High activity costs of behavioral buffering under negative warming tolerance in desert Liolaemus lizards

Anthropogenic climate change poses significant threats to desert reptiles already living near their physiological limits. While behavioral thermoregulation can buffer thermal stress, its effectiveness and associated costs under extreme conditions remain poorly understood in Neotropical psammophilous species. We evaluated thermoregulatory effectiveness (E) and thermal vulnerability in two sister species of sand-dwelling lizards (Liolaemus scapularis and Liolaemus riojanus) from the Argentine Monte desert. We integrated field body temperatures, operative temperatures (Te), preferred temperatures (Tset), and critical thermal maxima (CTmax) across dry and humid periods over four years. Both species exhibited high thermoregulatory effectiveness despite extremely poor environmental thermal quality; only 3–11% of available Te fell within Tset range. This behavioral buffering acts as a ‘behavioral shield’—consistent with the Bogert effect—that effectively protects lizards from lethal temperatures but imposes a severe opportunity cost through activity-time restriction. Paradoxically, E was significantly higher during the thermally harsher dry period, a pattern facilitated in L. scapularis by marked seasonal plasticity in Tset, representing a coordinated physiological-behavioral mechanism absent in Liolaemus riojanus. Despite this success, warming tolerance was consistently negative, forcing lizards to remain inactive for 18–50% of potential activity time. This restriction was most acute in the endangered L. scapularis during the humid season, when shaded microhabitats fail to provide a sufficient cooling margin. Our findings demonstrate that while behavioral plasticity can homogenize thermal responses to stress, such buffering masks an underlying physiological vulnerability. Under projected mid-century warming for arid South America, further compression of activity windows may exceed behavioral buffering capacity, threatening the long-term persistence of these desert specialists. Proactive conservation must prioritize microhabitat heterogeneity to preserve thermoregulatory refugia for these endemic reptile communities.

Arid ecosystems represent some of the most thermally extreme environments on Earth, where the survival of ectothermic organisms depends on their ability to navigate landscapes that frequently exceed their physiological safety margins. In these systems, climate change is accelerating thermal stress, with projections indicating temperature increases of 2-4 °C in arid regions of South America by mid-century under moderate to high emission scenarios (IPCC, 2021). Desert-dwelling reptiles are particularly vulnerable because many populations already experience environmental temperatures approaching or exceeding their critical thermal limits, leaving minimal safety margins to buffer further warming (Deutsch et al., 2008; Huey et al., 2010; Kingsolver et al., 2013; Neel et al., 2021). While behavioral thermoregulation—through microhabitat selection, activity timing, and postural adjustments—can temporarily buffer thermal stress (Hertz et al., 1993; Kearney et al., 2009; Grimm-Seyfarth et al., 2017; Moore et al., 2018), this buffering capacity has limits. When operative temperatures routinely exceed physiological tolerance thresholds, behavioral strategies become insufficient (Deutsch et al., 2008), forcing activity restrictions that compromise energy acquisition, reproduction, and ultimately population persistence (Sinervo et al., 2010; Cunningham et al., 2020; Domínguez-Godoy et al., 2024). However, recent critiques highlight that thermal vulnerability indices focusing solely on physiological limits without integrating behavioral plasticity, microhabitat heterogeneity, and acclimation capacity may overestimate or underestimate actual extinction risk (Clusella-Trullas et al., 2021; Neel et al., 2021; Carilo Filho et al., 2022; Stark et al., 2024). This emphasizes the need for integrative studies quantifying both thermoregulatory effectiveness—the precision with which organisms maintain body temperatures near preferred values—and thermal safety margins under current and projected climatic conditions.

Sand-dwelling (psammophilous) lizards represent an understudied but ecologically significant group in desert ecosystems. These species possess specialized morphological and behavioral adaptations for life in loose sandy substrates, including sand-burial behavior that provides thermal refugia unavailable to rock-dwelling or open-habitat congeners (Halloy et al., 1998; Kohlsdorf et al., 2008; Liz et al., 2019; Chen et al., 2021; Bolton, 2022). The Monte desert of Argentina—a biogeographic region characterized by extreme thermal amplitudes, low precipitation (<200 mm annually), and sparse vegetation cover—harbors several endemic psammophilous lizard species of the Liolaemus wiegmannii group (Etheridge, 2000; Olave et al., 2014; Salva et al., 2021). Within this clade, Liolaemus scapularis and L. riojanus are sister species inhabiting geographically disjunct sandy habitats separated by ~700 km: L. scapularis occurs in the Calchaquí Valleys (Salta Province) where it is sympatric with the rock-dwelling L. quilmes (which occupies compacted substrates and rocky microsites, exhibiting spatial niche segregation), while L. riojanus inhabits Talampaya National Park (La Rioja Province) in syntopy with two larger congeners— L. olongasta (rock-dwelling) and the sand-dwelling L. cuyanus (Villamil et al., 2019). Field observations indicate that L. riojanus (mean SVL ~54 mm, Laspiur et al., 2006) is consistently subordinate in agonistic interactions with both larger species (L. olongasta ~65 mm, L. cuyanus ~78–84 mm) (Laspiur and Acosta, 2007; Gallardo and Cruz, 2023), suggesting potential competitive constraints on microhabitat access. Despite their recent common ancestry and shared psammophilous ecology, these species exhibit highly similar body sizes (L. scapularis: 53.2 ± 7.0 mm; L. riojanus: 51.9 ± 7.1 mm SVL), yet differ in congeneric species richness (low for L. scapularis vs. high for L. riojanus), and competitive context (dominant vs. subordinate), raising the question: do thermal physiologies converge due to phylogenetic constraint and shared sand-burrowing behavior, or diverge in response to local ecological pressures (Sannolo et al., 2014; Bennett et al., 2019)? Both species face significant conservation threats: L. scapularis is categorized as Endangered due to habitat loss from urbanization, agriculture, and mining (Abdala, 2016b), while L. riojanus is listed as Least Concern but experiences local population fragmentation from vineyard expansion (Abdala, 2016a).

