Insights into the thermal ecology, physiology, and behavior of a threatened ectothermic specialist from a warming and drying ecoregion Authors: Brian R Blais, Maria Vittoria Mazzamuto, John L Koprowski Published: 2025 ABSTRACT ABSTRACTIncreased heat and drought from Anthropogenic climate change will challenge the adaptive capacity of species, underscoring the need to understand thermal ecology – how organisms behaviorally and physiologically respond to temperature. We used noninvasive infrared thermography (IRT) to examine the thermal ecology of threatened narrow-headed gartersnakes (Thamnophis rufipunctatus) in a conservation breeding program at the Arizona Center of Nature Conservation/Phoenix Zoo. From 718 microhabitat and 124 individual measurements, hierarchical models identified extrinsic and intrinsic factors influencing microhabitat usage, body temperature (Tb), and behavior. Gartersnakes exhibited regional heterothermy, with tails cooler than head and trunk segments. The Tb of T. rufipunctatus was shaped by perch temperature, perch-air temperature, and whether snakes were visibly exposed or hidden. We documented microhabitat aggregations (≥2 gartersnakes) in ca. 40% of observations, which was best predicted by Tb. Thamnophis rufipunctatus appeared to favor cavity-bearing microhabitats, consistent with wild populations. This first application of IRT to snakes in semi-natural environments, and for T. rufipunctatus specifically, provides novel insights to guide more effective field surveillance and conservation management, while demonstrating the broader value of IRT and collaborative ex situ studies for wildlife conservation.Keywords: Body temperature, Conservation, Endangered species, Infrared thermography, Reptiles, Zoos INTRODUCTION Contemporary rates of rising temperature and prolonged drought under Anthropogenic climate change will challenge the pace at which species can adaptively respond (Bennett et al., 2021; Hof et al., 2011; Mitchell et al., 2018; Radchuk et al., 2019; Visser, 2008), which presents a major threat to Earth's biodiversity (Bellard et al., 2012; Pacifici et al., 2015; Pecl et al., 2017). The impacts of rapid onset changes on bio-available water, metabolism, and reproduction in amphibians and reptiles, for example, can be severe or irreversible (Bickford et al. 2010, Davis et al. 2015, Mitchell et al. 2016). The thermal tolerance and responses to environmental stimuli of a species first require an understanding of its thermal ecology – an organism's behavioral and physiological response to temperature (Bogert, 1959; Camacho et al., 2018; Refsnider et al., 2019). Endothermic animals, such as birds and mammals, generate their own heat through internal metabolism and need mechanisms to conserve it; to maintain body temperature (Tb), heat produced in the core is distributed to the periphery (Hill et al., 2016; Scholander et al., 1950). Ectothermic reptiles rely on an interplay of environmental stimuli (e.g. conduction, convection, radiation), metabolic physiology, morphology, and behavior to thermoregulate Tb to optimal levels to perform natural activities (Christian et al., 2016; Cowles and Bogert, 1944; Cox et al., 2022; Goulet et al., 2017; Kearney et al., 2009). Many reptiles select microhabitats (e.g. burrows/refuges, basking/perch structures) to thermoregulate as well as perform regular natural behaviors such as foraging, escaping predation, gestating, and aestivation/overwintering among others (Dubiner et al., 2024; Goulet et al., 2017; Harvey and Weatherhead, 2010; Sannolo et al., 2019; Scheffers et al., 2014). Ambient conditions alone may not capture thermal heterogeneity within and among the microhabitat landscape; i.e. microclimate (Campobello et al., 2017; Kearney et al., 2009; Scheffers et al., 2016; Woods et al., 2015). Infrared thermography (IRT; Tattersall, 2016; McCafferty et al., 2021) tools can noninvasively and simultaneously capture the thermal properties of focal species and the microhabitat surfaces being occupied, which can help infer site selection (Blais et al., 2023a; Goller et al., 2014; Krochmal and Bakken, 2003; Signore et al., 2020; Taylor et al., 2021). Ecological insights garnered from IRT include various anatomical, physiological, and behavioral adaptations to thermal environments (Mazzamuto et al., 2023; McCafferty et al., 2021; Mitchell and Clarke, 2019; Monge et al., 2025). Understanding environmental influences of favored microhabitats can guide more effective conservation actions, such as species reintroduction site planning or habitat restoration initiatives (Huey et al., 1989; Luna and Font, 2013). Assessing thermal physiology in wild systems allow for natural behaviors but complexities and stressors exist (e.g. nutrient requirements, threat of predation), whereas (simple) ex situ arenas offer stronger experimental control but could influence behavior or other aspects of thermal ecology, including diel/seasonal changes, sex, gravidity, and age-class, especially for at-risk species (Tattersall, 2016; Taylor et al., 2021). Ex situ (e.g. zoos, aquariums) managed semi-natural environments and enrichment strategies that facilitate innate behavior and needs of focal species may provide a compromise through informational feedback (Blais et al., 2023b; Burghardt, 2013; Chiszar et al., 1993; Reading et al., 2013). Zoo populations can present opportunities to