Growth and survival outcomes for immature gopher tortoises in contrasting habitats: a test of drone-based habitat assessment

Juvenile growth rate is a critical demographic parameter, as it shortens time to maturity and often dictates how long individuals remain vulnerable to predation. However, developing a mechanistic understanding of the factors determining growth rates can be difficult for wild populations. The gopher tortoise ( Gopherus polyphemus ) is an ecosystem engineer threatened by habitat loss and deficient management of pinelands in the southeastern United States. We investigated the factors governing immature gopher tortoise growth and explored use of drone-based imagery for habitat assessment by comparing predictive models based on ground-based plant surveys versus drone-derived data. From 2021-2022, we tracked and measured immature tortoises in native sandhill and human-modified, ruderal habitat in south-central Florida. Using quarterly, high-resolution drone imagery, we quantified plant cover types and vegetation indices at each occupied burrow, and measured frequency of occurrence of forage species by hand. Annual growth rates of immature tortoises in ruderal habitat were higher than those in sandhill and were the highest published for this species. Models based on drone-derived data were able to explain similar proportions of variation in growth as ground-collected measures of forage, especially during the late dry season when both types of models were most predictive. Habitat differences in forage nitrogen content were also more pronounced during this season, when dominant ground cover in ruderal habitat (bahiagrass) had much higher nitrogen content than dominant ground cover in sandhill (wiregrass). Despite concerns about potential growth-survival trade-offs, tortoises in ruderal habitat did not exhibit lower apparent survival. Our findings indicate that habitat dominated by nutritious non-native grass can provide a valuable supplement to native sandhill through the mechanism of increased growth rates due to higher forage quality. Finally, our study demonstrates that drone technology may facilitate management by providing less labor-intensive ways to assess habitat quality for this and other imperiled herbivores.

Given the extent of anthropogenic habitat modification, the fate of many threatened species hinges on successful habitat management interventions to stabilize populations and reverse declines (Benayas et al., 2009). The effectiveness of such interventions depends on having an adequate understanding of the relationships between vital rates and habitat conditions - that is, identifying which vital rates and habitat characteristics are the most important targets for management – as well as knowing potential demographic trade-offs (de Silva & Leimgruber, 2019). Furthermore, entities implementing species management typically lack resources for intensive demographic monitoring and often rely on indicators of habitat quality to gauge whether management is having the desired beneficial effects (Benayas et al., 2009; de Silva & Leimgruber, 2019). This issue is particularly acute for long-lived species, such as many chelonians, owing to the mismatch between the multi-decadal pace of demographic change and the much shorter timescale of habitat management decisions (Congdon et al., 1993; de Silva & Leimgruber, 2019; Folt et al., 2021). Many long-lived animal species are susceptible to high predator-induced mortality when young (Wilson, 1991; Tucker , 2000). For example, mortality from depredation is particularly high for turtles during the first few years of life, when the shell has not fully hardened, and predators’ gape is not yet a limiting factor (Congdon & van Loben Sels, 1991; Stokes et al., 2006). Therefore, it can be beneficial for many chelonians to maximize growth while in early age classes to reduce the time at a vulnerable size, thus avoiding predation (Congdon & van Loben Sels, 1991; Tucker , 2000; Stokes et al., 2006; Mansfield et al., 2021). In addition to reducing time to maturity, rapid early growth may lead to larger size at maturity, further enhancing lifetime fitness (Landers et al., 1982; Congdon et al., 1993; Scott, 1994; Stamps et al., 1998). Specifically, larger body size generally correlates with increased fecundity and survivorship (Messerman et al., 2023, Otten & Refsnider, 2024), and reduced age at maturity often allows for more years being reproductively active (Iverson, 1992; Congdon et al., 1993; Harris, 2014). While beneficial to later survival, faster growth can be associated with higher mortality in the short term. Growth-survival trade-offs occur in diverse taxa, from plants (Inman-Narahari et al., 2014; Meira-Neto et al., 2019) to vertebrates (Lindeman, 1997; Mangel & Stamps, 2001; Wells & Harris, 2001). Negative consequences of rapid early growth may include reduced immune function or delayed ossification (Mangel & Stamps, 2001) or increased risk of predation while foraging and thus lower survivorship (Mangel & Stamps, 2001; Wells & Harris, 2001; Brown & Kotler, 2004; Quinn et al., 2018). For example, Jefferson salamanders ( Ambystoma jeffersonianum ) prioritize rapid larval growth at the expense of lower survival in the presence of predators (Wells & Harris, 2001). In contrast, larval four-toed salamanders ( Hemidactylium scutatum ) prioritize survival by reducing activity in the presence of predators, at the expense of growth (Wells & Harris, 2001). Similarly, most young Atlantic salmon ( Salmo salar ) feed during the night, which leads to lower growth rates but also greatly reduces risk of predation compared to those that feed during the day (Fraser & Metcalfe, 1997). As these examples demonstrate, there can be contrasting strategies to optimize fitness in response to unavoidable trade-offs between early growth and survival. Growth-survival trade-offs have been suggested but not thoroughly investigated in terrestrial turtles (Otten & Refsnider, 2024). Young terrestrial turtles may face a trade-off between forage utilization and predation risk, particularly in open habitats where there is sparse cover to hide from visual predators. This potentially occurs in gopher tortoises ( Gopherus polyphemus), a threatened, fossorial species restricted to Coastal Plain uplands of the southeastern United States (Pike, 2006; Guyer et al., 2012). The gopher tortoise is an important focal species for management because of its keystone functions and ecosystem engineering effects (Kinlaw & Grasmueck, 2012; Catano & Stout, 2015; Figueroa et al., 2021). Fire-maintained pine and sandhill communities on sandy, well-drained soils provide the highest quality natural habitat for this species (Auffenberg & Franz, 1982; Mushinsky et al., 2006). Regular burning is required to maintain the species’ preferred open-canopy conditions (Hermann et al., 2002; Ashton et al., 2008), which in turn promotes abundant herbaceous forage and enables faster growth (Mushinsky et al., 1994; Hermann et al., 2002). Although early growth is an important driver of demography via effects on age and size at maturity, modeling suggests gopher tortoise population growth rates are more sensitive to annual survival of juveniles and adult females (McKee et al., 2021; Folt et al., 2021, 2022). Hathaway (2012) and Hentges (2014) found that cattle pastures provide abundant, high-quality forage (resulting in large clutch sizes), but near-absence of juveniles suggested there was high mortality of eggs or young tortoises in such open habitat. Similarly, while adult gopher tortoises readily use roadsides to forage, anthropogenically created open areas may be ecological traps if there are concurrent risks (e.g., invasive predators, road mortality; Rautsaw et al., 2018). Our knowledge of habitat selection, growth, and survivorship of immature gopher tortoises is relatively limited compared to adults because of the difficulty of observing the small, cryptically colored, and typically less abundant juveniles (Wilson et al., 1994; Eubanks et al., 2003; Guyer et al., 2012). Estimates of habitat-specific growth and survival rates of immature gopher tortoises are needed to guide land management practices and conserve remaining populations. With this in mind, we examined how variation in forage resources affects growth of wild, immature (i.e., juvenile and subadult) gopher tortoises and whether they face trade-offs between growth and survival in high- and low-resource environments. Fine-scale habitat heterogeneity within our study area in south-central Florida enabled direct comparison of growth and survivorship between sandhill (low density of mostly native herbaceous plants) and ruderal habitat (high density of mostly non-native herbaceous plants). A previous study at this site found that adult tortoises residing in ruderal habitat attain larger body sizes (Howell et al., 2020); therefore, we hypothesized that the presence of non-native herbaceous plants in ruderal habitats provides supplemental forage for young tortoises and facilitates faster growth than experienced by similarly aged tortoises in sandhill habitats. However, we expected there would be lower survival in open, ruderal habitat compared to sandhill, which has more shrub cover. Additionally, we investigated temporal patterns of growth in this subtropical portion of the species’ range in relation to seasonal differences in forage abundance and quality. To enhance the applicability of our findings for species management, we examined how well immature growth rates can be predicted from traditional, ground-based measures of habitat quality versus data acquired using drones (i.e., unmanned aerial vehicles). Drones are becoming widely used for land management and wildlife conservation (Baratchi et al., 2013; Jiménez López & Mulero-Pázmány, 2019) and may enable more efficient monitoring of habitat quality and species’ responses to management (e.g., Scarpa & Piña, 2019). Thus, we explored the utility of drones for assessing differences in forage resources for this terrestrial, herbivorous reptile.