Here we evaluate thermoregulatory effectiveness and thermal vulnerability in these two sister species by integrating field body temperatures (Tb), environmental operative temperatures (Te), laboratory-determined preferred temperatures (Tpref), and critical thermal maxima (CTmax) across dry and humid climatic periods. Our general objective is to assess thermoregulatory effectiveness and warming tolerance in two sister species of sand-dwelling lizards that differ in community context, and competitive interactions, to determine their relative vulnerability to climate change in the Argentine Monte. Specifically, we aim to: (1) characterize the thermal biology (Tb, Te, Tpref, and CTmax) and thermoregulatory effectiveness (E = 1 - db/de, where db and de represent deviations of body temperature and environmental temperature from preferred temperature, respectively) of Liolaemus scapularis and L. riojanus; (2) evaluate the influence of seasonal environmental variation (dry vs. humid periods) on these traits; and (3) estimate their warming tolerance (Wt = CTmax - Te max) and potential activity restriction under current climate scenarios. While we initially considered the potential role of body size, its restricted variance in our sample led us to treat SVL primarily as a covariate in our statistical models rather than a central objective of the study. By comparing these sister species across different seasons, we aim to determine if behavioral and physiological traits are flexible enough to buffer the extreme thermal pressures of the Monte desert.

Study species and sites

Fieldwork was conducted during the austral spring-summer activity season (October–March, 2018–2022) at two sites in the Monte ecoregion of Argentina (Cabrera and Willink, 1980, Figure 1). Data for Liolaemus riojanus was collected across six field trips at Parque Nacional Talampaya (29°45’S, 68°00’W, La Rioja Province). The study area comprises a 45-ha plot of well-preserved low dune fields with reddish aeolian sand (up to 1.5 m high) and sparse vegetation cover (<30%) dominated by xerophilous and psammophilous shrubs (Bulnesia, Acacia, Atriplex, Larrea; Burkart et al., 1999). Data for L. scapularis was collected across four field trips at Reserva Natural Manejada Quebrada de Cafayate (26°02’S, 65°53’W, Salta Province), approximately 460 km north of Talampaya. This site consists of a 20-ha plot of whitish aeolian sand dunes (up to 2.5 m high) with ~50% vegetation cover comprising Sporobolus grasslands and mixed shrub steppes (Atriplex, Heliotropium, Gomphrena, Suaeda, Prosopis; Mares et al., 1985). Both sites experience extreme thermal fluctuations and low annual precipitation (~200 mm, concentrated October–March). Mean monthly temperatures range from -6.7 to 43.2 °C at Talampaya and -9.4 to 36.8 °C at Cafayate, with ground-level temperatures reaching 60 °C (Hueck, 1950; NASA LaRC POWER 1981–2020).

Field body temperatures and operative temperatures

We recorded body temperatures (Tb) from 376 active adult lizards (Liolaemus scapularis: n = 235; L. riojanus: n = 141) during activity periods (0800–2000 h, GMT-3) through active visual searches. Upon first sighting, individuals were captured by lassoing and Tb was measured immediately (within 20 s) using a quick-reading digital thermometer (Reed LM-8000, resolution 0.1 °C) with a type-K thermocouple (± 0.1 °C, 1.62 mm diameter) inserted 1 cm into the cloaca. Individuals were handled by the head to minimize heat transfer (Martori et al., 2002). Sex was determined based on sexually dimorphic characters (Cei, 1993; Etheridge, 2000), and snout-vent length (SVL) was measured to confirm adult status following species-specific maturity thresholds (L. scapularis: Ramirez Pinilla, 1994; L. riojanus: Laspiur et al., 2006). Individuals were marked dorsally with non-toxic enamel to prevent recapture and released at the capture site.

