogenic activity can negatively affect freshwater turtle abundance, alter community structure, and modify key behaviors through habitat alteration, degradation, or loss (Garber and Burger 1995; Donner Wright et al. 1999; Gibbons et al. 2000; Conner et al. 2005). Riparian woodlands are frequently fragmented in urban landscapes (Décamps et al. 1988; Maisonneuve and Rioux 2001), which further degrades associated aquatic habitats. Freshwater turtles inhabit a variety of microhabitats within these riparian corridors (Moll and Moll 2004).
Freshwater turtles use the entire riparian corridor, so shoreline stressors may be as influential as aquatic ones. Highly urbanized landscapes can disrupt turtle behaviors on terrestrial and aquatic fronts by removing vegetation, increasing human intrusion, increasing pollutants, and altering river geomorphology via damming and channelization (Bodie 2001; Tucker et al. 2001; Moore and Seigel 2006). Urbanization may reduce shoreline vegetation and deadwood inputs into the river and thereby constrain turtle basking (Spinks et al. 2003). Deficient habitat fragments turtle populations in areas as large as a drainage basin or as small as an isolated tributary (Shively and Jackson 1985), and intolerant species may be reduced or eliminated by urbanization.
Life-history characteristics of freshwater turtles make them especially susceptible to the effects of urbanization (Baldwin et al. 2004). High juvenile mortality and delayed sexual maturity (Ernst et al. 1994) exacerbate the effects of urban stressors such as habitat modification (Browne and Hecnar 2007), disturbance (Dodd and Dreslik 2008), and capture for commercial sale (Ceballos and Fitzgerald 2004). This synergy underscores the importance of assessing and monitoring the influence of urbanization on turtle community structure.
Characterizing urbanization is particularly challenging because anthropogenic stressors vary and affect systems at multiple scales. Categorizing a location as urban is often based on human population densities, land use classes, or percent impermeable surfaces (Snodgrass et al. 2008). Authors have previously characterized the effects of urbanization on turtles by analyzing the effects of pollutants (Neuman-Lee and Janzen 2011), identifying factors that increase road mortality (Cureton and Deaton 2012), quantifying nest predation by human-subsidized predators (Ner and Burke 2008), and monitoring demographic shifts (Eskew et al. 2010). To combine characterizations of the urban environment and the effects on the turtle community, we developed urbanization indices that were relevant to basking turtle behavior. The present study focuses on the effects of shoreline urbanization from multiple sources and at multiple scales on the structure of a freshwater turtle community in a central Texas river, their basking behavior, and the availability of potential basking sites. We addressed the following questions: 1) How does shoreline urbanization relate to the spatial distribution of basking turtles? 2) What are the scales and sources of urbanization that impact turtle basking abundances? 3) Is the response of freshwater turtles to shoreline urbanization species-specific?
METHODS
Indexing Urbanization
Research was conducted on 14 km of the Brazos River in McLennan County, Texas (Fig. 1). Shoreline development ranged from an intact, active riparian corridor of native vegetation to the complete absence of vegetation and the addition of impenetrable surfaces such as concrete. We independently characterized 28 km of opposing shorelines for the level of development with 3 distinct indices: 1) Shoreline Modification, 2) Disturbance Frequency, and 3) Building Density. These 3 urbanization indices encompassed 2 spatial scales: local and broad. Local-scale shoreline urbanization included activity immediately adjacent to a basking turtle, while broad-scale urbanization included activity along tens of meters of the shoreline adjacent to a basking turtle. The local-scale urbanization indices were Shoreline Modification and Disturbance Frequency and the broad-scale urbanization index was Building Density.
Citation: Chelonian Conservation and Biology 12, 2; 10.2744/CCB-0961.1
Shoreline Modification measures the degree of altered shoreline vegetation. Shoreline Modification was ranked as extreme (1), high (2), moderate (3), slight (4), or unmodified (5). Extremely modified shorelines had native woody vegetation removed with addition of ornamental trees and evidence of landscaping, as in a residential yard. Highly modified shorelines had patchy native vegetation with evidence of shoreline management. Moderately modified shorelines had reduced but intact native vegetation with evidence of management. Slightly modified shorelines had reduced but contiguous native vegetation with no evidence of shoreline management. Unmodified shorelines had extensive native vegetation with no evidence of shoreline management. The number of basking observations for each category was standardized per shoreline kilometer of that category.