Study site We studied immature gopher tortoises in the “Red Hill” study population at Archbold Biological Station (Highlands Co., FL), which occupies 5 ha of ruderal habitat surrounded by approximately 84 ha of historically fire-suppressed sandhill (wiregrass-pine-oak habitat; Fig. 1). The grass-dominated ruderal area is kept open by mowing and occasional burning and has higher densities of tortoises than adjacent sandhill (Howell et al., 2020). As of 2021, the adjacent sandhill units had a time-since-fire of 1–5 years, following 1–2 prescribed burns since 2013. Prescribed burns were typically preceded by mechanical clearing, i.e., mulching of sand pine ( Pinus clausa ), oaks ( Quercus spp.), and other woody species, which aided reintroduction of fire and restoration of the desired open vegetation structure. Fig 1 Tortoise Data Collection Upon initial capture, each tortoise was measured, photographed, and given a unique set of marks by drilling or notching multiple marginal scutes. We captured tortoises by searching smaller burrows (≤24 cm wide) mapped during annual line-transect surveys of the study area. To enable monitoring of individual habitat use and growth, we initiated radio tracking of immature tortoises (1–5 years old) in Spring 2021. In total, we tagged and tracked 32 tortoises two times per week for up to six seasons (Table S1). We deployed Advanced Telemetry Systems radio transmitters weighing 0.75 g (model R1635), 1.20 g (R1655), or 3.60 g (R1680), depending on size of the tortoise (range 68–1320 g, 68–175 mm initial carapace length). We recaptured tortoises quarterly for comprehensive morphometric measurements starting March 2021 until September 2022. For purposes of this study, we divided the year into early dry season (December to March), late dry season (March to June), early wet season (June to September), and late wet season (September to December; Fig. 2). Morphometric measurements included carapace length (CL) and width, plastron length (PL) and width, and mass. Tortoise age was determined by counting the number of major annuli on the plastral scutes (Mushinsky et al., 1994). Fig. 2 Habitat data collection Forage sampling For the first three seasons of the study, we determined plant community composition by sampling 42 1-m 2 plots within 14 m of each burrow used by the radio-tagged tortoises. Seven plots were spaced at 2-m intervals along each of six transects radiating at 60-degree intervals from the burrow. We used 14 m as a radius because this is a realistic daily foraging distance for immature gopher tortoises (Wilson et al., 1994; Stemle et al., 2022). We conducted a new vegetation survey whenever a tortoise moved to a new burrow (provided it stayed > 1-2 weeks) and once every season at burrows that were used for multiple seasons. Variables calculated from these surveys included frequency of non-native bahiagrass ( Paspalum notatum ), native wiregrass ( Aristida stricta ), sedges, other non-leguminous herbaceous species, native legumes, non-native legumes, and tall woody vegetation (Table 1). Frequency was calculated as the proportion of the 42 plots surrounding each burrow in which the plant type was present. During the late dry season (May) and early wet season (August) of 2022, we collected plant tissues to determine forage quality, i.e., nitrogen content. We sampled the three most abundant forage species in each of three randomly placed 1-m 2 plots within a 14-m buffer surrounding the burrow of each radio-tagged tortoise. The plot locations were chosen using random numbers generated for distance (1–14 m) and bearing (0–360°) from the burrow. We collected enough leaves from each dominant species to reach ~2 mg of dry weight. In most cases, this meant collecting nearly all the leaves from multiple plants; however, we did not collect buds, flowers, or seeds. We placed the samples in paper bags and dried them in an oven at 50° C for 72 hrs, then sent the dried samples to the University of Georgia’s Stable Isotope Lab for processing to determine total nitrogen (N) and carbon:nitrogen ratio. Supervised classification of drone imagery We acquired high-resolution Red-Green-Blue (RGB) aerial images (3.5-cm resolution) with a DJI Phantom 4 Pro of the study area during five seasons (Fig. 2), excluding the winter season (December 2021-March 2022) when there was negligible tortoise growth. Because of cloudy conditions in the late wet season of 2022, we had to conduct two drone flights (on 12 and 20 September) to obtain shadow-free images of every burrow that were high enough quality for classification. The RGB images were post-processed using ESRI Drone2Map to produce georectified orthomosaics for each date. In addition to this initial post-processing, we used photogrammetry to produce a digital surface model (DSM) showing the elevation of the tallest object in each pixel and a digital terrain model (DTM), which categorizes the resulting point cloud and interpolates ground elevation values. We derived a vegetation height layer by subtracting the DTM from the DSM. To improve the accuracy of the classification, we used multiple subclasses to maximize spectral separation and then grouped those values into the final classes (for example, we classified white sand and yellow sand separately, then merged these into one bare ground class). At the completion of the classification, we used the vegetation height layer to help correct obviously erroneously classified pixels (for example, pixels classified as “tree” with a height value of <1 m were reclassified to woody vegetation). To further improve the accuracy of our classification, we manually edited shadow pixels within the burrow buffers to estimate which classes were not visible because of shadows. We then employed supervised classification in ArcGIS Pro (v 2.9, ESRI, Redlands, CA) to determine percent cover of bare ground, brown herb, green herb, trees, and woody vegetation within 14 m of each burrow (Fig. 3). Fig. 3 Calculation of vegetation indices We calculated two vegetation indices (VIs) from RGB drone imagery for all five seasons. Modified Green Red Vegetation Index (MGRVI) and Green Leaf Index (GLI) were calculated with the raster calculator tool in ArcGIS Pro and the mean, max, and 90 th percentile were extracted from the 14-m buffers with the zonal statistics tool. In addition to the RGB imagery, we acquired multispectral imagery in May and September 2022 using a DJI Matrice 300 RTK quadcopter equipped with a MicaSense RedEdge-MX sensor and downwelling light sensor. The flights occurred within 2 hr of solar noon at a height of 120 m