Lizard thermoregulation revisited after two decades of global warming

Read the free Plain Language Summary for this article on Journal blog. Temperature is one of most relevant environmental factors that control biological processes and determine distribution all life-forms (Angilletta, 2009). Thermal effects span levels organization, from catalytic activity enzymes (Daniel & Danson, 2010) to whole-organism performance traits such as sprint speed (Bauwens et al., 1995; Van Berkum, 1986; Zamora-Camacho 2015), escape efficiency (Christian Tracy, 1981) or foraging (Angilletta 2002; Díaz, 1994; Hu 2019; Damme 1991). shapes both relationship between phenotypic (e.g., by controlling velocity, force duration skeletal muscle contraction; Niu 2021) fitness affecting growth, energy budgets, reproductive output survivorship). Thus, thermal influence life histories, population dynamics conservation prospects (Adolph Porter, 1993; Buckley, 2008; Kearney 2009; Sinervo 2010). It therefore not surprising global warming stands with habitat loss important causes current biodiversity crisis (Ceballos 2015; Ceballos 2020; Garcia 2014). Warming has sped up in past decade, records surface temperature 2016 2020 (Voosen, 2021). Terrestrial ectotherms are particularly prone suffer consequences (Deutsch 2008) because they rely environment, much more heterogeneous space variable over time than aquatic habitats, regulate their body temperature. Most species projected lose range due limited dispersal ability. For instance, suitable climate likely occur southwestern Europe, where hot, dry conditions increase (Araújo 2006). On other hand, many terrestrial can behavioural means values almost high stable those endotherms Cowles Bogert, 1944; Díaz Cabezas-Díaz, 2004; Hertz 1993). In fact, thermoregulation been regarded a factor inhibits, rather drives, adaptation warming, it may reduce impact change organisms, thus minimizing intensity selection sensitivity physiological (Huey 2003). Studies were fostered early 90s proposal an elegant protocol evaluate small, field-active (Hertz The starting point was long-held view mechanisms extent thermoregulation, indeed very existence active thermoregulatory process, cannot be evaluated without appropriate null hypotheses (Heath, 1964; Hertz, 1992; Such obtained estimating operative temperatures (Tes), defined equilibrium inanimate objects same heat-transfer properties study animals (Bakken Gates, 1975). These estimated directly using physical models organism (‘operative thermometers’; Bakken Angilletta Jr., 2014; 1975; Shine Kearney, 2001; Vickers Schwarzkopf, 2016; Vitt Sartorius, 1999). random Tes environment describes (Tbs) expected nonregulating animals. al. (1993) combined data about Tes, Tbs selected ranges (Tsel; lizards attempt maintain absence biotic constraints; Licht 1966) estimate quantitative indexes precision (standard deviation Tbs), accuracy (average distance Tsel) effectiveness (extent which closer Tsel Tes) thermoregulation. Although some these measurements, especially effectiveness, have criticized (Blouin-Demers Weatherhead, 2001), still provide standard parameters field studies. Despite advances, there paucity empirical studies comparing biology inhabiting unmodified habitat, after long enough period measuring analytical methods. Within framework, we took advantage long-term database Psammodromus algirus, commonest lacertid areas western Mediterranean region. includes taken 1987, 1992, 1994, 1997 2017, capture, manipulation measurement. More specifically, procedure developed replicated summers (July–August) 2017. This allowed us analyse statistics, instead relying predictions envelop modelling. Importantly, our area (Monte de El Pardo) Europe's best preserved forests, royal hunting ground centuries before became part Spain's national heritage. Therefore, landscape remained relatively unchanged, between-year differences availability sunlit shaded patches (see Results) must attributed small variations vegetation cover among sampled transects southern facing slopes open northern ones). consistent fact inter-annual variation shade 2017 magnitude seasonal (Díaz 2005; 2004). methods minimize confounding effects, providing direct link biology. Our