Disturbance Frequency refers to the frequency of human presence on the shoreline. Disturbance levels were ranked as daily (1), weekly (2), monthly (3), semiannually (4), and undisturbed (5). Disturbance Frequency was assessed via evidence of recent human presence, such as structures, trails, litter, or vandalism. A city park exemplifies daily shoreline disturbance, while a rural fishing trail typifies monthly disturbance. The number of basking observations for each category was standardized per shoreline kilometer of that category. For Shoreline Modification and Disturbance Frequency, both shorelines were characterized from a boat for the entire length of the study site (14 km), 10 m landward from where the active river channel meets the land.
Analysis of the broad-scale index of urbanization, Building Density, was performed using ArcView GIS, with each location of a basking turtle projected onto a geo-referenced image of the study site. The number of buildings within 30 m of each incidence of basking on the adjacent shoreline was recorded. Building data were gathered from 1-m resolution Digital Orthophoto Quarter Quads (August 2004) from the National Agricultural Imagery Program (Fig. 1).
Survey of Potential Basking Sites
Potential basking sites in the study site were surveyed from a boat in August 2006, April 2007, and August 2007. Emergent stationary woody vegetation, alive and dead, with circumferences > 15 cm and angles to the water's surface ≤ 90 degrees were recorded as potential basking sites. Potential basking site abundances were standardized for the distance of shoreline characterized by each urbanization index.
Basking Surveys
Sixteen surveys of basking turtles were conducted by boat from August 2006 to September 2007 and were restricted to warm (> 20°C) sunny days with < 30% cloud cover. For each survey, potential basking sites on a single shoreline were monitored for basking turtles. To avoid interobserver variability, all surveys were performed by a single researcher (the lead author). Survey starting and end points were fixed and observations were made from a boat moving upstream while using binoculars (Eagle Optics 8 × 42). The boat was kept in the middle of the active river channel and maintained a slow velocity in order to allow time for identification of basking turtles. Sampling effort was consistent among surveys, averaging 5 person-hrs per survey. Location and species of all basking turtles were recorded. In cases where multiple turtles occupied the same basking site, abundances of each species were recorded in order to analyze intra- and interspecific interactions. For each basking occurrence the shoreline was characterized for the categories of Shoreline Modification, Disturbance Frequency, and Building Density. Basking turtle abundances were standardized by the distance of shoreline characterized for each urbanization index.
Trapping
Trapping occurred from 11 October 2004 to 6 December 2004, 14 October 2005 to 2 December 2005, and 13 November 2006 to 7 July 2007 and was not concurrent with the surveys for basking turtles. Trapping sites were randomly selected and traps were typically deployed for 2 wk. Recorded turtle abundances were standardized for differences in trapping effort. Basking traps were polyvinyl chloride frames (167 × 167 × 30.5 cm) with wire basking platforms attached to the outer edge of the frame and rising 20 cm above the water. A central mesh basket filled the interior of the frame and caught turtles leaving the basking platforms toward the interior of the frame. Trapped turtles were identified to species, sexed, and marked individually by notching the marginal scutes (Cagle 1939). Carapace length (midline curve) and width (maximum curve) as well as plastron length and width were measured. After processing, turtles were released at the point of capture.
Data Analysis
The relationship between basking frequency and categories of Shoreline Modification, Disturbance Frequency, and Building Density were explored with the SAS statistical package, specifically the GENMOD procedure that fits data to General Linear Models (GLM) as defined by Nelder and Wedderburn (1972; also see McCullagh and Nelder 1989; Dobson 2001). The assumed distribution of the response variable (basking abundance) was Poisson with a log-link function. We deemed analyses with a more traditional least-squares regression model to be inadequate because neither the response variable nor its log transformation was normally distributed. In addition, the independent predictor variables were categorical and not appropriate for parametric analysis. The GENMOD multivariate procedure with a Poisson distribution (sometimes referred to as a Poisson regression) provides strong power of assessment. The proposed model for each turtle species is first assessed with appropriate goodness-of-fit tests. This procedure 1) tests the goodness-of-fit of the multivariate model, although it does not provide a stepwise path; 2) accepts count data and categorical data to be mixed at all positions of the model; 3) assesses and corrects for overdispersion (increases in variance as the mean increases); 4) uses Deviance and Chi-square to test the overall goodness-of-fit for the entire model; and 5) assigns a probability statement for the significance of each predictor variable in the model. Use of a Poisson-based analysis in a multivariate model, along with assessment of each variable and its coefficient, also minimizes bias in the biological conclusions resulting from mild covariance of the predictor variables (McCullagh and Nelder 1989).