AGL, yielding a resolution of ~10 cm. The resulting imagery was processed in Agisoft Metashape Professional Edition to obtain radiometrically calibrated orthomosaics consisting of blue, green, red, red edge, and near-infrared bands. These orthomosaics were then used to calculate eight multispectral VIs using the raster package in R version 4.2.1 (R Core Team, 2022) similar to Walsh et al., (2018). These multispectral VIs were loaded as rasters into ArcGIS Pro and georeferenced to the RGB imagery for each season. The multispectral VIs were then extracted to the 14-m buffers around each burrow using the zonal statistics as table tool in ArcGIS Pro to get the maximum, mean, and 90 th percentile within the buffers. Statistical analyses Growth analyses For tortoises tracked at least 12 mo (March 2021–March 2022), we used a two-sample t-test to compare annual growth (CL % growth per day) between habitats. For individuals tracked all six seasons (see Table S1), we ran a repeated-measures analysis of variance (ANOVA) to examine effects of habitat, time (i.e., season), and their interaction on growth measured at quarterly intervals. The repeated-measures ANOVA was conducted in R Studio using the dplyr package (Wickham et al., 2022). Post-hoc tests consisted of one-way ANOVAs comparing growth between habitats within each season using the Benjamini-Hochberg procedure to correct for multiple comparisons. To evaluate which variables were the best predictors of tortoise growth (CL % growth per day), we utilized model selection using the dredge function in R Studio (Bartoń, 2022) to identify the best-supported model based on AICc. We compared each set of predictors and their R 2 values for each season, as different subsets of predictors were available for each season (Table 1). We included initial CL as a covariate to account for ontogenetic change in growth rate. To ensure no variables were correlated in each top model, we verified that all VIFs were <10. Forage analyses We used a factorial ANOVA to examine effects of season and habitat (ruderal versus sandhill) on the N content of forage species averaged across the three plots at each burrow. Additionally, we compared percent total N between habitats for each season separately using one-way ANOVAs. Survival analyses Using a POPAN formulation of the Jolly‐Seber model in package marked (R Studio v 4.0.0, R Core Team, 2022; Laake et al., 2013), we estimated the quarterly apparent survival of the radio-tracked tortoises from March 2021 to September 2022. We considered models with both constant and time-varying values for apparent survival (phi) and recapture probability (pent), resulting in four candidate models. We used the model with the lowest AICc to estimate survival of all immature tortoises, as well as small (126 mm CL) tortoises separately. In addition, we estimated apparent survival separately for each habitat, excluding three tortoises that spent significant amounts of time in both habitats (see Table A1). For habitat-specific analysis, we only included tortoises that spent >80% of their time in a single habitat during the 18-mo study period. Tortoise growth Seasonal and habitat-specific growth rates From March 2021 to March 2022, tortoise growth rate differed significantly between ruderal and sandhill habitat (F 1,13 = 12.0, p = 0.004; Fig. 4). Mean annual growth rates were 34.4 ± 3.5 mm (n = 6) in ruderal habitat and 19.5 ± 3.4 mm (n = 8) in sandhill. Quarterly growth rates differed significantly between habitats (F 1,10 = 6.9, p = 0.003) and across seasons (F 5,50 = 40.4, p < 0.001; Fig. 5, Table S2), including a significant habitat x season interaction (F 5,50 = 4.4, p < 0.001; Fig. 5). Growth was consistently higher in ruderal habitat, with the strongest effect in the late dry season and smaller effects in the early wet and late wet seasons (Tables S2 and S3). There was no difference in growth between habitats in the early dry season, because growth was near zero during this coldest part of the year. Growth also did not vary significantly between habitats during the early wet season of 2022, which may be due to the lower sample size compared to the early wet season of 2021 (Tables S2 and S3). Fig. 4 Fig. 5 Predictors of tortoise growth When available, data from ground-based plant surveys were always more predictive of tortoise growth than remotely acquired data. We found positive effects of BAHIAGRASS, SEDGE, and NN LEGUME on growth and negative effects of TALL WOODY (Table 2). The environmental variables were most predictive of tortoise growth in the late dry season 2021 (Figs 5 and 6) and the early wet season 2022 (i.e., after the first significant rainfall in those respective years). During these seasons, drone-derived indices explained a similar percentage of the variation in immature tortoise growth (R 2 = 0.71 for ground-based plant surveys; R 2 = 0.64 for drone-derived indices). Among the drone-derived indices, GREEN HERB was the best predictor of tortoise growth (top model in three of five seasons; Fig. 6). WOODY VEG, which is largely the inverse of GREEN HERB (r = -0.89, p < 0.001), was also informative, as were a couple of the VIs based on RGB data. Specifically, GLI, which measures density of greenness (oak leaves high value, soil negative value), and MGRVI, which measures just the Green and Red bands, provided useful drone-derived predictors of immature tortoise growth. In contrast, none of the VIs based on multispectral data were good predictors of growth (Table 2). Fig. 6 Forage composition and quality Bahiagrass was the dominant forage plant in ruderal habitat (42.8% of plots), whereas wiregrass was dominant in sandhill (45.5% of plots; Tables S4 and S5). Thus, the difference in forage quality between the two habitat types can largely be attributed to these two species. Wiregrass consistently had the lowest N content and bahiagrass consistently had higher N than wiregrass (two-sample t-test, p < 0.001; Table S5). However, the difference in forage quality between the two habitat types decreased between the late dry and early wet seasons for two reasons. First, N content of bahiagrass decreased over this time interval (two-sample t-test, p < 0.001), thus diminishing N availability in the ruderal habitat. Second, the diversity of forage species in the sandhill increased (late dry Shannon diversity = 2.40; early wet Shannon diversity = 3.69), decreasing the influence of the N-poor wiregrass. In both seasons, leguminous (native or non-native) species had higher N than other species (May: F 1,15 = 6.3, p = 0.024; August: F 1,18 = 7.2, p = 0.015). Collectively, the