Skip to main content

Time of activity is a better predictor of the distribution of a tropical lizard than pure environmental temperatures

Cite this dataset

de Oliveira Caetano, Gabriel Henrique et al. (2020). Time of activity is a better predictor of the distribution of a tropical lizard than pure environmental temperatures [Dataset]. Dryad.


Environmental temperatures influence ectotherms’ physiology and capacity to perform activities necessary for survival and reproduction. Time available to perform those activities is determined by thermal tolerances and environmental temperatures. Estimates of activity time might enhance our ability to predict suitable areas for species’ persistence in face of climate warming, compared to the exclusive use of environmental temperatures, without considering thermal tolerances. We compare the ability of environmental temperatures and estimates of activity time to predict the geographic distribution of a tropical lizard, Tropidurus torquatus. We compared 105 estimates of activity time, resulting from the combination of four methodological decisions: (1) How to estimate daily environmental temperature variation (modeling a sinusoid wave ranging from monthly minimum to maximum temperature, extrapolating from operative temperatures measured in field or using biophysical projections of microclimate)? (2) In which temperature range are animals considered active? (3) Should these ranges be determined from body temperatures obtained in laboratory or in field? and (4) Should thermoregulation simulations be included in estimations? We show that models using estimates of activity time made with the sinusoid and biophysical methods had higher predictive accuracy than those using environmental temperatures alone. Estimates made using the central 90% of temperatures measured in a thermal gradient as the temperature range for activity also ranked higher than environmental temperatures. Thermoregulation simulations did not improve model accuracy. Precipitation ranked higher than thermally related predictors. Activity time adds important information to distribution modeling and should be considered as a predictor in studies of the distribution of ectotherms. The distribution of T. torquatus is restricted by precipitation and by the effect of lower temperatures on their time of activty and climate warming could lead to range expansion. We provide an R package “Mapinguari” with tools to generate spatial predictors based on the processes described herein.


Distribution data. We used 359 distribution records from the literature and scientific collections spanning the range of Tropidurus torquatus. To minimize the effects of spatial autocorrelation and sampling bias, we used function ‘clean_points’ from the R package Mapinguari to eliminate records within 40 km from each other, leaving us with 144 records. We empirically determined the size of the buffer area fitting Random Forest models under different buffers (1, 5, 10, 20, 30, 40 and 50 km) and comparing Moran’s I index (Gittleman and Kot 1990) calculated from the models’ residuals. We selected buffer distance based on the smaller distance resulting in no spatial autocorrelation. We estimated Moran’s I using the ‘Moran.I’ function from R package ‘ape’ (Paradis et al. 2004). Thirty percent of the distribution data, 44 records, was set aside for model cross-validation.

Physiological data. Between 2013 to 2017, we obtained physiological data from five populations of T. torquatus sampled during monitoring studies and field expeditions. Monitoring took place in Brasília, Distrito Federal (15.7998°S, 47.8645°W, 24 individuals) and Nova Xavantina, Mato Grosso (14.6644°S, 52.3585°W, 4 individuals). Short-term field sampling occurred at Gaúcha do Norte (12.9656°S, 53.5636°W, 13 individuals) and Alta Floresta, Mato Grosso (9.8765°S, 56.0855°W, 3 individuals); and Lagoa da Confusão, Tocantins (10.9201°S, 50.1833°W, 8 individuals). We captured animals using pitfall traps, lassos and by hand.