In order to characterize the intra- and interspecific interactions of an assemblage of turtles on a single basking site and to determine whether species preferentially basked together, a Chi-square goodness-of-fit test compared the proportions of species basking together with the overall proportions in the basking community.
RESULTS
Degree of Urbanization
All levels of Shoreline Modification and Disturbance Frequency were represented by > 1 km (Table 1). Extreme Modification and Unmodified were the most highly represented levels of modification. Daily Disturbance and Undisturbed were the most highly represented levels of disturbance. The shoreline was predominantly devoid of buildings (23.8 km) with 3.2 km of shoreline with 1 building/km, 0.8 km of shoreline with 2 buildings/km, and 0.2 km of shoreline with ≥ 3 buildings/km.
Potential Basking Sites
Potential basking sites (n = 7972) were most abundant on shorelines that were moderately (29%) or highly (28%) modified (Fig. 2a) and on shorelines that were undisturbed (40%) or disturbed semiannually (24%; Fig. 2b). Most potential basking sites (77%) occurred in areas with no buildings on the shoreline.
Citation: Chelonian Conservation and Biology 12, 2; 10.2744/CCB-0961.1
The number of available basking sites and its predictor variables were modeled with the GENMOD procedure described in the “Data Analysis” section of the “Methods”. Goodness-of-fit of the proposed multivariate model (response variable: number of available basking sites; predictor variables: Shoreline Modification, Disturbance Frequency, and Building Density) was significant (p < 0.05) for both Deviance (df = 3096, Dev. = 3487) and Chi-square (df = 3096, χ2 = 4361). The multivariate model fit well.
The significance of each of the predictor variables was also assessed individually. Chi-square analysis determines the significance of the variable. If the coefficient of the variable could be randomly chosen with no effect on the goodness-of-fit of the equation, then the variable was deemed insignificant. Disturbance Frequency was the only statistically significant predictor of basking-site abundance (χ2 = 65.15, p < 0.001; Table 2).
Basking Surveys
Observations of basking turtles (n = 1597) revealed a community consisting of the endemic Texas river cooter, Pseudemys texana (n = 650, 40.7%); the red-eared slider, Trachemys scripta elegans (n = 336, 21.0%); Mississippi map turtle, Graptemys pseudogeographica kohnii (n = 51, 3.2%); and a member of the Kinosternidae family (n = 7, 0.4%) that could not be identified to species during basking surveys. The midland smooth softshell (Apalone mutica mutica) and the pallid spiny softshell (Apalone spinifera pallida) were indistinguishable from a distance; therefore, observations were recorded collectively and represented 1.6% (n = 26) of the basking community. The remaining observations of the basking community (n = 527, 33.0%) were of turtles that retreated into the water before their identification. These unidentified turtles were included in analyses of combined turtle taxa but omitted from analyses of species-specific responses.
