Realistic projections of the biological impacts of climate change require predicting fitness responses to variations in environmental conditions. For ectotherms, this challenge requires methods to scale-up microclimatic information into actual body temperatures, Tb, while dealing with uncertainty regarding individual behaviors and physiological constraints. Here, we propose an information-theoretical model to derive microhabitat selection and Tb distributions of ectotherm populations from microclimatic data. The model infers the most probable allocation of individuals among the available microenvironments and the associated population-level Tb distribution. Using empirical Tb data of 41 species of desert lizards from three independently evolved systems-Western North America, Kalahari Desert, and Western Australia-we show that the model accurately predicts empirical Tb distributions across the three systems. Moreover, the framework naturally provides a way to quantify the importance of thermoregulation in a thermal environment and thereby a measurement for the constraint imposed by the climatic conditions. By predicting Tb distributions of ectotherm populations even without exhaustive information on the underpinning mechanisms, our approach forms a solid theoretical basis for upscaling microclimatic and physiological information into a population-level fitness trait. This scaling process is a first step to reliably project the biological impacts of climate change to broad temporal and spatial scales.