specific goals address following questions: (1) How changed two decades well-conserved forest? (2) changes affected behaviour (microhabitat use patches) realized (body thermoregulation)? (3) To what buffer temperature, hence physiology, ectotherms? Because interested making comparable, used revised version described Cabezas-Díaz (2004), summarized next sections. algirus medium-sized (adult snout-vent length 60–85 mm; mass 6–15 g) heliothermic inhabits shrub woodland habitats Carrascal, Lizards at ‘El Pardo’ (Madrid, central Spain: 40°30′N, 03°45′W; 650–700 m elevation), holm oak (Quercus ilex) perennial forest offshoots Q. ilex dominate layer together rockroses Cistus ladanifer. July–August, mean ambient 25.2°C (daily nearby meteorological stations Barajas-Aeropuerto Cuatro Vientos, averaged 40-year 1978 2017), daily maximum 33.0°C. within different categories sun exposure (full sun, filtered vegetation, full shade) simultaneously collected during species. Animals captured under licence ‘Dirección General Biodiversidad y Recursos Naturales’ Community Madrid, care conducted conformity principles approved Animal Experimentation Ethical Committee Universidad Complutense Madrid. 1997, sampling place 17 31 July six transects, per day 19 25 August, again but each transect days. 15 Te thermometers transect, provided covering whole new scatter devices placed randomly intervals (1–9 m, determined one-digit numbers table numbers). 20 below). measured throughout 90 locations 120 unpainted copper cylinders (5 cm × 1 wide mm thick) closed ends except fissure inserting sensing tip electronic digital thermometer (digi-thermo®; ±0.1°C precision; examined Belliure (1996), who showed radiant heat loads (100 W bulb heights), recently dead highly correlated statistically indistinguishable (r2 > 0.99; p 0.25 assumptions identity regression line slope intercept do differ significantly zero respectively). However, validation method absorption emissivity living ones, also complex laboratory. years carried out work Jr. (2014) argued simple geometries ones lead systematic errors 2°–4°C, should match shape, size colour live avoid errors. reason, June 2022 second calibration experiment compared pair identical (and 2004) matching detailed hollow lizard casts built malleable 0.2 thick sheet, moulded contours museum specimen, painted visually lizards. done conditions, ensuring validated outdoor natural exposed modes exchange. We chose sites (12 8 measure models, tethered (one male female) 08:00 14:00 h. Temperatures allow exchange rates, registered 5 min, 1-min minutes 2 case Equilibrium (i.e., rates ≈ 0) sometimes difficult reach, temperatures, when heated slowly types devices, probably prevent lethal overheating (in did four highest [>50°C] death). type entity (simple lizards) site, resulting subsequent statistical analyses. models: r2 = 0.980; casts: 0.974; N 16 << 0.001 cases). produced datasets 0.969, 20; 0.53 respectively; average difference 0.6°C; repeated measures ANOVA: F1,19 2.92, 0.1). results support conclusion statistics metrics only comparable years, accurate. Also, indicate accurate prediction shifts actively shade, Tb (35.9°C) close (35.3°C) May random, shuttling frequently patches, (33.5°C), equilibrating any patch, grand (33.7°C) (46.7°C), (29.8°C) (24.7°C; Kearney's (2001) assumption placement increased predicted Te, suggest distributed consistent. hourly 20:00 h, recorded shade; occasional cloudy excluded analyses) hypothesis against lizards' selectivity sunlit/shaded could tested (Díaz, 1997; lassoed nearest 0.1°C Miller–Weber quick-reading cloacal thermometer. As first sighting (also detected captured). August scattered across (08:00–20:00 h). devices. First, above, since obtaining datasets. Second, Thermochron® iButtons (model DS1921G-F), loggers monitor every larger number estimates Te. pilot establish so (Vitt purpose, matched forming nine pairs three vegetation. 0.978, < 0.001), allowing transformation into model equation (Te [model] 5.4069 + 0.895 TiButton). measurement 10 along line. Thermochrons set register whereas h (five hour). 12 readings iButtons, leading single iButton hour. obtain noted focal observation middle interval. Again, analysis requires sympatrically collected. 