Basking-site occupancy ranged from a solitary individual to substantially larger densities. Single occupancy on a basking site was common for all turtle species (61.8%). In cases of multiple occupancy of a single basking site, 53% of P. texana basked with other individuals of the same species. Pseudemys texana basked with T. scripta 38% of the time, with G. pseudogeographica 7% of the time, and with Apalone spp. 2% of the time. Trachemys scripta basked with P. texana (61%) more often than with conspecifics (32%), rarely basked with G. pseudogeographica (6%) and the kinosternid (1%), and was never observed basking with Apalone spp. As was true for T. scripta, G. pseudogeographica basked more frequently with P. texana (61%) than with conspecifics (4%) or with T. scripta (35%). Graptemys pseudogeographica was never observed basking with Apalone spp. or the kinosternid. Apalone spp. exclusively basked with P. texana (83%) or with congeners (17%). Only one observation was made of the kinosternid basking with any other species, which was T. scripta. Pseudemys texana (χ2 = 7.44, df = 4, p = 0.11) and T. scripta (χ2 = 2.67, df = 3, p = 0.45) basked with other species on the same basking site in the same proportions as each species is found within the basking community. The rarest observations for multiple occupancy in the T. scripta analysis (basking with Apalone spp. or kinosternids) were combined to meet the criterion of 5 observations per category, but limited sample size made Chi-square analysis of multiple basking by G. pseudogeographica, A. mutica, A. spinifera, and the kinosternids inappropriate. There is no evidence that P. texana or T. scripta preferentially bask with any species in this community. Rather, the proportion of species sharing a basking site is comparable to the proportions observed basking alone.
Cumulatively, freshwater turtles basked most frequently (38%) in areas of high modification, with similar abundances in moderately (19%), extremely (17%), and slightly (17%) modified shorelines (Fig. 3a). Interestingly, turtles basked least frequently (9%) near unmodified shorelines. Basking by P. texana and T. scripta was most frequent near highly modified shorelines (Fig. 3a). Graptemys pseudogeographica basked in similar proportions adjacent to moderately (28%), slightly (22%), and highly (21%) modified shorelines and less prevalently near extremely (11%) and unmodified (18%) shorelines (Fig. 3a). Apalone mutica and A. spinifera basked most frequently next to highly modified shorelines (63%) and much less adjacent to slightly (16%), extremely (8%), moderately (8%), and unmodified shorelines (6%). The kinosternids basked almost exclusively (86%) adjacent to the most urbanized shorelines (Fig. 3a), but limited sample size should temper conclusions about the response of kinosternids to Shoreline Modification.
Citation: Chelonian Conservation and Biology 12, 2; 10.2744/CCB-0961.1
The responses of turtle basking to Disturbance Frequency (Fig. 3b) differed strikingly from the response to Shoreline Modification. Pseudemys texana basked most in areas of low disturbance (semiannually = 28%, undisturbed = 22%) and less in areas of weekly (19%), monthly (17%), and daily (15%) disturbance. Trachemys scripta basked in all ranges of Disturbance Frequency (Fig. 3b). As with Shoreline Modification, the response of G. pseudogeographica to Disturbance Frequency was markedly different from that of P. texana, with G. pseudogeographica basking adjacent to shorelines with higher levels of disturbance. Choice of disturbance frequencies by A. mutica and A. spinifera was bimodal. They most often chose areas with weekly (41%) and semiannual (37%) disturbance. As with the trend in Shoreline Modification, the kinosternids basked most frequently in areas of greatest disturbance (weekly = 47% and daily = 34%).
Turtle basking for all species was greatest near shorelines with buildings. Specifically, basking across all species was greatest (32%–60%) in areas with one building. Surprisingly, shorelines without buildings were the least conducive for turtle basking. Only 2%–26% of the basking observations were made adjacent to shorelines without buildings.
Basking frequency was modeled with the multivariate combination of urbanization indices of Shoreline Modification, Shoreline Disturbance, and Building Density for P. texana, T. scripta, G. pseudogeographica, Apalone spp., and the kinosternids. The goodness-of-fit for the model was significant for P. texana, T. scripta, and G. pseudogeographica (Table 2). The model with urbanization indices failed to predict basking abundances for Apalone spp. or the kinosternids (Table 2).
Trapping
Trapping at 47 randomly selected locations within the study site yielded 383 turtles. The community of turtles trapped was similar in proportions to the species observed basking. Trapped turtles included P. texana (66%), T. scripta (29%), G. pseudogeographica (3%), A. mutica (1%), and A. spinifera (1%).
Trapping success for all species was highest adjacent to highly modified shorelines (Fig. 4a), which was consistent with the basking observations. Only the 2 most abundant species in the community, P. texana and T. scripta, were trapped in every category of Shoreline Modification.