To characterize the thermal environment, we used copper models (length: 60 mm; diameter: 15 mm) to record operative temperatures (Te). Operative temperatures (Te), representing thermal environments available to non-thermoregulating ectotherms (Bakken, 1992), were recorded at 5-minute intervals using ten temperature dataloggers (Maxim DS1921G Thermochron iButton ®, ± 0.1 °C). Dataloggers were housed in polyvinyl chloride (PVC) tube models matching species’ body dimensions, with 1.5 cm silicone segments at each end and painted gray. Following a stratified sampling design, these models were placed in different microhabitats (full sun, filtered shade, and deep shade) during the dry (late September to early October) and humid (February to early March) periods to capture the full thermal mosaic available for behavioral buffering, regardless of the total area covered by each microhabitat type. In each period, Te models were deployed for 10–15 consecutive days, recording data 24 h/day to capture the full range of daily thermal fluctuations. This sampling effort ensured that our estimates of Te max reflected the chronic thermal landscape and characteristic midday heat peaks typical of the region, rather than isolated or atypical weather events. Models were calibrated against live lizards prior to deployment (Dzialowski, 2005) and placed in microhabitats where lizards were observed active. Dataloggers recorded temperatures synchronously with field body temperature sampling (0800–2000 h). We collected 8,405 Te records (L. scapularis: 3,050 from 4 field trips; Liolaemus riojanus: 5,355 from 6 field trips). However, this dataset was filtered to include only the diurnal activity window of the species (08:00–20:00 h). To ensure a balanced representation of the thermal landscape and avoid bias arising from the non-proportional placement of sensors relative to microhabitat availability, records were averaged by exposure category (full sun, filtered, and full shade) for each time step. This filtering and balancing process resulted in 3,927 standardized thermal records used for all subsequent analyses. Mean and maximum operative temperatures (Te and Te max) were calculated for each species and period. Standard errors (SE) for Te max were estimated using a non-parametric bootstrap (1,000 iterations) to account for uncertainty in extreme values, while SE for Te followed standard analytical procedures. All metrics are reported as mean ± SE.

Preferred temperatures

We determined preferred temperatures (Tpref) for 65 individuals (Liolaemus riojanus: n = 28, 12 dry + 16 humid period; L. scapularis: n = 37, 15 dry + 22 humid period) using different individuals for each climatic period. Lizards were transported to the laboratory and housed individually in terraria (35×20×20 cm) with water ad libitum and photoperiod matching field daylight hours for 7 days prior to trials. During this period, room temperature was maintained constant at 25 ± 2 °C to standardize physiological baselines. Food was withheld 24 hours before experiments.

Preferred temperatures were assessed using a wooden thermal gradient (1.4×0.6×0.3 m) divided into four 0.15 m lanes by opaque panels, with a 0.05 m layer of fine sand covering the bottom. The cold end (20 °C) was maintained using ambient air conditioning (18 °C, HITACHI HS3000FC) reinforced with cooling gel compresses. The hot end (60 °C) was generated by four 150 W incandescent bulbs suspended at varying heights (0.2–0.4 m) above the substrate. After 1 h acclimation, body temperatures were recorded at 30-min intervals for 4 h using a quick-reading digital thermometer (Reed LM-8000) with type-K thermocouple inserted 1 cm into the cloaca. All trials were conducted during species-specific field activity periods. Preferred temperature (Tpref) was calculated as the mean of all recorded temperatures per individual, and the thermal set-point range (Tset) was defined as the interquartile range (25th–75th percentiles) encompassing 50% of observations (Hertz et al., 1993).

Critical thermal maximum

We determined critical thermal maxima (CTmax) for a subsample of individuals (Liolaemus scapularis: n = 11; L. riojanus: n = 26) at least 24 h after preferred temperature trials. Each lizard was placed in a plastic container beneath a 100 W incandescent lamp (PHILIPS Classictone) suspended 40 cm above. Individuals were heated at 0.8 °C/min and monitored continuously for behavioral changes. Lizard body temperature (Tb) was checked every 2 min via an infrared thermometer (Testo 830-T1, Lenzkirch, Germany) until reaching 40 °C (approximate annual maximum temperature at study sites; NASA LaRC POWER 1981–2020), then every 30 s until loss of righting response. This initial monitoring approach was implemented to minimize handling stress and prevent potential cloacal injury from repeated probing, both of which could prematurely exhaust the animal and bias the physiological endpoint. CTmax was recorded when individuals exhibited panting, loss of coordinated locomotion, and muscle spasms (Pough and Gans, 1982; Cruz et al., 2005; Kubisch et al., 2015). The righting response—the ability to correct posture when placed supine—was used as the primary endpoint criterion, which may be preceded by evident locomotor deterioration (Kaufmann and Bennett, 1989; Brown, 1996; Labra et al., 2001). Final record was made using a quick-reading digital thermometer (Reed LM-8000) with type-K thermocouple inserted 1 cm into the cloaca.

Thermoregulatory metrics

We calculated thermoregulatory precision (db) as the absolute deviation of each field body temperature (Tb) from the species’ thermal set-point range (Tset). When Tb fell within Tset, db = 0; otherwise, db = | Tb - nearest Tset boundary|. Environmental thermal quality (de) was calculated as the absolute deviation of each operative temperature (Te) from Tset during lizard activity periods. High de values indicate thermally challenging environments, while values near zero represent optimal thermal conditions (Hertz et al., 1993). Thermal quality of the environment was also expressed as the percentage of Te records falling within Tset. Thermoregulatory effectiveness (E) quantifies the degree to which organisms reduce thermal deviations relative to environmental availability, calculated as E = 1 - (db mean / de mean), where db mean and de mean represent the mean values of all precision and quality records, respectively, for each species and period. Values approaching 1 indicate precise thermoregulation, values near 0.5 suggest moderate thermoregulation, and values near 0 indicate thermoconformity (Hertz et al., 1993).