closely monitor and gain insight into many natural parameters (e.g. behavior, life history traits, physiology) of a focal species that may otherwise be data deficient or onerous to track in situ (Blais et al., 2023b; Minteer et al., 2018; Pritchard et al., 2011; Spooner et al., 2023). The narrow-headed gartersnake, Thamnophis rufipunctatus (Cope, in Yarrow, 1875), is a piscivorous specialist endemic to montane, perennial, cool-water riparian systems (700–2430 m asl) in the Gila River Watershed of central Arizona and western New Mexico (Holycross et al., 2020; Rossman et al., 1996; Wood et al., 2018). Limited information exists on the thermal physiology of T. rufipunctatus (USFWS, 2014). Body temperature has been shown to strongly relate to air and surface temperature (Fleharty, 1967) but also water temperature (Hibbitts et al., 2009). This species is active in a wide range of air temperature (Ta, 11–32°C) and water temperature (Tw, 12–22°C; USFWS, 2014); active T. rufipunctatus can operate at cooler conditions than several congenerics (Rosen, 1991), including sympatric species (Fleharty, 1967). Range-wide population declines (Holycross et al., 2020; Wood et al., 2025a) culminated in the species being listed as threatened under the Endangered Species Act (USFWS, 2014; Wood et al., 2018). Projected loss of suitable environment across time in a warming and drying region presents additional challenges the species is likely to face (Blais and Koprowski, 2024). In response, the Arizona Center for Nature Conservation/Phoenix Zoo (hereafter ACNC) has managed an ex situ conservation breeding program for T. rufipunctatus since 2006 (Allard and Wells, 2018). The ex situ colony at ACNC has provided opportunities to study and learn about this enigmatic species, including insights into behavior and reproductive biology (Blais et al., 2023b; Wood et al., 2025b). Herein, we conducted a noninvasive IRT study on the ACNC colony to better understand the thermal ecology and physiology of T. rufipunctatus. Our objectives were to: 1) model environmental and temporal factors in snake-used versus available microhabitats in two complex naturalistic microcosms; 2) derive Tb and thermal physiology metrics; 3) analyze how behavior and environment influence Tb; and 4) explore relationships of intrinsic and extrinsic factors that may influence aggregation behavior. We expected gartersnakes to select microhabitat types similar to those used in the wild (e.g. refuges with cavities; Holycross et al., 2020; Rosen, 1991) and in relation to thermoregulatory needs. We expected Tb measures to be near or slightly above measurements recorded in the wild due to the controlled environment maintained in the ex situ location. We use ex situ ecological and physiological data to guide conservation of an imperiled ectotherm and inform broader thermal ecology research. Our methods incentivize IRT use in naturalistic settings and strengthen feedback loops and collaboration across the ex situ–in situ spectrum. Understanding how body temperature, behavior, and microclimate interact can reveal thermal preferences and adaptive challenges terrestrial ectotherms face in a warming, drying climate. RESULTS General results and temporal conditions We completed 36 total surveys (1–2 per enclosure per sampling date) between 24 May and 27 September 2019 (sampling date interval: 14±0.7 d); we ceased the project in conjunction with planned maintenance by ACNC in preparation for the overwinter brumation period. Morning shift surveys occurred at 10:12 h on average and afternoon shift surveys at 13:04 h on average (shift interval: 178.1±35.2 min); we omitted two afternoon surveys due to logistical restraints/husbandry priorities. The average survey duration was 13.9 (±5.8) min. Ambient conditions varied temporally between morning and afternoon survey shifts for Pb (Wilcoxon test: P<0.001; mean change=−2.2 mb), rH (P<0.001; mean change=−2.9%), Ta (P<0.001; mean change=2.7°C), and Tw (P<0.001; mean change=2.1°C; Fig. S2A,D). Ambient conditions also varied by month (Kruskal–Wallis: all P<0.001). Microhabitat selection by narrow-headed gartersnakes We generated a dataset of n=718 microhabitat assessments. The thermal heterogeneity of microhabitats varied within and among types (Ts: 29.8±4.55°C, Ts-Ta: 0.4±3.35°C; Table S1, Fig. S3A,B). Although we knew gartersnake total N per enclosure survey, we did not always observe all present individuals (detection rate: 60.1±29.7%) and made no attempt to search beyond surface visuals (i.e. beneath substrates, leafy plants, or underwater). Gartersnakes occupied microhabitat types as follows: subterranean (62.9%), cover (22.4%), water (10.3%), ground (2.3%), and plants (2.0%). Twenty-seven of 42 unique microhabitats remained unoccupied during study observation. Four models were within two corrected Akaike Information Criterion (AICc) units to explain microhabitat selection (Table S2). Only microhabitat type was identified as an important variable at the ≥0.80 threshold score and was the only factor in the top microhabitat usage model (AICc=312.6, weight=0.188, R2=0.71). After investigating the internal-only microhabitat dataset, there were seven models within two AICc units (Table S2). Again, type was important at ≥0.80 threshold although Ta and (internal) Ts had scores 0.40–0.60. The model that included those three terms yielded some trends (0.05