seasonal and habitat-specific patterns of N content of forage species mirror the observed patterns of immature tortoise growth. Forage species in ruderal habitat had significantly higher total N than forage in sandhill habitat in the 2022 late dry season (F 1,84 = 6.1, p = 0.015), when tortoise growth was similarly greater in ruderal habitat (Fig. 5). Total N did not differ between habitats in the 2022 early wet season (F 1, 133 = 0.1, p = 0.770), when tortoise growth was also statistically equivalent between the two habitats (Table S3). Tortoise survival For all time periods and analyses, the top model had constant survival across time periods (phi~1). Quarterly apparent survival of tagged immature tortoises across both habitats from March 2021–September 2022 was 94.2% ± 2.2 (95% CI: 88.0%–97.3%; n = 32). Quarterly apparent survival was similar for immature tortoises residing in ruderal habitat (94.9% ± 3.5, 95% CI: 82.1%–98.7%; n = 12) and in sandhill (93.1% ± 3.3, 95% CI: 83.0%–97.4%; n = 16). These estimates correspond to annual apparent survival rates of 81.1% and 75.1% for ruderal and sandhill habitats, respectively, or 79.3% for all tortoises combined. Quarterly apparent survival did not differ significantly between size classes, although point estimates were lower for small tortoises (91.6% ± 3.5, 95% CI: 81.6%–96.4%; n = 19) than for large tortoises (96.4% ± 2.5, 95% CI: 86.8%–99.1%; n = 13). These estimates correspond to annual apparent survival rates of 70.4% (tortoises 126 mm CL). For all time periods and analyses, the top model had constant survival across time periods (phi~1). Quarterly apparent survival of tagged immature tortoises across both habitats from March 2021–September 2022 was 94.8% ± 2.1 (95% CI 88.8%–97.7%; n = 32). Quarterly apparent survival was similar for immature tortoises residing in ruderal habitat (95.7% ± 3.1, 95% CI 83.5%–99.1%; n = 12) and in sandhill (93.1% ± 3.3, 95% CI 83.0%–97.4%; n = 16). These estimates correspond to annual apparent survival rates of 83.9% and 75.1% for ruderal and sandhill habitats, respectively, or 80.8% for all tortoises combined. Time investment per method Drone imagery required a ~0.5-hr flight if it was for a RGB-derived vegetation index or ~2 hr for a multispectral index, and many hours of post-processing and computer run time; the estimated total time for calculating spectral indices was approximately 30 hr, with multispectral taking closer to 35 hr (25–40 hr). Supervised classification required a ~0.5-hr drone flight, post-processing and computer run time; the estimated total time for supervised classification was approximately 38 hrs. Plant surveys required ~60 person-hr of fieldwork (30–60 min per burrow with 2–4 people) plus 5 hr entering data, for a total estimate of 65 hr per season. Thus, considering it took ~65 hr for plant surveys and ~34 hr for drone data acquisition (averaging the 30 hr for spectral indices and 38 hr for supervised classification), drone-based methods required ~48% less time.

Our examination of individual variation in growth rates of immature gopher tortoises with respect to forage resources provides valuable new insights for wildlife managers aiming to maintain population resilience under anthropogenically modified disturbance regimes. We found that growth of immature gopher tortoises varied significantly by habitat and by season. In ruderal habitat, we observed the highest annual growth rates ever reported for wild gopher tortoises, an average of 34.4 mm/yr compared with 5–19 mm/yr at other sites throughout the species’ range (Harris, 2014). In comparison, immature tortoises in sandhill habitat grew only slightly faster than those at another sandhill site in central Florida (19.5 mm/yr vs. 19 mm/yr; Mushinsky et al., 1994). Within our study population, there is a strong positive relationship between immature tortoise growth and abundance of bahiagrass and other non-native grasses/legumes (found mostly in ruderal habitats), whereas growth is negatively correlated with abundance of woody vegetation (found mostly in sandhill). These patterns were apparent not only from ground-based plant surveys, but also using supervised classification or vegetation indices derived from high-resolution drone imagery (see Table 2), enabling similar estimates of forage quality over much larger areas in the future. We further demonstrate that the advantage in forage quality of the ruderal habitat is most prevalent during the late dry season, a time when native food resources are scarce. Additionally, we have developed a mechanistic understanding of growth rate variation by finding an identical habitat-by-season pattern in the N content of available forage. In support of this conclusion, previous work has shown N-rich plants are important for growing gopher tortoises (Mushinsky et al., 2003; Hathaway, 2012). Our study also clarifies effects of habitat quality on demography by demonstrating how habitat heterogeneity contributes to within-population differences in age-specific body size. Specifically, our linear model of size versus age projects that tortoises in ruderal habitat at Archbold Biological Station can attain the minimum size for sexual maturity and likely earlier than reported for any other population (Harris, 2014; Meshaka et al., 2019). Rapid growth enables young tortoises to reach sexual maturity faster, thus increasing fitness by reducing time at a vulnerable size and expanding the number of reproductive years (Mangel & Stamps, 2001; Harris, 2014). Just as fine-scale differences in habitat quality create divergent growth trajectories within a population, site-to-site differences in habitat quality combined with latitudinal variation in length of the activity season, result in widely varying mean growth rates among populations throughout the species’ range (Mushinsky et al., 2006). Our hypothesis of a growth-survival trade-off was not supported because apparent survivorship rates showed no evidence of negative effects of ruderal habitat on survival of this age class (1–7 years). If anything, survival tended to be higher in the ruderal habitat and thus in the same direction as that habitat’s higher growth potential. This contrasts with previous hypotheses that faster growth could be associated with trade-offs in fitness, such as increased predation risk while foraging or reduced immune function (Lindeman, 1997; Mangel & Stamps, 2001), which still needs to be examined. Generally, mowed grass habitats are considered an ecological trap for turtle species, as they are often associated with roadsides, where turtles face higher