Warming tolerance (Wt) was calculated as CTmax - Te max for each field trip, representing the thermal buffer between maximum environmental temperatures and physiological limits. Negative Wt values indicate that environmental maxima exceed critical thermal thresholds (Deutsch et al., 2008). Temporal thermal restriction was quantified as the percentage of potential activity time (0800–2000 h) when Te exceeded the upper limit of Tset, indicating periods when lizards must seek thermal refugia to avoid overheating.

Statistical analysis

We used Linear Mixed Models (LMM) to analyze field-derived thermal traits (Tb and E), with species and climatic period (dry vs. humid) as fixed effects and year as a random effect to account for interannual variability across the 2018–2022 sampling period. Laboratory-measured traits (Tpref and CTmax) were analyzed using Two-way ANOVAs (Species × Period), as these measurements were conducted during a single sampling event per species and period, precluding the inclusion of ‘Year’ as a random factor. Effects of body size (SVL) on Tb were evaluated using ANCOVA with species and period as cofactors. Relationships between SVL and CTmax were assessed using Pearson correlations within each species and a linear model testing SVL × species interaction across the full dataset. Warming tolerance (Wt) was compared between species and periods using independent t-tests.

Non-parametric Mann-Whitney U tests were applied to compare thermoregulatory precision (db) and environmental thermal deviation (de) between species and periods due to non-normal distributions confirmed by Kolmogorov-Smirnov tests. Thermal quality (percentage of Te within Tset) was compared descriptively. Post-hoc comparisons for significant interactions used Tukey HSD tests. Statistical significance was set at α = 0.05. All analyses were conducted in R v.4.3.0 (R Core Team, 2023) using packages lme4 (Bates et al., 2015), lmerTest (Kuznetsova et al., 2017), car (Fox & Weisberg 2019), emmeans (Lenth, 2023), and ggplot2 (Wickham, 2016).

Thermal environment characteristics

After the standardized filtering and microhabitat-balancing process described in the methods, a total of 3, 927 Te records were analyzed across 10 field trips during lizard activity periods (Liolaemus scapularis: 1,321 records from 4 field trips, 2018-2020; L. riojanus: 2,607 records from 6 field trips, 2018-2022, Table 1; Figure 2). Operative temperatures did not differ significantly between species (Linear Mixed Model, F 1,7.03 = 0.16, p = 0.705) or between dry and humid periods (F 1,7.02 = 0.18, p = 0.683). Liolaemus scapularis experienced a narrower thermal range (8-63.9 °C) with lower mean temperatures (30.0 ± 0.3 °C) than L. riojanus (5.6-68.2 °C, mean: 31.8 ± 0.2 °C), though this difference was not statistically significant due to high thermal variability between field trips (Table 1; Figure 2).

Despite similar mean temperatures, thermal deviation from preferred temperatures (de = | Te - Tset |) was significantly higher during dry periods in both species (Mann-Whitney, Liolaemus scapularis: W = 224, 950, p = 0.007; L. riojanus: W = 925, 559, p < 0.001; Figure 3A). Median de values were 8.6 °C (dry) vs 6.9 °C (humid) in L. scapularis and 9.1 °C (dry) vs 7.0 °C (humid) in L. riojanus. Environmental thermal quality was consistently poor, with only 3.2-10.8% of available operative temperatures falling within preferred temperature ranges. L. scapularis showed marked seasonal variation in thermal quality (dry: 3.2%, humid: 10.8%), while L. riojanus exhibited consistently low quality across both periods (6.1-6.6%).

Body temperature and thermoregulatory precision

We measured body temperatures (Tb) of 376 individuals (Liolaemus scapularis: 235; L. riojanus: 141) with balanced sex ratios (Table 2). The Linear Mixed Model (LMM) revealed a significant interaction between species and period (LMM, F 1, 185.4 = 6.79, p = 0.009) indicating that seasonal shifts in Tb differed between species. While L. riojanus increased its Tb during the dry period (36.8 ± 3.2 °C) compared to the humid period (35.8 ± 4.2 °C), L. scapularis maintained more stable body temperatures (34.4 ± 3.3 °C dry vs. 34.9 ± 3.9 °C humid; Figure 2). The random effect of ‘Year’ accounted for a variance of 2.18 (SD = 1.47 °C), justifying its inclusion to account for interannual environmental variation.

Preferred temperatures determined from thermal gradient experiments (Liolaemus scapularis: n = 37; L. riojanus: n = 28) showed a significant species by period interaction (ANOVA, F 1,61 = 17.38, p <0.001; Table 2). Liolaemus scapularis exhibited marked seasonal plasticity, selecting higher temperatures in dry period (35.6 ± 1.1 °C) than humid period (34.0 ± 1.0 °C; post-hoc p <0.001), a difference of 1.6 °C. In contrast, L. riojanus showed relatively stable preferences across periods (dry: 35.1 ± 1.0 °C; humid: 35.8 ± 1.0 °C; p = 0.39). Neither sex (ANOVA, L. scapularis p = 0.451, L. riojanus p = 0.277) nor body size (correlations, L. scapularis r = -0.12, p = 0.49; L. riojanus r = 0.09, p = 0.65) significantly affected Tpref within species.