Citation: Chelonian Conservation and Biology 12, 2; 10.2744/CCB-0961.1
Disturbance Frequency affected trapping success (Fig. 4b). Pseudemys texana was trapped most frequently in areas of weekly (33%) to monthly (29%) disturbance; proportions were similar for G. pseudogeographica. Trachemys scripta was trapped most frequently in areas of monthly (43%) and daily (23%) disturbance.
Trapping success was greatest in areas with 1–2 buildings on the shoreline, accounting for 72% to 82% of all captures of P. texana, T. scripta, and G. pseudogeographica. Apalone mutica was captured exclusively in areas with no buildings. Apalone spinifera was never captured in areas with more than 1 building on the shoreline.
DISCUSSION
Basking Surveys
To a degree, shoreline urbanization facilitates turtle basking behavior. Selection of basking location relative to Shoreline Modification was similar for P. texana and T. scripta; they preferentially basked in areas of high Shoreline Modification. This is interesting to note because it would be an intuitive assumption that turtles avoid basking near greater levels of urbanization. Preference for highly modified shorelines may relate to greater solar radiation reaching basking sites near modified shorelines. The dense canopy of unmodified riparian vegetation typically extends over near-shore areas and filters direct sunlight from reaching many potential basking sites. Shoreline Modification, by definition, reduces the number of trees and shade, allowing for direct sunlight to reach many potential basking sites. Graptemys pseudogeographica was the exception to greater basking frequencies near highly modified shorelines. It basked most often in areas of moderate to low modification.
Pseudemys texana and T. scripta, the most abundant turtles in this community, were intolerant of frequent disturbance, while Apalone spp. and the kinosternids, the rarest turtles in this community, basked almost exclusively near moderate to high disturbance. Interestingly, while G. pseudogeographica was sensitive to shoreline modification, this species preferred more disturbed areas. Moore and Seigel (2006) found that another species of map turtle, Graptemys flavimaculata, tolerated disturbance by watercrafts. Although basking G. flavimaculata were dislodged from basking sites by passing watercrafts, reemergence was common.
Turtles responded more strongly to local-scale urbanization indices of Shoreline Modification and Disturbance Frequency than to the broad-scale measure of Building Density. Shoreline Modification significantly predicted basking presence for P. texana, T. scripta, and G. pseudogeographica. Disturbance Frequency predicted basking presence for P. texana and G. pseudogeographica, and Building Density did not predict the presence of any members of the basking community.
Previous research has shown substantial variation in the responses of freshwater turtles to urbanization. In an urban canal, Peterman and Ryan (2009) observed a robust community of emydid turtles; however, their use of basking sites was reduced when shoreline grass was mowed, because they were less sheltered from human disturbance in adjacent recreational areas. Foster et al. (2004) found less turtle species richness and reduced abundances in an urban setting when compared with an adjacent rural environment. In an urban stream, A. spinifera was monitored before and after pronounced habitat alteration; and, while abundances were only modestly reduced postmodification, the previously localized population increased long-distance movements (Plummer and Mills 2008). Phrynops geoffroanus had significantly higher leech loads in urban portions of the Uberabinha River, indicating poor population health (Brites and Rantin 2004). However, abundances of P. geoffroanus in the highly urbanized Ribeirão Preto Stream were remarkably high (Souza and Abe 2000). Even in areas protected from urbanization, freshwater turtle populations may be in decline. Browne and Hecnar (2007) documented the extirpation of the spotted turtle (Clemmys guttata) and reduced abundance of Blanding's turtles (Emydoidea blandingii) with skewed sex ratios and reduced juvenile recruitment, despite long-term habitat protection.
Intra- and interspecific interactions may influence assemblage structure of freshwater turtles on a single basking site. The number of turtles occupying a basking site varied from a single individual to abundances so great that turtles stacked on top of each other. In cases of multiple occupancy, species composition on a single basking site may range from a lone species to a more complex assemblage structure, consisting of several species. Multiple occupancy on basking sites provides an opportunity for social interaction (Lovich 1988, 1990; Lindeman 1999a). On the Brazos River, the majority of observed basking consisted of one solitary individual per basking site. Though observations of multiple occupants on a basking site were not uncommon, multiple occupancy was observed less frequently than solitary basking. Previous studies have documented aggressive interactions between basking turtles (Lovich 1988; Lindeman 1999b). Interference competition appeared to be at a minimum in this study, because most potential basking sites remained unused and multiple occupancy was observed for several species in the community, albeit at a lower frequency than solitary basking. Species basked alone most frequently and when basking together, they did so in frequencies comparable to the proportion at which they were found in the community; therefore, strong social interactions, either attraction or repulsion, were not supported. This suggests that selection of basking sites is not driven solely by social interactions and may be influenced by environmental conditions, such as changes in the shoreline due to urbanization.