mortality risk (Aresco, 2005; Shepard et al., 2008; Rautsaw et al., 2018). However, at our study site, there is high driver awareness of tortoises and mowing is restricted to early morning when tortoises are less active aboveground. Additionally, survivorship could differ among sites due to differences in predator community composition or differences in vegetation structure and cover (e.g., grass height, shrub density). It is also important to note that our study did not assess the survival of neonates; thus, it remains unclear if gopher tortoises would experience a growth-survival trade-off during the first year post-hatching. Given ongoing urbanization and projected population declines (USFWS, 2021; Folt et al., 2022), we expect more intensive management interventions will be required to stabilize gopher tortoise populations. Mitigation translocations of gopher tortoises are already common (Seigel & Dodd, 2000), particularly in Florida, where thousands of permits have been issued in recent years for onsite relocation or offsite translocation of tortoises from areas slated for development to permitted recipient sites with suitable habitat (USFWS, 2021). Both translocated and natural populations may be subject to increased predation rates, either from subsidization of mesopredators (Harden et al., 2009) or spread of novel invasive predators (Offner et al., 2021). These rising threats lend urgency to research on factors influencing recruitment and ways to enhance juvenile growth and survival to boost population resiliency. In the presence of greater predation pressure, promoting faster juvenile growth might improve demographic resilience by reducing the window of extreme vulnerability (Folt et al., 2021). Along these lines, our results point to intentional manipulation of forage resources as a useful avenue for future research. As we have shown, the availability, abundance, and higher N content of non-native grasses and other plants during the late dry season and into the early wet season enhances tortoise growth in ruderal habitat. Thus, interspersion or maintenance of some ruderal habitat could be beneficial to gopher tortoises in certain situations. For example, habitat management options are often limited in urbanizing landscapes due to restrictions on prescribed burning (Yager et al., 2007; USFWS, 2021; Folt et al., 2022). We suggest it would be worthwhile to test the demographic effects of establishing small, interspersed plots of perennial, non-native forage, particularly in areas where native groundcover is lacking or difficult to establish. Plants that have high forage value, such as legumes, could also be incorporated. While forage supplementation alone might offer limited benefits, it could be a useful, low-cost tool when combined with other intensive management efforts, for example, to enhance post-release growth and site fidelity of head-started or translocated tortoises (Tuberville et al., 2021; McKee et al., 2021). Manipulation of forage availability and cover is a common management practice for quail and other game birds (Brunk et al., 2023), though supplemental feeding has had varied success (Henry et al., 2022) . Importantly, we are not advocating conversion of high-quality native habitat to artificially created “food plots” to support gopher tortoises, as this would be a species-specific viewpoint that fails to consider the conservation value of the entire sandhill community. It is also critical to avoid introduction or facilitation of invasive predators, such as red imported fire ants ( Solenopsis invicta ) or invasive plants, but bahiagrass has not shown a tendency to invade sandhill at our study site (Rothermel, pers. obs). In general, we see a need to investigate this and other novel habitat management techniques that might benefit this keystone species. Use of drone imagery is becoming prevalent in ecological and conservation research (Blake et al., 2013; Jiménez López & Mulero-Pázmány, 2019; Marcelino et al., 2020). Although ground-based plant surveys always provided the strongest predictors of immature tortoise growth in our study, drone data were able to closely approximate the same result in the late dry/early wet season when plant growth is stimulated by increased precipitation. During this time of year, the drone data identified green herbaceous cover (or GLI, MGVRI, NDVI) as the primary driver of immature tortoise growth, which encompasses the more refined driver of bahiagrass identified by ground-based plant surveys. Moreover, while both conventional field methods and drone data provided weaker predictors of growth during the remaining seasons, our repeated-measures analysis indicated that ruderal habitat supported the fastest growth throughout the entire year, and it was just the magnitude of this advantage that varied seasonally, not the direction of the effect. Thus, identifying the best growth habitat using a drone flight in the late dry/early wet season when the predictive power is highest would not lead one to identify habitat that was suboptimal at other times of year. However, given differences in rainfall and vegetation phenology throughout the range of gopher tortoises, the optimal timing of flights would likely vary as well. For areas north of central Florida, if there is some supplemental forage available during the driest part of the active season (when temperatures are high enough for foraging), then we would expect to see similar effects on growth. A similar study conducted in the northern part of the species’ range could clarify. In summary, our results generated an improved mechanistic understanding of juvenile growth, which informs lifetime fitness. As the enhanced growth of immature (1- to 7-year-old) tortoises does not trade off with lower survival, well-timed drone imagery can be used to identify high-quality habitat. Although drone-based habitat assessment requires access to a drone and sufficient data storage capability, as well as significant post-processing time, we found a roughly 50% time savings compared to ground-based surveys. We therefore suggest that supervised classification or vegetation spectral indices (not necessarily multispectral indices) of high-resolution drone imagery are useful methods for assessing gopher tortoise forage resources. We especially see a role for drones when there is a need to assess large areas, such as entire preserves or recipient sites, as these remote methods scale up more efficiently than ground-based plant surveys. Thus, our study adds another example to a growing list of potential applications of drone technology that can facilitate more effective, cost-efficient habitat monitoring and management to benefit imperiled species (Baratchi et al., 2013).

