Michael Hilton

Social network analyses are sparse, despite having great potential to illuminate intricate details of wildlife behavioral ecology and to inform basic conservation practices. Using social interactions recorded during 1 year of 5-second interval photography, we conducted social network analyses of Gopher Tortoises (Gopherus polyphemus). G. polyphemus are charismatic and declining mid-sized tortoises that are habitat specialists endemic to the southeastern United States. We also conducted a simultaneous radio-telemetric study of tortoises contained within our study population to ascertain whether home range location is consistent with membership in distinct tortoise social network communities. We found strong statistical support for the presence of nonrandom social networks that were derived from male-female mating relationships. The most parsimonious social network included two distinct “cliques” that were spatially segregated. Each clique contained a similar number of males and females. Understanding this basic aspect of tortoise behavior should be key in basic population biology, not only of turtles but also other reptiles. Our results should influence protocols for successful conservation of this keystone species.

The overlap between spatial and physiological ecology is generally understudied, yet both fields are fundamentally related in assessing how individuals balance limited resources. Herein, we quantified the relationships between spatial ecology using two parameters of home range (annual home range area and number of burrows used in one year) and four measures of physiology that integrate stress and immunity (baseline plasma corticosterone concentration [CORT], plasma lactate concentration, heterophil:lymphocyte ratio [H:L], and bactericidal ability [BA]) in a wild free-ranging population of the gopher tortoise (Gopherus polyphemus) to test the hypothesis that space-usage is correlated with physiological state. We also used structural equation modeling (SEM) to test for causative relationships between the spatial and physiological parameters. We predicted that larger home ranges would be negatively correlated to traditional biomarkers of stress and positively correlated with immunity, consistent with our hypothesis that home ranges are determined based on individual condition. Males had larger home ranges, used more burrows, and higher baseline CORT than females. We found significant negative correlations between lactate and home range (r = -0.456, df = 21, P = 0.029). CORT was negatively correlated with number of burrows used in both sexes (F = 7.322, df = 2,20, P = 0.003, Adjusted R² = 0.383). No correlations were observed between space use and BA or, notably, H:L. SEM models suggested that variation in number of burrows used was a result of variation in baseline corticosterone. The lack of a relationship between H:L and home range suggests that home range differences are not associated with differences in chronic stress, despite the pattern between baseline CORT and number of burrows used. Rather, this study indicates that animals balance tradeoffs in energetics, likely by way of baseline corticosteroid, in such a way as to maintain function across continuously variable home range strategies.

Large-scale Gopher Tortoise translocations are a widespread practice to remove tortoises from imminent dangers of habitat loss for human development. However, translocation for explicit conservation goals remains an infrequently-utilized and untested tool to conserve this at-risk species. Given the effects of density on tortoise movement and population viability, we conducted a conservation effort to increase the population density of a low-density tortoise population on a public property in southern Alabama. Efforts included consolidating the resident tortoises from a fire-suppressed landscape into a temporary enclosure located in a high-quality restored sandhill. We also headstarted two cohorts of hatchlings from the site for one and two years. In total, 100 resident tortoises were consolidated into the enclosure and 98 headstarted juvenile tortoises were released into the nearby area into both soft- and hard-release conditions. Using radiotelemtry of a sub-population of consolidated adults tracked for two seasons (before and after the enclosure was removed), we found that site fidelity of the adults was 69% in the area enclosed by the pen and 93% in the wider-managed site. Adult home range area was not significantly different between the years that adult tortoises were enclosed in the pen and the year after the pen was removed. A subpopulation of two-year old headstarts was also radiotracked. Radiotracked headstarts had 100% site fidelity, regardless of release condition, and home ranges significantly decreased over the study period. Mortality was extremely low in all groups. We further discuss how this research may contribute to effective strategies for the demographic management of low-density tortoise populations, particularly in the species’ distributional periphery.

Recent technological advances have allowed wildlife ecologists to begin to understand the complexity of relationships among individuals within populations. Given gopher tortoises have high site fidelity, naturally exist in high population densities, and are long lived, their population social structures may be among the most complex within Reptilia. Prior studies have demonstrated that gopher tortoises form non-random interactions with each other, described as “cliques” in a southern Georgia population of tortoises. Herein, we tested for the presence of this complex social structure at Boyd Hill Nature Preserve in Pinellas County, FL, using camera traps placed at 12 active tortoise burrows and set to record every 5 seconds for an entire year during daylight hours. We found that cliques are not present within this population when we only considered social interactions as co-occurrence of two individuals. However, when we only considered interactions to be positive, namely burrow sharing between same- or opposite-sex pairs and burrow chasing between opposite-sex pairs, we found that cliques were significantly present. This study has implications for basic understanding of sociality in turtles and also conservation efforts for this intensively-managed species. 

Camera traps are increasingly being used as a research tool in behavioral ecology to observe animals in the field. One challenge of continuous data collection by camera traps is that the amount of data produced can be overwhelming to process manually. Machine learning techniques show promise for becoming important tools to meet this challenge in a time- and cost-effective way. Over the past year, we collected images of over 950 social interactions among gopher tortoises from twelve active tortoise burrows at Boyd Hill Nature Preserve in St. Petersburg, Florida. The individual tortoises in each interaction must be identified to comprehensively analyze this data. To automate the re-identification step, we developed a machine learning algorithm to distinguish between tortoises. It takes a large amount of data to train the machine learning algorithm to recognize patterns. Therefore, we describe how our training dataset was collected and created. We also recount how we trained a Siamese neural network to identify individual tortoises based on their carapace markings in a manner similar to how facial recognition is used on humans. This research demonstrates that machine learning has the potential to be a powerful tool to aid data processing and analysis in behavioral ecology. 

Camera trap time-lapse recordings can collect vast amounts of data on wildlife in their natural settings. Transforming these data into information useful to ecologists is a major challenge. Machine learning techniques show promise for becoming important tools in the cost-effective analysis of camera trap data, but only if they become readily available to researchers without requiring advanced computing skills and resources. We present a new suite of software tools that reduce the amount of human effort needed to segment time-lapse, camera trap recordings in preparation for analysis. The tools incorporate a convolutional neural network trained to detect a focal species and to generate a draft video segmentation indicating the ranges of time when the focal species is present. We evaluated the utility of our neural network by comparing manual and automatic segmentations of 64 time-lapse recordings of gopher tortoise (Gopherus polyphemus) burrows, recorded in Pinellas County, Florida, USA between 25 November 2020 and 30 November 2020. The neural network correctly found 130 of the 145 segments containing tortoises (89.7%), whereas student graders found 135 segments (93.1%). A year of experience using the new software suite in an ongoing study of gopher tortoises deploying 12 camera traps indicates one person, assisted by machine learning algorithms, can segment a week's worth of time-lapse recordings-11.5 hours of standard-speed video-in under 3 hours. We concluded that the use of machine learning algorithms is practical and allows researchers to process large volumes of time-lapse data with minimal human effort.

A new open-source image processing pipeline for analyzing camera trap time-lapse recordings is described. This pipeline includes machine learning models to assist human-in-the-loop video segmentation and animal re-identification. We present some performance results and observations on the utility of this pipeline after using it in a year-long project studying the spatial ecology and social behavior of the gopher tortoise.