Insects in temperate regions are predicted to be at low risk of climate change relative to tropical species. However, these assumptions have generally been poorly examined in all regions, and such forecasting fails to account for microclimatic variation and behavioural optimisation. Here, we test how a population of the dominant ant species, Iridomyrmex purpureus, from temperate Australia responds to thermal stress. We show that ants regularly forage for short periods (minutes) at soil temperatures well above their upper thermal limits (upper lethal temperature = 45.8 ± 1.3°C; CT(max) = 46.1°C) determined over slightly longer periods (hours) and do not show any signs of a classic thermal performance curve in voluntary locomotion across soil surface temperatures of 18.6-57°C (equating to a body temperature of 24.5-43.1°C). Although ants were present all year round, and dynamically altered several aspects of their thermal biology to cope with low temperatures and seasonal variation, temperature-dependence of running speed remained invariant and ants were unable to elevate high temperature tolerance using plastic responses. Measurements of microclimate temperature were higher than ant body temperatures during the hottest part of the day, but exhibited a stronger relationship with each other than air temperatures from the closest weather station. Generally close associations of ant activity and performance with microclimatic conditions, possibly to maximise foraging times, suggest I. purpureus displays highly opportunistic thermal responses and readily adjusts behaviour to cope with high trail temperatures. Increasing frequency or duration of high temperatures is therefore likely to result in an immediate reduction in foraging efficiency. In summary, these results suggest that (1) soil-dwelling temperate insect populations may be at higher risks of thermal stress with increased frequency or duration of high temperatures resulting from climate change than previously thought, however, behavioural cues may be able to compensate to some extent; and (2) indices of climate change-related thermal stress, warming tolerance and thermal safety margin, are strongly influenced by the scale of climate metrics employed.