This work was generously supported by a Disney Conservation Fund award to Archbold Biological Station. We thank Archbold Herpetology staff Rachel Fedders, Alonso Reyes, Amanda West, Jared Draxler, Madison Molter, Kate Schmid, Kameron Voves, and Theresa Fonseca for assisting with data collection. Volunteers Marc Behrendt, Gabriel Kamener, Linda Gette, Hunter Howell, Jackie Hunt, and Abigail Arnashus also provided invaluable help in the field. Vivienne Sclater and Ryan Howell (Archbold) assisted with image classification and data management, and members of the Searcy Lab provided feedback on the manuscript. We conducted research under Florida Fish and Wildlife Conservation Commission permits LSSC 10-00043G and 10-00043H. Animal handling was approved by Archbold’s animal use committee and conformed to ASIH (2004) Guidelines for Use of Live Amphibians and Reptiles in Field and Laboratory Research, as well as University of Miami IACUC protocol #21-054. Author contributions LRS, BBR and CAS conceived the ideas and designed the methodology. BBR obtained funding. LRS, BBR, CAS, JMS, CLM, and JTC collected the data. JMS performed image analysis. LRS, BBR and CAS analyzed the data. LRS led the writing of the manuscript. BBR and JTC assisted with figures and tables. All authors contributed to the drafts and gave final approval for publication. Conflict of interest The authors state no conflicts of interest. Data availability statement Data and code for much of the project can be found on GitHub: https://github.com/lstemle/Immature_GT_growth_survival.git . Other data is available upon request. Table 1. Predictor variables included in models of immature tortoise growth, and seasons when each variable was measured. All variables were measured within 14 m of burrows used by radio-tagged tortoises. If a tortoise used two burrows within the same season, we calculated a weighted average of each metric based on the proportion of time the tortoise used each burrow. GV = ground-based vegetation survey, D-SC = supervised classification of RGB drone imagery, D-RGB = calculated from RGB drone data, D-MS = calculated from multi-spectral drone data, VI = vegetation index. | Variable name | Method | Description | Mar-Jun | Jun-Sep | Sep-Dec | Mar-Jun | Jun-Sep | | BAHIAGRASS | GV | Frequency of non-native bahiagrass | • | • | • | || | WIREGRASS | GV | Frequency of native wiregrass | • | • | • | || | OTHER HERB | GV | Frequency of other herbs | • | • | • | || | LEGUME | GV | Frequency of native legumes | • | • | • | || | NN LEGUME | GV | Frequency of non-native legumes | • | • | • | || | SEDGE | GV | Frequency of sedges | • | • | • | || | TALL WOODY | GV | Frequency of tall woody plants | • | • | • | || | GREEN HERB | D-SC | % cover of green herbaceous veg | • | • | • | • | • | | BROWN HERB | D-SC | % cover of brown herbaceous veg | • | • | • | • | • | | WOODY VEG | D-SC | % cover of wood vegetation | • | • | • | • | • | | TREE | D-SC | % cover of trees | • | • | • | • | • | | BARE GROUND | D-SC | % cover of bare ground | • | • | • | • | • | | MGRVI | D-RGB | Modified Green Red VI | • | • | • | • | • | | GLI | D-RGB | Green Leaf Index | • | • | • | • | • | | NDVI | D-MS | Normalized Difference VI | • | • | ||| | NDRE | D-MS | Normalized Difference Red Edge Index | • | • | ||| | EVI2 | D-MS | 2-band Enhanced VI | • | • | ||| | SRre | D-MS | Red-Edge Simple Ratio | • | • | ||| | CIgreen | D-MS | Green Chlorophyll Index | • | • | ||| | CIred-edge | D-MS | Red Edge Chlorophyll Index | • | • | ||| | MTCI | D-MS | MERIS Terrestrial Chlorophyll Index | • | • | ||| | RTVICore | D-MS | Red-Edge Triangulated VI | • | • | Table 2. Best models of tortoise growth by season (see Table 1 for variable descriptions and seasons measured). Values in parentheses are standardized regression coefficients and significant relationships (P < 0.05) are indicated in bold. Early dry season is not included due to negligible growth during this period. | Season | Ground-based plant survey | Drone - supervised classification | Drone - RGB indices | Drone - multispectral indices | | Late dry 2021 (Mar-Jun) | R 2 = 0.71 BAHIAGRASS (0.839) Initial CL (-0.153) | R 2 = 0.64 GREEN HERB (0.798) Initial CL (-0.211) | R 2 = 0.59 90% GLI (-1.687) 90% MGRVI (1.104) Initial CL (-0.289) | NA | | Early wet 2021 (Jun-Sep) | R 2 = 0.55 SEDGE (0.516) NN LEGUME (0.651) Initial CL (-0.367) | R 2 = 0.28 GREEN HERB (0.429) Initial CL (-0.310) | R 2 = 0.23 MGRVI Max (0.3719) Initial CL (-0.279) | NA | | Late wet 2021 (Sep-Dec) | R 2 = 0.43 SEDGE (0.358) TALL WOODY (-0.370) Initial CL (-0.320) | R 2 = 0.24 GREEN HERB (0.410) Initial CL (-0.363) | null model | NA | | Late dry 2022 (Mar-Jun) | NA | R 2 = 0.31 WOODY VEG (-0.438) Initial CL (-0.252) | R 2 = 0.27 MGRVI Max (0.392) Initial CL (-0.336) | null model | | Early wet 2022 (Jun-Sep) | NA | R 2 = 0.44 WOODY VEG (-0.371) Initial CL (-0.544) | R 2 = 0.64 MGRVI Max (0.844) 90% GLI (-0.739) Initial CL (-0.426) | R 2 = 0.601 NDVI Max (0.424) Initial CL (-0.613) |

Aresco, M.J. (2005). The effect of sex-specific terrestrial movements and roads on the sex ratio of freshwater turtles. Biol. Conserv. 123, 37–44. Ashton, K.G., Engelhardt, B.M., & Branciforte, B.S. (2008). Gopher tortoise ( Gopherus polyphemus ) abundance and distribution after prescribed fire reintroduction to Florida scrub and sandhill at Archbold Biological Station. J Herpetol. 42, 523–529. Auffenberg, W., & Franz, R. (1982). The status and distribution of the gopher tortoise ( Gopherus polyphemus ). In North American tortoises: conservation and ecology : 95–126. Bury, R.B. (Ed.). Washington, D.C: Wildlife Research Report 12, U.S. Fish and Wildlife Service. Baratchi, M., Meratnia, N., Havinga, P., Skidmore, A., & Toxopeus, B. (2013). Sensing solutions for collecting spatio-temporal data for wildlife monitoring applications: a review. Sensors 13, 6054–6088. Bartoń, K. (2022). MuMIn: Multi-Model Inference. R package version 1.46.0. https://CRAN.R-project.org/package=MuMIn. Benayas, J.M.R., Newton, A.C., Diaz, A., & Bullock, J.M. (2009). Enhancement of biodiversity and ecosystem services by ecological restoration: a meta-analysis. Science 325, 1121–1124. Blake, S., Yackulic, C.B., Cabrera, F., Tapia, W., Gibbs, J.P., Kümmeth, F., & Wikelski, M. (2013). Vegetation dynamics drive segregation by body size in Galapagos tortoises migrating across altitudinal gradients. J. Anim. Ecol. 82, 310–321. Brown, J.S., & Kotler, B.P. (2004). Hazardous duty pay and the foraging cost of predation: foraging cost of predation. Ecol. Lett. 7, 999–1014. Brunk, K.M., Gutiérrez, R.J., Peery, M.Z., Cansler, C.A., Kahl, S., & Wood, C.M. (2023). Quail on fire: changing fire regimes may benefit mountain quail in fire-adapted forests. Fire Ecol . 