Thermoregulation in tropical lizards

Thermoregulation is critically important for ectotherms, and there is a large body of literature on the topic. Much of the theory predicting aspects of behavioural thermoregulation stems from lizard biology, and there have been numerous developments in theoretical understanding over the past 30 years. Naturally, as the data increases, and more systems are understood, some of the basic theories and methods underpinning the study of any subject must be updated. New techniques in data collection and modeling improve predictive capacity and confidence in predictions, and can be used to further the understanding of behavioural thermoregulation. As with many organismal processes, the execution of behavioural thermoregulation relies on balancing the costs and benefits, as described in the classic cost benefit model for behavioural thermoregulation written by Huey & Slatkin (1976). Costs of behavioural thermoregulation include the energetic cost of moving to locations at appropriate temperatures, but also more difficult-to-quantify costs, such as home range maintenance (assuming, for example that 'good' home ranges have many good temperatures), and interruption to foraging or mating time while thermoregulating. The benefit derived from body temperature control occurs because processes such as locomotion, digestion, and cognition are dependent on body temperature, which means, ultimately, that fitness is thermally dependent. The relationship between fitness and temperature is described by the thermal performance curve, which is a characteristic asymmetrical inverted U shape, with the peak of the fitness hump (thermal optimum) nearer to the hot end than the cold. This asymmetry means overheating is more costly than overcooling and, I present a new cost-­‐benefit model for behavioural thermoregulation that includes this asymmetry, considers the cost of failing to thermoregulate, and describes these costs and benefits in terms of fitness, rather than energetic cost alone. The key predictions of the model were that a) organisms should invest more (not less, as the previous model predicted) effort in thermoregulation as environmental temperature deviates from the thermal optimum; and b) to offset the increased cost of failing to thermoregulate at high temperatures, organisms will thermoregulate more effectively when it is hot than when it is cold (the previous model did not consider the case in which the environment is too hot). Both predictions of my new model were supported by data from three sympatric rainbow (Carlia spp.) skinks in tropical Australia, and by existing literature that did not support the predictions of the previous null model. Quantifying the costs and benefits of thermoregulation relies on comparison of an organism's body temperature with a null model, or model of a non-thermoregulator. Hertz, Huey & Stevenson (1993) formalized the first null model for behavioural thermoregulation, in which they standardized terminology: body temperature was defined as the temperature of a real organism; operative temperature was defined as the temperature of the null model (or non-thermoregulator). Here, I discuss the limitations of their null model, and propose a new null model that aims to overcome these limits. Originally, operative temperature was determined by placing static models randomly around in the environment, and creating indices describing the available environmental temperatures relative to the preferred body temperature of the modeled species. These indices were calculated using mean absolute differences between the temperature achieved by the physical null model and the organism's preferred body temperature, which was measured in a thermal gradient. Mean absolute differences do not correctly model the influence of the asymmetrical shape of the thermal performance curve, and its fitness consequences. Also, using overall mean temperatures obtained from static physical models assumes there is no spatial or temporal structure to environmental temperatures, and that the organism will reach thermal equilibrium at each site it attends. Foraging strategies of organisms range from ambush to searching, which has broad implications for the level of activity, i.e., the tendency and frequency of movement, of individuals. Rather than static models, I created a computer model that used random walks through a detailed, spatio­temporally realistic thermal landscape to sample environmental temperature. To account for different foraging modes, walk rate can be tailored to match the organism of interest. Rather than using physical models of the organism to collect operative temperature, I used a differential equation to estimate operative temperature from the environmental temperatures that were measured using data loggers placed around the environment (iButtons™). Operative temperature was calculated as function (validated against real lizards) of environmental temperature, recent operative temperature, and the rate of temperature change. Using my model, operative temperature (i.e., the 'body' temperature of the non-thermoregulating computer model) can be calculated as often as desired, and I calculated it every second. Operative temperature and lizard body temperature were converted to a new metric, which I called Thermal Benefit, by transforming operative temperature using the thermal performance curve. Due to its relation to the thermal performance curve, thermal benefit incorporates the asymmetrical effects of being too hot versus too cold. The thermal benefit I calculated for the null model was a temporally integrated estimate of habitat thermal quality, and could be used to determine the difficulty of behavioural thermoregulation, in the sense that if null model thermal benefit was low, it is difficult for a behavioural thermoregulator to achieve preferred body temperature, and vice versa. If the assumptions of the null model are upheld, comparing the estimate of habitat quality obtained from a model with the thermal benefit obtained by real lizards indicates the real benefit gained by thermoregulation. I found that the thermal benefit gained by thermoregulation increased towards the middle of the day, as thermal quality decreased due to high environmental temperatures, suggesting that active lizards worked harder to maintain their preferred temperatures as the temperature of the environment increased over the day. My random walk null model can be used to describe habitat in detail, and test specific hypotheses in thermal ecology. Global temperatures are apparently increasing, but the likely future changes in cloud cover are less clear. The increase in temperature is predicted to reduce the amount of potential daily activity time for ectotherms, which will impact their fitness, and may cause populations to decline in abundance. Cloud cover, on the other hand, influences lizard activity rates by changing the amount of solar radiation reaching the ground, and therefore, local temperatures, sometimes reducing available activity time by decreasing temperature. I used my null model to quantify the implications, in terms of thermal quality of the habitat, of cloudy versus sunny days in contemporary conditions, using 3 years of temperatures of lizard environments, in both winter and summer. The descriptions of habitat quality provided by the model were also calculated for a scenario with 3°C of climate warming. The climate-­‐warming scenario was also run with scenarios including a 30% increase, and a 30% decrease, in number of cloudy days per year. Overall, winter days had higher summed total thermal benefit scores than summer days and, in both seasons, cloudy days had higher total thermal benefit scores than sunny days. Thermal quality in summer decreased when I included climate warming, but in winter thermal quality increased enough to offset summer's decrease over the entire year: i.e., total annual thermal quality of the habitat was better when there was climate warming. Increasing cloud cover linearly increased thermal benefit in summer, because it buffered lethally hot conditions somewhat, although never enough to balance the negative effects on thermal benefit caused by increased environmental temperatures. If summer is a critical period in terms of activity, the effect of high temperatures may cause problems. This example demonstrates variability in the effect of changing temperature interacting with other environmental factors, such as cloud cover. Studies predict that tropical ectotherms are at particular risk from climate change. Tropical species tend to be thermal specialists, and live in environments near, or even above, their thermal optimum. In the tropics, increasing temperatures should cause fitness to decrease, because environmental temperatures often exceed thermal optima and maxima of most species. Using concepts generated while defining cost benefit model for thermoregulation and the new null model, I argue that, as thermal specialists, tropical ectotherms tend to be highly precise thermoregulators, and are particularly adept at thermoregulating in high temperature environments. I suggest that these traits could provide the behavioural buffer required to filter the negative effects of increased temperature, altering our expectations of the effect of climate change. The uncertainty of some aspects of future climate, together with the unknown extent of buffering due to behavioural thermoregulation means that predictions of gloom or success for ectotherms are premature