26 3 restricted (transects 3) analyses derived indices although conclusions would unchanged had complete dataset Results). gathered entire period, sample sizes log-linear) previously published (Tsel, 80% thermogradient) 32.8–36.9°C males females look possible preferred temperatures. transported laboratory individually housed terraria (40 60 30 cm) white, opaque walls tops covered green net (0.5-cm mesh) prevented created shrubby-like shelter. daylight ventilation. Terraria filled moistened earth leaf litter layer. Some rocks thin fallen wood structural complexity additional refuge. Food water supplied ad libitum. A thermogradient 100 suspended above end height c. cm. gradient offered available (24.3–47.7°C), times (at 10:30–11:00, 13:00–13:30 15:00–15:30 CET), field, 4 (six lizard). five occasions hidden cool thermogradient, thereby considered inactive (not thermoregulating). final 43 field. Tbs, recovered July–August thermometers. compare summer 1989, (Carrascal 1989; study). quality proportion ( d e ¯ $$ \overline{d_{\mathrm{e}}} ) absolute deviations (des). Similarly, quantified b \overline{d_{\mathrm{b}}} (dbs). actual variability Tbs) distributions Tbs. Finally, respect indexes: original formula proposed (1993), E – / \overline{d_{\mathrm{b}}}/\overline{d_{\mathrm{e}}} ); − \overline{d_{\mathrm{e}}}-\overline{d_{\mathrm{b}}} , Blouin-Demers Weatherhead mathematically simpler, better alternative. relative importance (Tes des) weighted patterns 1996; While represents dbs proportional ‘no thermoregulation’ hypothesis), its behaviour. weighting coefficients Li/Mi, Li Mi proportions ith category period. Descriptive presented ± deviation. Both air Vientos (Figure S1; http://www.aemet.es), spanning (mean temperatures: T −82.82 0.054 year, F1,38 24.33, 0.001; −60.14 0.047 12.20, 0.001). trend ca. 0.05°C/year (2.16 1.86°C, respectively, examined; Figure S1). Mean rising 30-years −174.26 0.104 F1,3 131.84, 0.0014), 0.11°C/year, 3.3°C 1987 S2). Lizard differed (one-way F4,327 11.42, reaching 1). percentage (32.8–36.9°C; see 2005) higher (52%) previous (6, 9, 7 9% 1989 Moreover, 18 (94.7%) 38.2°C (the 95th percentile 332 2017) belonged including eight 39°C. average, (two-way ANOVA year factors; year: F1,108 21.43, day: F11,108 12.64, interaction: 2.07, P 0.029; Table 1; 2). raised morning basking remain 09:00 12:00 2; S5). (F1,42 0.41, 0.524; After time, however, clearly (F1,86 54.74, Whereas managed keep rest day, exceeded 13:00 18:00 mentioned paragraph Given pattern, significant interaction interval (a levels, vs. h) two-way (F1,128 6.11, 0.015), categorical predictor facilitate interpretation results. \overline{T_{\mathrm{b}}} (N) 32.5 3.0 1.2 2.2 (22) 35.0 2.0 0.3 0.7 (45) 34.2 2.6 0.6 1.4 (67) 33.1 3.2 2.1 37.7 1.1 0.9 (43) 36.1 3.1 (65) \overline{T_{\mathrm{e}}} 24.0 8.3 10.6 5.3 (360) 41.4 12.1 7.8 9.6 (720) 35.6 13.7 8.7 8.50 (1080) 30.7 10.0 7.4 5.2 (120) 50.0 11.7 13.3 11.5 (240) 43.6 14.4 11.3 10.2 observed rise concomitant shift Tsel, (field: 2.1°C, 65; thermogradient: 34.5 1.7, 43; F1,106 9.60, 0.002; Levene´s test: 7.18, 0.009). 2006) smaller (14, 84, 2% below, respectively) 34, 52% χ2 31.9, df 2, result even clearer if restrict (0%, 23%, 77% 50.8, 26–27 (transect 2) 2–3 3), present days fulfil dataset, might hottest Appendix Supplementary Material). similar 14:00–15:00 followed slower decrease peak Overall, ANOVA; F1,1416 181.32, F11,1416 109.23, 1.42, 0.156; Sun marked effect Tes. reached noon, peaking midday ANOVAs ps exposure, interaction). mostly importantly, frequencies specially (Table 51% below meaning lower ‘heat sinks’. 67% measurements 7% Tsel. frequency pooling (χ2 19.26, 0.001) 93.02, truly challenging 92% Precision sensu 1): (Levene's F1,1145 70.36, (F1,423 82.90, indicating degree test 2017: F1,130 0.75, 0.389). picture considering (before F1,1436 4.10, 0.043; interval: 7.37, 0.007; 57.00, far 3.2°C opposite true, being 5.5°C Thermoregulation s \overline{d_{\mathrm{b}}}\ \mathrm{s} : 4.44, 0.037; marginally non-significant 0.078), F1,86 22.91, 0.8°C 0.02, 0.887), 42.8, 34 45 (75.6%) (15.5%) (8.9%) no (23.3%) remaining 33 (76.7%) 1), is, (1997: 61.72, F1,423 64.31, unequivocal evidence independent ; (F1,1145 0.69, 0.407), \overline{T_{\