19, 1–13. Catano, C.P., & Stout, I.J. (2015). Functional relationships reveal keystone effects of the gopher tortoise on vertebrate diversity in a longleaf pine savanna. Biodivers. Conserv. 24, 1957–1974. Congdon, J.D., & Van Loben Sels, R.C. (1991). Growth and body size in Blanding’s turtles ( Emydoidea blandingi ): relationships to reproduction. Can. J. Zool. 69, 239–245. Congdon, J.D., Dunham, A.E., & Van Loben Sels, R.C. (1993). Delayed sexual maturity and demographics of Blanding’s turtles ( Emydoidea blandingii ): implications for conservation and management of long-lived organisms. Conserv. Biol. 7, 826–833. De Silva, S., & Leimgruber, P. (2019). Demographic tipping points as early indicators of vulnerability for slow-breeding megafaunal populations. Front. Ecol. Evol. 7, 171. Eubanks, J.O., Michener, W.K., & Guyer, C. (2003). Patterns of movement and burrow use in a population of gopher tortoises ( Gopherus polyphemus ). Herpetologica 59, 311–321. Figueroa, A., Lange, J., & Whitfield, S.M. (2021). Seed consumption by gopher tortoises ( Gopherus polyphemus ) in the globally imperiled pine rockland ecosystem of Southern Florida, USA. Chelonian Conserv. Biol. 20, 27–31. Folt, B., Marshall, M., Emanuel, J.A., Dziadzio, M., Cooke, J., Mena, L., Hinderliter, M., Hoffmann, S., Rankin, N., Tupy, J., & McGowan, C. (2022). Using predictions from multiple anthropogenic threats to estimate future population persistence of an imperiled species. Glob. Ecol. Conserv. 36, e02143. Folt, B., Goessling, J. M., Tucker, A., Guyer, C., Hermann, S., Shelton‐Nix, E., & McGowan, C. (2021). Contrasting patterns of demography and population ability among gopher tortoise populations in Alabama. J. Wildl. Manag. 85, 617–630. Fraser, N.H.C., & Metcalfe, N.B. (1997). The costs of becoming nocturnal: feeding efficiency in relation to light intensity in juvenile Atlantic Salmon. Funct. Ecol. 11, 385–391. Guyer, C., Johnson, V.M., & Hermann, S.M. (2012). Effects of population density on patterns of movement and behavior of gopher tortoises ( Gopherus polyphemus ). Herpetol. Monogr. 26, 122–134. Harden, L., Price, S., & Dorcas, M. (2009). Terrestrial activity and habitat selection of Eastern mud turtles ( Kinosternon subrubrum ) in a fragmented landscape: implications for habitat management of golf courses and other suburban environments. Copeia 2009, 78–84. Harris, B.B. (2014). Ecology of juvenile gopher tortoises (Gopherus polyphemus) on a Georgia barrier island. MS Thesis, University of Georgia. Hathaway, A.L. (2012). Availability and quality of vegetation affects reproduction of the gopher tortoise (Gopherus polyphemus) in improved pastures . Dissertation, University of South Florida. Henry, B.J., Brym, M.Z., Henry, C., & Kendall, R.J. (2022). Supplemental feeding of northern bobwhite ( Colinus virginianus ) and dietary requirements: a review. Wildl. Res. 49, 89–99. Hentges, T.W. (2014). Are gopher tortoises (Gopherus polyphemus; Daudin) compatible with cows? . MS. Thesis, University of South Florida. Hermann, S.M., Guyer, C., Hardin Waddle, J., & Greg Nelms, M. (2002). Sampling on private property to evaluate population status and effects of land use practices on the gopher tortoise, Gopherus polyphemus . Biol. Conserv. 108, 289–298. Howell, H.J., Rothermel, B.B., White, K.N., & Searcy, C.A. (2020). Gopher tortoise demographic responses to a novel disturbance regime. J. Wildl. Manag. 84, 56–65. Inman-Narahari F., Ostertag R., Asner G.P., Cordell S., Hubbell S.P., & Sack L. (2014). Trade-offs in seedling growth and survival within and across tropical forest microhabitats . Ecol Evol . 4, 3755–67. Iverson, J.B. (1992). Correlates of reproductive output in turtles (Order Testudines). Herpetol. Monogr. 6, 25–42. Jiménez López, J. & Mulero-Pázmány, M. (2019). Drones for conservation in protected areas: present and future. Drones 3, 10. Kinlaw, A., & Grasmueck, M. (2012). Evidence for and geomorphologic consequences of a reptilian ecosystem engineer: the burrowing cascade initiated by the gopher tortoise. Geomorphology 157, 108–121. Laake, J.L., Johnson D., & Conn, P. (2013). marked: an R package for maximum-likelihood and MCMC analysis of capture-recapture data. Methods Ecol. Evol . 2013, 885–890. Landers, J.L., McRae, W.A., & Garner, J.A. (1982). Growth and maturity of the gopher tortoise in southwestern Georgia. Bull. Florida State Mus. 27, 81–110. Lindeman, P.V. (1997). Contributions toward improvement of model fit in nonlinear regression modelling of turtle growth Herpetologica 53, 179–191. Mangel, M., & Stamps, J. (2001). Trade-offs between growth and mortality and the maintenance of individual variation in growth. Evo. Ecol. Res. 3, 583–593. Mansfield, K.L., Wyneken, J. & Luo, J. (2021). First Atlantic satellite tracks of ‘lost years’ green turtles support the importance of the Sargasso Sea as a sea turtle nursery. P. Roy. Soc. B. 288, 20210057. Marcelino, J., Silva, J.P., Gameiro, J., Silva, A., Rego, F.C., Moreira, F. & Catry, I., (2020). Extreme events are more likely to affect the breeding success of lesser kestrels than average climate change. Sci. Rep. 10, 7207. McKee, R.K., Buhlmann, K.A., Moore, C.T., Hepinstall‐Cymerman, J., & Tuberville, T.D. (2021). Waif gopher tortoise survival and site fidelity following translocation. J. Wildl. Manag. 85, 640–653. Meira‐Neto, J.A.A., Nunes Candido, H.M., Miazaki, Â., Pontara, V., Bueno, M.L., Solar, R. and Gastauer, M. (2019). Drivers of the growth–survival trade‐off in a tropical forest. J. Veg. Sci. 30, 1184–1194. Meshaka Jr., W.E., Layne, J.N., & Rice, K.G. (2019). The effects of geography, habitat, and humans on the ecology and demography of the gopher tortoise in the southern Lake Wales Ridge region of Florida. Herpetol. J. 29, 95–114. Messerman, A.F., Clause, A.G., Gray, L.N., Krkošek, M., Rollins, H.B., Trenham, P.C., Shaffer, H.B., & Searcy, C.A. (2023). Applying stochastic and Bayesian integral projection modeling to amphibian population viability analysis. Ecol. App. 33, p.e2783. Mushinsky, H.R., Stilson, T.A., & McCoy, E.D. (2003). Diet and dietary preference of the juvenile gopher tortoise ( Gopherus polyphemus ). Herpetologica 59, 475–483. Mushinsky, HR., McCoy, E.D., Berish, J.E., Ashton, R.E., & Wilson, D.S. (2006). Gopherus polyphemus – Gopher Tortoise. In Biology and Conservation of Florida Turtles : 346–371. Meylan (Ed.), Arlington, VT: Chelonian Research Foundation. Mushinsky, H.R., Wilson, D.S., & McCoy, E.D. (1994). Growth and sexual dimorphism of Gopherus polyphemus in central Florida Herpetologica 1994, 119–124 . Offner, M.T., Campbell, T.S., & Johnson, S.A. (2021). Diet of the invasive Argentine black and white tegu in central Florida . Southeastern Nat. 20, 319–337. Otten, J.G., & Refsnider, J.M. (2024). Bigger is better: age class‐specific survival rates in long‐lived turtles increase with size. J. Wildl. Manag . 