Critical thermal maxima differed significantly between species during the humid period (t = 6.16, df = 19.3, p < 0.001), with Liolaemus riojanus exhibiting higher heat tolerance (45.6 ± 0.9 °C) than L. scapularis (43.8 ± 0.5 °C). For L. riojanus, CTmax remained stable across seasons (p = 0.076). Notably, the maximum body temperatures observed in the field for L. scapularis (44.0 °C) exceeded its average CTmax, highlighting extreme thermal vulnerability. Regarding the influence of body size, SVL did not significantly correlate with CTmax in either L. scapularis or L. riojanus (all p > 0.05). Furthermore, when included as a covariate in the analysis, SVL had no significant effect on critical thermal limits, confirming that the observed differences in CTmax are driven by species identity and seasonality rather than body size.

Body temperature deviation from Tset (db = | Tb - Tset |) was was analyzed using LMM to account for interannual variance (Random effect variance = 0.064). Precision (db) remained consistent across species (F 1, 37.2 = 1.23, p = 0.274) and periods (F 1, 12.9 = 2.59, p = 0.131), with median values of 0.81–1.67 °C (Figure 3A). Approximately 20–25% of individuals maintained Tb within their Tset (Table 2), representing a 2.3–7.8-fold improvement over environmental thermal quality (Figure 2).

Body size (SVL) showed no significant effect on body temperature (ANCOVA, F 1,373 = 0.75, p = 0.454) nor on any other thermal parameter (p > 0.10 in all cases) and was thus excluded from further behavioral analyses. Mean SVL differed by only 1.3 mm between species (L. scapularis: 53.2 ± 7.0 mm; Liolaemus riojanus: 51.9 ± 7.1 mm).

Thermoregulatory effectiveness

Both species exhibited high thermoregulatory effectiveness (E>0.75) calculated from 376 individuals (E = 1 - db / de; Table 3). The LMM (with ‘Year’ as a random effect) showed that mean E did not differ significantly between species (F 1, 22.8 = 0.86, p = 0.364; Liolaemus scapularis: 0.789 ± 0.23; L. riojanus: 0.766 ± 0.25). However, E was significantly higher during dry periods (0.82 ± 0.20, n = 161) than humid periods (0.747 ± 0.26, n = 215; F 1, 7.7 = 7.54, p = 0.026; Table 3, Figure 3B). This seasonal pattern was consistent in both species with no significant interaction (p = 0.411): L. scapularis showed E of 0.829 ± 0.20 (dry) vs 0.760 ± 0.24 (humid), while L. riojanus showed 0.816 ± 0.21 (dry) vs 0.721 ± 0.26 (humid; Table 3, Figure 3B). The random effect of ‘Year’ showed negligible variance (< 0.001), indicating that the observed increase in effectiveness during the dry season is a robust physiological response across years.

Thermal vulnerability

Thermal restriction analysis revealed that lizards were forced to seek refuge 18-50% of potential activity time when operative temperatures exceeded the upper limit of their preferred range (Te > Tset max; Table 4, Figure 4). Restriction time varied by species and period: L. scapularis experienced 18.3% restriction in dry period but 50.0% in humid period, while L. riojanus showed more consistent restriction across periods (31.8% dry, 38.5% humid; Table 4; Figure 5A).

Warming tolerance (Wt = CTmax - Te max) calculated using species-specific critical limits, was negative across all field trips, ranging from -6.7 to -22.9 °C (mean: -14.7 °C; Table 4, Figure 5B). These negative values represent a theoretical environmental potential for overheating, indicating that maximum operative temperatures routinely exceeded critical thermal limits. Despite the significant difference in CTmax between species, both faced extreme environmental heat loads that resulted in deeply negative Wt values, with no significant differences in these potential safety margins between species (t-test, t 7.98 = -0.11, p = 0.914; Liolaemus scapularis: -14.5 ± 5.1 °C; L.| riojanus: -14.9 ± 6.8 °C) or periods (t 7.67 = -0.01, p = 0.992; dry: -14.7 ± 4.6 °C; humid: -14.7 ± 6.3 °C) as environmental heat overwhelmed physiological differences.

High thermoregulatory effectiveness and the resilience of psammophilous lizards

Both Liolaemus species exhibited remarkably high thermoregulatory effectiveness (E > 0.75), a result that is particularly striking given the chronically poor thermal quality of the Monte desert. In this environment, only 3–11% of available operative temperatures (Te) fall within the preferred thermal range (Tset), meaning that an organism moving randomly would almost never reach its physiological optimum. The 2.3–7.8-fold improvement in thermal quality achieved by active individuals (with 20–25% of Tb records within Tset) highlights the critical role of behavioral precision (Hertz et al., 1993).