We brought captured lizards to the laboratory, housed them individually and performed the thermal gradient experiments no longer than 24 h after capture. We measured the preferred temperature of each lizard using a thermal gradient, which consisted of a terrarium made of MDF plywood (Medium Density Fiberboard, 100 cm x 15 cm x 30 cm – L x W x H), open at the top and with 2 cm of substrate composed of sand and vermiculite. We generated a thermal gradient approximately between 15 °C – 50 °C by placing a 60-W incandescent lamp at one end and an ice pack on the other (Paranjpe et al. 2013). Lizards were placed in the gradient for one hour while their body temperature was recorded every minute by a 1 mm thermocouple attached with tape to their abdomen and connected to a data logger (Eltek® 1000 Series Squirrel Meter Data Logger 64K, 10 Channel 1001WD). We allowed lizards to acclimate to the gradient for 10 min (Paranjpe et al. 2013) before recording body temperatures. We calculated for each individual and for the whole sample: (1) the range between the 5th and 95th temperature percentile (T90), (2) range between the 25th and 75th temperature percentile (T50) and (3) average temperature (Tmean). Tmean and T50 have been used in previous studies (Sinervo et al. 2010, Kubisch et al. 2016, Piantoni et al. 2016), and the broader range, T90, was chosen under the hypothesis that lizards spend almost all of their time in the gradient at preferred temperatures.

We obtained field-active body temperatures from lizards at Brasília, Distrito Federal, from natural populations occurring within the city’s Zoo (15.8512°S, 47.9379°W, 1158 samples, 640 individuals, details in Wiederhecker et al. 2002), which was visited weekly from March 1996 to September 1998, from 8 am to 6 pm, and at Santa Terezinha, Mato Grosso (10.3705°S, 50.5145°W, 9 samples, 9 individuals) in April 1999, from 12 pm to 2 pm. Active animals (i.e., those in the open, basking or moving) were captured, individually marked by toe-clipping, and had their cloacal temperature measured with a Miller & Weber T-6000 quick reading cloacal thermometer (0.02 ºC precision) immediately after capture. We then performed the same calculations for laboratory T90, T50 and Tmean on the aggregated field body temperatures.

The different methods of collecting body temperatures result in very different data structures. While the laboratory experiments allow extensive sampling of fewer individuals, field sampling allows the collection of many individuals, but fewer replicates per individual. In the laboratory, we sampled 52 individuals with a median of 65 samples per individual (standard deviation = 9.48), whereas in the field we sampled 649 individuals with a median of 1 sample per individual (standard deviation = 1.58). This presents a challenge when comparing data from the two sources, because we could calculate temperature ranges for each individual from the thermal gradient, but not from individuals in the field. Therefore, for the thermal measurements collected in the wild, we pooled the data and assumed the estimated thermal tolerances characterized the individuals from the entire sample. For laboratory data, we estimated temperature ranges as both averages of individual values or from the data aggregated from the whole sample, and then assessed which choice generated better predictors of distribution. We performed an analysis of variance with repeated measures to evaluate whether the body temperatures measured in the thermal gradient differed among individuals between populations and analysis of variance to see if there were differences between individuals in the same population.

Operative Environmental Temperatures. We recorded operative temperatures using dataloggers (HOBO® U23 Pro v2 2x External Temperature Data Logger U23-003) with sensors attached to PVC models of equivalent size and color as Tropidurus torquatus. This methodology has been validated by previous studies with small ectotherms (Adolph 1990, Lara-Reséndiz et al. 2015, Kubisch et al. 2016, Kirchhof et al. 2017). We placed models adjacent to pitfall trap arrays, in the locations where lizards were captured for the physiological trials, uniformly distributed in microhabitats where they were observed in activity—shaded and open spots on the ground, on termite mounds, and at the base of trees. Data loggers were deployed during months of August 2013 and April to July 2014 in Brasília, August 2015 to August 2016 in Nova Xavantina, August 2015 in Gaúcha do Norte, July to August 2016 in Lagoa da Confusão and July to August 2017 in Alta Floresta. Data loggers recorded temperatures every 10 min during the trapping period at each location. Variation in air temperature was also measured at the same time and locations, using another data logger without a PVC model attached to sensors (HOBO® U23 Pro v2 Temperature/Relative Humidity Data Logger U23-001), which was protected from rain and solar radiation by a PVC case suspended about 30 cm from the ground and open at the bottom to expose the sensor to the air.


Coordenação de Aperfeicoamento de Pessoal de Nível Superior, Award: 99999.013716/2013-01

United States Agency for International Development, Award: AID-OAA-A-11-00012

National Science Foundation, Award: EF-1241848