Trapping
As with the basking data, abundances of trapped freshwater turtles were greatest in areas of high modification. This was especially applicable for A. mutica and A. spinifera, which showed a high degree of association with areas having reduced vegetation. An exception was G. pseudogeographica, which was trapped evenly across all levels of modification. All turtle species were trapped in greatest abundances near shorelines where disturbances were fairly infrequent. Apalone spinifera were trapped most frequently in undisturbed and extremely disturbed areas, which is probably an artifact of limited trapping success for this species.
Basking Sites
Potential basking sites were more abundant along shorelines with reduced riparian vegetation (i.e., moderately and highly modified shorelines contained more potential basking sites). Availability of potential basking sites may also be influenced by deposition and scouring within the active river channel, which may modify the intuitive relationship between shoreline vegetation and deadwood inputs into the river. Alternatively, sparse woody vegetation in a thinned forest may increase the weathering effects on the trees, causing greater tree falls and lost branches. Lindeman (1999a) found a statistically significant relationship between basking Graptemys spp. frequency and basking-site abundance, suggesting that ample basking substrates may increase habitat suitability.
Conclusions and Management Implications
The influence of urbanization on the basking ecology of freshwater turtles varies for each species, across urbanization indices, and at different spatial scales. The basking ecologies of P. texana and T. scripta were strikingly similar and often distinct from the remaining community. Freshwater turtles respond more strongly to local-scale measures of urbanization. Pseudemys texana and T. scripta tolerated the removal of shoreline vegetation, but P. texana preferred to bask in areas of low human disturbance. In contrast, G. pseudogeographica preferred shaded, unmodified shorelines but tolerated human disturbance, and the kinosternids basked near substantially modified and disturbed shorelines.
Potential basking-site abundance was greatest in areas of high and moderate modification and low disturbance. P. texana and T. scripta bask most often in these areas, so plentiful basking sites may facilitate greater abundances of these species.
The influence of urbanization on turtle abundances and basking is not entirely negative. Portions of the Brazos River are significantly urbanized, yet they support large populations of basking turtles. All species in this basking community tolerated some level of urbanization and 2 species basked exclusively in highly urbanized areas. Even the rarest species in the community was found basking in the most metropolitan portion of this study site.
To maintain a basking turtle community in an urban environment, riparian corridor management plans should focus on creating a mosaic of shoreline development with varying degrees of riparian vegetation and human presence. Removal of potential basking sites should be minimized (Bodie 2001). Shoreline development and removal of shoreline vegetation need not be eliminated entirely to support a large community of basking turtles.
Study site on the Brazos River in McLennan County, Texas, illustrating a rural to urban gradient. Example rural area and example urban area, as well as upstream and downstream boundaries of the study site, are indicated. Photograph courtesy of National Agricultural Imagery Program.
Densities of potential freshwater turtle basking sites per kilometer of shoreline of the Brazos River in McLennan County, Texas, for each of 5 levels of Shoreline Modification (panel a) and Disturbance Frequency (panel b).
Densities of basking freshwater turtle species per kilometer of shoreline of the Brazos River in McLennan County, Texas, for each of 5 levels of Shoreline Modification (panel a) and Disturbance Frequency (panel b). All = total turtles, TRC = Texas river cooters, RES = red-eared sliders, MM = Mississippi map turtles, SS = softshells, and K = kinosternids.
Trapping success of freshwater turtle species standardized by number of trapping days and quantity of shoreline of the Brazos River in McLennan County, Texas, for each of 5 levels of Shoreline Modification (panel a) and Disturbance Frequency (panel b). TRC = Texas river cooters, RES = red-eared sliders, MM = Mississippi map turtles, SmSS = smooth softshells, and SpSS = spiny softshells.
Contributor Notes