These effectiveness values place these sand-dwelling species among the most efficient thermoregulators within the Liolaemus genus (≈0.6–0.9; Labraga and Villalba, 2009; Rodríguez-Serrano et al., 2009; Moreno Azócar et al., 2013; Valdecantos et al., 2013; Artacho et al., 2017; Gómez Alés et al., 2017). Unlike highland congeners that often face “cold stress” and exhibit lower E due to the high metabolic cost of searching for rare heat sources (Gutiérrez et al., 2010; Miranda-Calle et al., 2021; Montoya-Cruz et al., 2024), these desert species face “heat stress.” Here, the high E values suggest that the lizards are not merely “shuttling” to get warm, but are actively managing a dangerous heat load through sophisticated microhabitat selection. Studies in congeners demonstrate that exploitation of vegetation-generated thermal mosaics—transitions between sun-exposed patches, filtered light under shrubs, and shaded microsites—enables Liolaemus to maintain body temperatures near optima in thermally heterogeneous environments (Block et al., 2013; Stellatelli et al., 2013, 2020; Vivas et al., 2019; Brizio et al., 2021, 2025; Laspiur et al., 2021; Salva et al., 2021; Méndez Osorio et al., 2024). In our study sites, sparse to moderate vegetation cover (30-50%) dominated by xerophilous shrubs (Larrea, Atriplex, Prosopis) creates spatially heterogeneous thermal landscapes where lizards can behaviorally track preferred temperatures. The specialized sand-burrowing behavior documented in the Liolaemus wiegmannii group (Halloy et al., 1998) may provide additional refuge during extreme thermal events, though shallow burial depth (typically 2–3 cm based on field observations) likely provides limited thermal insulation and may function primarily as an antipredatory response (Bujes and Verrastro, 2008; Block et al., 2013; Maia-Carneiro and Rocha, 2013; Stellatelli et al., 2018a; Dematteis et al., 2022).

The counterintuitive pattern of stress-induced precision

A significant finding of this study is that thermoregulatory effectiveness was higher during the dry period (E = 0.82) than the humid period (E = 0.75, p = 0.027), despite the environment being objectively more thermally hostile (de was higher). This “counterintuitive” pattern supports cost-benefit models of thermoregulation (Huey and Slatkin, 1976; Vickers et al., 2011; Sears and Angilletta, 2015; Chan et al., 2024). When environmental quality is extremely low, the “cost” of thermoconforming (i.e., allowing body temperature to track the environment) increases exponentially because it leads to temperatures near or exceeding CTmax (Kearney et al., 2009; Buckley et al., 2015). Consequently, lizards are “forced” to be more precise. The temporal restriction analysis provides a mechanistic explanation: Liolaemus scapularis experienced 50% activity time restriction during the humid period compared to only 18% during the dry period. This suggest that when temporal windows for activity compress, the availability of thermally suitable microhabitats becomes the primary constraint on thermoregulatory precision (Lara-Reséndiz et al., 2022; Doucette et al., 2023).

The seasonal Tpref plasticity observed in Liolaemus scapularis (a 1.6 °C increase during the dry season, p <0.001) provides a physiological mechanism for this shift. By elevating their thermal set-point during the hottest months, these lizards effectively reduce the “thermal gap” between their internal requirements and environmental reality (Gvoždík, 2012; Stellatelli et al., 2018b; Domínguez-Guerrero et al., 2019; Liz et al., 2019; Gómez Alés et al., 2022; Giacometti et al., 2023). This 1.6 °C shift is substantial given that Tpref is generally considered phylogenetically conserved and exhibits limited intraspecific variation in Liolaemus (Labra et al., 2009; Medina et al., 2012; Moreno Azócar et al., 2013; Ibargüengoytía et al., 2021). This “labile” strategy contrasts with the “static” strategy of Liolaemus riojanus, which showed no such plasticity (0.7 °C difference, p = 0.39). This divergence suggests that even within sister species, the capacity to adjust physiological set-points to environmental predictability can vary, potentially reflecting different evolutionary histories of exposure to thermal variability (Logan et al., 2018; Ryan and Gunderson, 2020; Laspiur et al., 2024). This species-specific difference in thermal acclimation capacity has important implications for predicting vulnerability trajectories under climate change. While both species currently exhibit equivalent overall effectiveness, L. scapularis ’ ability to adjust Tpref may confer greater resilience to progressive warming by continuously “tracking” shifting thermal environments, whereas the static strategy of L. riojanus could lead to progressively widening mismatches between physiology and environment under novel thermal regimes (Morley et al., 2019; Carilo Filho et al., 2022). Testing this hypothesis requires long-term monitoring across multiple warming cycles.

Behavioral buffering and thermal vulnerability: the bogert effect

The most alarming result for the long-term persistence of these species is the consistently negative warming tolerance (Wt), which averaged -14.7 °C across all field trips. While these negative values represent a theoretical environmental potential for overheating, they indicate that maximum environmental heat loads routinely surpass the physiological “ceiling” of these lizards (Deutsch et al., 2008; Sunday et al., 2014), aligning with predictions that desert ectotherms already experience environmental temperatures near or exceeding critical thermal thresholds (Lara-Resendiz et al., 2014; Doucette et al., 2023; Dufour et al., 2024). However, the discrepancy between negative Wt (based on environmental potential) and the absence of field body temperatures exceeding CTmax reveals the critical role of behavioral thermoregulation in current survival. Despite Liolaemus riojanus having a significantly higher CTmax (1.8 °C higher than L. scapularis during the humid period), this physiological advantage is often insufficient to provide a positive safety margin under extreme desert heat, as the environmental potential remains overwhelmingly higher than their thermal limits.