88, e22544. Pike, D.A. (2006). Movement patterns, habitat use, and growth of hatchling tortoises, Gopherus polyphemus . Copeia 2006, 68–76. Quinn, D.P., Buhlmann, K.A., Jensen, J.B., Norton, T.M., & Tuberville, T.D. (2018). Post-release movement and survivorship of head-started gopher tortoises. J. Wildl. Manag. 82, 1545–1554. Rautsaw, R.M., Martin, S.A., Vincent, B.A., Lanctot, K., Bolt, M.R., Seigel, R.A., & Parkinson, C.L. (2018). Stopped dead in their tracks: the impact of railways on gopher tortoise ( Gopherus polyphemus ) movement and behavior. Copeia 106, 135–143. R Core Team (2022). R : A language and environment for statistical computing . Vienna, Austria: R Foundation for Statistical Computing. https://www.R-project.org/. Scarpa, L.J., & Piña, C.I. (2019). The use of drones for conservation: a methodological tool to survey caimans nests density. Biol. Conserv. 238, 108235. Scott, D.E. (1994). The effect of larval density on adult demographic traits in Ambystoma opacum . Ecology 75,1383–1396. Seigel, R.A., & Dodd, Jr. C.K. (2000). Manipulation of turtle populations for conservation. Halfway technologies or viable populations? In Turtle Conservation : 218–238. Klemens M.W. (Ed.). Washington, D.C.: Smithsonian Institution Press. Shepard, D.B., Kuhns, A.R., Dreslik, M.J., & Phillips, C.A. (2008). Roads as barriers to animal movement in fragmented landscapes. Anim. Conserv. 11, 288–296. Stamps, J.A., Mangel, M.M., & Phillips, J.A. (1998). A new look at relationships between size at maturity and asymptotic size. Am. Nat . 152, 470–479. Stemle, L., Rothermel, B.B., & Searcy, C.A. (2022). GPS technology reveals larger home ranges for immature gopher tortoises. J. Herpetol. 56, 172–179. Stokes, L., Wyneken, J., Crowder, L. B., & Marsh, J. (2006). The influence of temporal and spatial origin on size and early growth rates in captive loggerhead sea turtles ( Caretta caretta ) in the United States. Herpetol. Conserv. Biol. 1 , 71-80 . Tuberville, T.D., Quinn, D.P., & Buhlmann, K.A. (2021). Movement and survival to winter dormancy of fall-released hatchling and head-started yearling gopher tortoises. J. Herpetol. 55, 88–94. Tucker, J.K. (2000). Body size and migrations of hatchling turtles: inter- and intraspecific comparisons. J. Herpetol. 34, 541–546. USFWS. (2021). Species status assessment report for the gopher tortoise (Gopherus polyphemus), version 0.4 . Atlanta: U.S. Fish and Wildlife Service. Walsh, O.S., Shafian, S., Marshall, J.M., Jackson, C., McClintick-Chess, J.R., Blanscet, S.M., Swoboda, K., Thompson, C., Belmont, K.M., & Walsh, W.L. (2018). Assessment of UAV based vegetation indices for nitrogen concentration estimation in spring wheat. Adv. Remote Sens. 7, 71–90. Wells, C.S. & Harris, R.N. (2001). Activity level and the tradeoff between growth and survival in the salamanders Ambystoma jeffersonianum and Hemidactylium scutatum . Herpetologica 57, 116–127. Wickham, H., François, R., Henry, L., & Müller, K. (2022). Dplyr: A Grammar of Data Manipulation . https://dplyr.tidyverse.org. Wilson, D.S. (1991). Estimates of survival for juvenile gopher tortoises, Gopherus polyphemus . J. Herpetol. 25, 376–379. Wilson, D.S., Mushinsky, H.R., & McCoy, E.D. (1994). Home range, activity, and burrow use of juvenile gopher tortoises (Gopherus polyphemus) in a central Florida population. In Biology of North American Tortoises : 147–160. Bury, R.B., & Germano, D.J. (Eds.). Washington, D.C.: U.S. Department of the Interior. Yager, L.Y., Heise, C.D., Epperson, D.M., & Hinderliter, M.G. (2007). Gopher tortoise response to habitat management by prescribed burning. J. Wildl. Manag. 71, 428–43. Supporting Information Table S1. Body size, age, tracking history, and seasonal habitat assignments for every gopher tortoise included in the study of immature growth and survival at Archbold Biological Station, Florida, in 2021-2022. Seasons included in the growth analysis are in bold. Table S2. Seasonal mean percent growth in carapace length (CL % growth per day) in each habitat, based on every radio-tagged tortoise we obtained measurements for in that season (provided they could be classified as living in a single habitat type). Table S3. Results of separate one-way ANOVAs testing for effects of habitat (ruderal vs. sandhill) on immature tortoise growth in each season. P-values were corrected for multiple comparisons using the Benjamini-Hochberg procedure. Table S4. Frequency of different plant taxa within 14 m of burrows used by immature gopher tortoises in ruderal and sandhill habitats at Archbold Biological Station, Florida. Vegetation surveys occurred in 2021 during the late dry season (12 May–7 June; n = 23 burrows), early wet season (6 August–27 September; n = 36), and late wet season (8 November–4 January; n = 31). Frequency was calculated as the mean proportion of plots across all burrows sampled containing each type of plant, averaged across all burrows sampled within that season. Table S5. Mean percent total nitrogen (N) and mean carbon:nitrogen ratio (C:N) of forage species in late dry season (May) and early wet season (August) of 2022 in ruderal and sandhill habitat at Archbold Biological Station, Florida. Dashes (–) indicate no plants of that species were sampled during that month. Non-native species are in bold. Supplementary Material File (figure_captions_final.docx) - Download - 17.69 KB Information & Authors Information Version history Peer review timeline Published Ecology and Evolution Version of Record3 Nov 2024Published Copyright This work is licensed under a Non Exclusive No Reuse License. Collection

Authors Funding Information Metrics & Citations Metrics Article Usage 540views 192downloads Citations Download citation Leyna Stemle, Julie M Sorfleet, Chelsea L Moore, et al. Growth and survival outcomes for immature gopher tortoises in contrasting habitats: a test of drone-based habitat assessment. Authorea. 03 September 2025. DOI: https://doi.org/10.22541/au.172114985.51304046/v2 DOI: https://doi.org/10.22541/au.172114985.51304046/v2 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu.