Abstract For ectotherms such as lizards, the importance of behavioural thermoregulation in avoiding thermal extremes is well‐established and is increasingly acknowledged in modern studies of climate warming and its impacts. Less appreciated and understood are the buffering roles of retreat sites and activity phase, in part because of logistical challenges of studying below‐ground activity. Burrowing and nocturnal activity are key behavioural adaptations that have enabled a diverse range of reptiles to survive extreme environmental temperatures within hot desert regions. Yet, the direct impact of recent global warming on activity potential has been hypothesised to have caused extinctions in desert lizards, including the Australian arid zone skink Liopholis kintorei. We test the relevance of this hypothesis through a detailed characterisation of the above‐ and below‐ground thermal and hydric microclimates available to, and used by, L. kintorei. We integrate operative temperatures with observed body temperatures to construct daily activity budgets, including the inference of subterranean behaviour. We then assess the likelihood that contemporary and future local extinctions in this species, and those of similar burrowing habits, could be explained by the direct effects of warming on its activity budget and exposure to thermal extremes. We found that L. kintorei spent only 4% of its time active on the surface, primarily at dusk, and that overall potential surface activity will be increased, not restricted, with climate warming. The burrow system provides an exceptional buffer to current and future maximum extremes of temperature (≈40°C reduction from potential surface temperatures), and desiccation (burrows near 100% humidity). Therefore, any climate warming impacts on this species are likely to be indirect. Our findings reflect the general buffering capacity of underground microclimates, therefore, our conclusions for L. kintorei are more generally applicable to nocturnal and crepuscular ectotherms, and highlight the need to consider the buffering properties of retreat sites and activity phase when forecasting climate change impacts.