This decoupling of heat tolerance and vulnerability exemplifies the Bogert effect (Huey et al., 2003; Sunday et al., 2014; Fey et al., 2019; Muñoz, 2022). Efficient behavioral thermoregulation acts as a “buffer” that shields organisms from environmental selective pressures. Because these lizards are effective at exploiting thermal heterogeneity, their CTmax is rarely “tested” by natural selection, potentially leading to evolutionary stasis in thermal limits (Sunday et al., 2014; Buckley et al., 2015; Gunderson and Stillman, 2015; Muñoz and Bodensteiner, 2019).The similarity in thermoregulatory capacity and thermal vulnerability between sister species despite differences in physiological thresholds (CTmax) and competitive context suggests that behavioral flexibility can homogenize responses to thermal stress across species facing similar thermal challenges. However, in the context of rapid anthropogenic climate change, this behavioral shield becomes problematic: environments are warming faster than CTmax can evolve, making the behavioral response the primary, yet increasingly constrained, line of defense.

The apparent contradiction between effective behavioral buffering (field Tb < CTmax) and negative warming tolerance (Te max > CTmax) highlights that while the lizards are biologically successful at avoiding lethal heat (as evidenced by the midday activity drop in Figure 4), they face a severe “opportunity cost”. This negative Wt should be interpreted as quantitative measure of the environmental pressure that forces lizards into extensive periods of inactivity, effectively trading potential foraging and mating time for thermal safety. In our study area, maximal Te reached values far exceeding physiological limits (up to 58.0 °C), creating a thermal landscape where safe activity is only possible through intense buffering. For Liolaemus scapularis, whose CTmax is significantly lower than its sister species, this margin is virtually non-existent, as maximum field Tb can reach or even exceed its absolute lethal threshold. Therefore, buffering is a vital survival mechanism, but it represents a long-term ecological constraint by limiting the species’ activity window and overall fitness in a warming world.

Thermal window narrowing: the hidden energetic cost

While behavioral thermoregulation enables survival under negative Wt conditions, it imposes substantial opportunity costs through activity time restriction. Both species must remain inactive 18-50% of potential activity time when operative temperatures exceed thermal tolerance limits, directly constraining foraging opportunities, limiting mate-searching windows, and potentially elevating predation risk when multiple individuals synchronize emergence during brief thermally suitable periods (Huey and Slatkin, 1976; Blouin-Demers and Nadeau, 2005; Doucette et al., 2023). We documented acute “thermal window narrowing,” particularly in Liolaemus scapularis during the humid period, where the species faced 50% activity restriction. This severe restriction is not merely a consequence of higher absolute heat peaks, but rather a result of ‘thermal homogenization’ in the landscape. During the humid season, shaded microhabitats (thermal sinks) degrade in quality as their operative temperatures (Te shade) remain high for longer durations, frequently approaching the species’ CTmax. Consequently, lizards are forced into inactivity not because the sun is hotter, but because the shade is no longer ‘cool enough’ to provide a safe physiological retreat. This means that for half of the daylight hours, lizards are forced into refugia, unable to forage, defend territories, or seek mates.

This 50% loss of activity time represents not merely a behavioral shift but a potential demographic threat. When activity restriction exceeds critical thresholds (typically 60-70% of potential activity time), populations may become energetically unsustainable as individuals can no longer meet daily metabolic requirements, leading to reproductive failure (Sinervo et al., 2010, 2018; Doucette et al., 2023). The fact that Liolaemus scapularis already experiences 50% restriction during the humid season suggests this endangered species is approaching an energetic tipping point. In Liolaemus riojanus populations syntopic with larger congeners (L. olongasta, L. cuyanus), thermal restriction may compound competitive disadvantages by reducing temporal access to shared resources, though quantifying this interaction requires experimental verification (Laspiur et al., 2024; Bonavita et al., 2025).

Under projected climate warming of 2-4 °C for arid South America by mid-century (IPCC, 2021), further increases in Te max would compress viable activity windows beyond current levels (Labraga and Villalba, 2009; Blanco and Doyle, 2024). Given that maximum operative temperatures already exceed CTmax (negative Wt), additional warming will not merely make environments “less comfortable” but will directly expand the temporal window during which surface conditions are physiologically lethal, forcing increased refuge use and further constraining energy acquisition.

