Activity niches outperform thermal physiological limits in predicting global ant distributions

Abstract

Aim Thermal physiology is commonly used in mechanistic models to predict species distribution and project distribution change. Such thermal constraints for ants are often measured under laboratory conditions as critical thermal limits (CTmax and CTmin), but have also been observed in the field as foraging thermal limits (FTmin and FTmax). Here we compared distribution projections based on ant physiological and behavioural thermal limits with their realized distributions to assess the validity of using ecophysiological models to predict species ranges over large scales. Location Global. Taxon Ants. Methods From published literature, we compiled a database of 148 lab‐measured critical thermal limits and 137 field‐observed foraging thermal limits for 20 ant genera distributed around the world. We projected their potential ranges using active hour thresholds by matching their thermal limits or preferred thermal breadth (incorporating thermal optima) with hourly surface temperature data. We then compared the projections against a comprehensive global database of ant distributions to assess the validity of physiologically and behaviourally based models. Results We found that surface temperature plays a clear role in generic‐level ant biogeography. Projections based on foraging thermal limits were more conservative in constraining ranges to observed distributions, while projections based on critical thermal limits often overestimated ranges. Furthermore, thermal limit models can be improved by incorporating behaviourally derived thermal optima. Main conclusions Our results suggest that temperature‐dependent activity niches and thermal optima better reveal species realized thermal niches than critical thermal limits, and hence can improve predictions of species distributions and future distribution changes under climate change.

Realised rather than fundamental thermal niches predict site occupancy: Implications for climate change forecasting

“…Explained deviance for patterns of relative abundance was, however, generally low across both sets of data. Our findings contrast with those for other insects where critical limits are related to geographic range (García‐Robledo et al., 2016; Sheldon & Tewksbury, 2014), but build on a growing understanding of the complexity of the relationship between critical limits and abundance and occupancy of ants (Bujan, Roeder, de Beurs, et al., 2020; Guo et al., 2020), which may also apply to ectotherms more broadly. Previous work has shown that in ants the relationship between critical limits and geographic range varies between local and global scales (Diamond & Chick, 2018; Nowrouzi et al., 2018).…”

“…Rather, the range of realised foraging temperatures explained much of the observed variation. This finding, that realised niches are of greater value for predicting occupancy and abundance than their fundamental counterparts, is corroborated by recent work by Guo et al (2020) where mechanistic species distribution models were found to fit better with observed distributions when built using the foraging limits rather than the physiological limits of a species.…”

“…It also implies that even simple estimates of changing assemblage composition may improve forecasts of the impacts of environmental change. Further support of the forecasting benefits of considering realised thermal niches in addition to, or in place of, their fundamental counterparts is afforded by the substantial improvements that realised niche data can provide to mechanistic models of present‐day distributions of ants (Guo et al., 2020).…”

“…Temperature Thermal niche type between critical limits and abundance and occupancy of ants (Bujan, Roeder, de Beurs, et al, 2020;Guo et al, 2020), which may also apply to ectotherms more broadly. Previous work has shown that in ants the relationship between critical limits and geographic range varies between local and global scales (Diamond & Chick, 2018;Nowrouzi et al, 2018).…”

“…Reasons exist, however, for caution when applying trait‐based ecological forecasting when such traits represent the fundamental rather than realised niche of a species. Forecasts relying exclusively on the fundamental thermal niche fail to consider behavioural responses either to the abiotic environment or to other species that might effect significant differences between the fundamental and realised thermal niches (Arnan & Blüthgen, 2015; Guo et al., 2020; Huey et al., 2012; Mitchell & Angilletta, 2009; Sunday et al., 2014). Hence, environmentally mediated interactions among species (Chase, 1996; Park, 1962; Sinclair et al., 2016) may result in multi‐species outcomes very different to those predicted on a single species basis.…”