The importance of fine-scale refugia and behavioral thermoregulation in the resilience of intertidal limpet populations.

Fine-scale spatial variability can play a key role in determining the distribution and abundance of organisms living in heterogenous habitats, where small-scale spatial variation in temperature can often exceed daily variation at any single location. However, many models of species distributions ignore such organism-scale abiotic variability and instead focus only on large-scale biogeographic patterns. Here, we investigated the importance of fine-scale temperature variability in population resilience of intertidal limpets, which are widely studied sentinels of climate change. To do this, we used a heat-budget model coupled with fine-scale reef-surface models to predict individual-scale limpet body temperatures. Initial modeling for 12 years (2009-2022) showed an extremely hot day during which the predicted body temperatures of an exposed limpet exceeded 39°C, which is lethal for all four of the limpet species studied (Cellana spp.) based on published thermal tolerances. Using this day as an exemplar thermal event, we then incorporated fine-scale (0.02 × 0.02 m resolution) topographic models of five New Zealand intertidal rocky reefs into the heat-budget model to quantify the effects of small-scale topographic variation. Predicted body temperatures of limpets during this exceptional day were highest on horizontal and equator-facing surfaces. Homing species (Cellana flava and Cellana ornata) tend to occupy these hot surfaces but have higher thermal tolerances and relatively high average estimates of survival (>75%). Species with lower thermal tolerances (Cellana radians and C. denticulata) would have lower survival if scattered randomly across the reef (65 or 72%, respectively), but their behavioral tendency to move to poleward-facing surfaces is estimated to increase survival by 38%-46% (to 95 or 99%). Estimates of survival generally agreed with our long-term (six years) limpet population data in which no extreme declines were detected. When the heat-budget model was presented with a smoothed version of the topography, reducing variation caused by microhabitats, sitewide modeled survival of one species decreased from ~68% to 38%. This study demonstrates the importance of incorporating relevant individual-scale topographical, physiological, and behavioral information to accurately estimate resilience and long-term persistence of populations following extreme events.