Conservation strategies: beyond habitat area to thermal quality

Our results demonstrate that conservation status based on habitat range alone can be deceptive when thermal vulnerability is considered. Both species face equivalent thermal risks despite their different IUCN classifications (L. scapularis Endangered, Liolaemus riojanus Least Concern). This equivalence in vulnerability between sister species carries important conservation implications: protecting one species without conserving the other would overlook shared thermal risks that transcend current threat assessments based primarily on habitat degradation rates. The preservation of microhabitat heterogeneity is the primary viable mitigation strategy. The shrubs that provide shade— Larrea, Atriplex, Prosopis, and other xerophilous species—are not merely “habitat components”; they function as biological “coolers” that generate thermal mosaics allowing lizards to persist despite Wt approaching -20 °C in some field trips. While these extreme negative margins represent a theoretical environmental potential for overheating rather than actual physiological exposure—as lizards successfully exploit micro-heterogeneity to maintain Tb within safe limits—they quantify the extreme magnitude of the thermal barrier in this system. Such values are consistent with other high-temperature deserts, such as the Chihuahuan Desert (Wt = -17.6 °C; Gadsden et al., 2020) and the Sonoran Desert (margins exceeding -25 °C; Lara-Resendiz et al., 2014).

Conservation priorities should focus on: (1) protecting vegetation structure that maintains thermal heterogeneity, as habitat degradation reducing vegetation cover (overgrazing, land clearing for vineyards, urban expansion) will eliminate thermal refugia and immediately expose lizards to lethal operative temperatures (Scheffers et al., 2014; Arenas-Moreno et al., 2021; Tucker et al., 2023; Brizio et al., 2025); (2) maintaining habitat connectivity to facilitate thermal tracking as surface temperatures increase (Gonçalves et al., 2016; McGuire et al., 2016; Keeley et al., 2018; Pellerin et al., 2022); and (3) monitoring population demography during extreme heat events to detect early warning signals of thermal stress approaching critical thresholds (Tinkle and Ballinger, 1972; Vargas-García et al., 2019; Clifton et al., 2023; Templeton et al., 2023).

Sister species as natural experiments in thermal ecology

The comparative approach using sister species differing in specific ecological axes (community structure, competitive context) while sharing evolutionary history and specialized morphology provides a natural experiment to disentangle phylogenetic constraints from ecological determinants of thermal vulnerability (Warren et al., 2008). Our finding that both species exhibit equivalent overall vulnerability (E, Wt, activity restriction) despite physiological divergence (CTmax difference, Tpref plasticity presence/absence) demonstrates that behavioral thermoregulation can homogenize thermal responses across species facing similar absolute thermal challenges, exemplifying the Bogert effect (Huey et al., 2003). However, the subtle differences we detected—seasonal Tpref adjustment in the competitively dominant species versus static physiology in the subordinate species, and contrasting temporal restriction patterns potentially linked to competitive asymmetries—suggest that microevolutionary divergence in thermal strategies may be underway but currently masked by the overwhelming effect of extreme environmental conditions that push both species to their physiological limits (Labra et al., 2009; Moreno Azócar et al., 2013; Araya-Donoso et al., 2021). Under progressive climate warming, these nascent differences in acclimation capacity and competitive interactions may amplify, leading to divergent vulnerability trajectories despite current equivalence (Pontes-da-Silva et al., 2018). Thus, studying both species is essential not only to characterize present-day thermal ecology but to identify which populations harbor adaptive potential for future thermal regimes, informing conservation prioritization under climate uncertainty.

Limitations and future directions

A key limitation of this study is the “active lizard bias”—our Tb measurements represent only individuals that emerged and were detectable. The temperatures of buried or completely inactive lizards during extreme thermal conditions remain unknown, which might lead to overestimation of E during the most extreme hours. Additionally, we did not quantify specific microhabitat use concurrent with body temperature measurements, limiting our ability to identify which behavioral mechanisms (shade exploitation, subsurface retreat, postural adjustments) contribute most to the observed thermal buffering. Future research should: (1) quantify thermal performance curves to calculate safety margins relative to optimal performance temperatures rather than survival limits (Angilletta et al., 2002; Rezende and Bozinovic, 2019), as Wt based on CTmax may underestimate vulnerability if sublethal temperatures impair fitness-related performance; (2) evaluate seasonal plasticity in CTmax to determine if thermal tolerance adjusts alongside Tpref (Stellatelli et al., 2018b; Domínguez-Guerrero et al., 2021; Giacometti et al., 2023); (3) integrate physiological data with mechanistic species distribution models to project range shifts under climate scenarios (Kearney et al., 2009, 2010); and (4) utilize implantable dataloggers to track body temperatures continuously and characterize thermal conditions experienced by inactive individuals (Virens and Cree, 2019).

Our study demonstrates that while Liolaemus riojanus and L. scapularis are efficient thermoregulators, their survival in the Monte desert hinges on significant ecological trade-offs. The negative warming tolerance (Wt) reported here represents a theoretical environmental pressure rather than immediate mortality; it reflects a thermal landscape where safe activity is only possible through intense behavioral buffering.

This success carries a high “opportunity cost”, as lizards must trade essential foraging and mating time for thermal safety. This constraint is particularly severe for Liolaemus scapularis, which possesses a lower physiological ceiling and operates much closer to its lethal limits. Ultimately, while the “Bogert effect” may provide a behavioral shield against lethal heat, the resulting chronic activity restriction under a warming climate poses a serious long-term threat to the fitness and persistence of these desert populations.