The impact of climate warming on biodiversity loss is exacerbated not only by changes in mean but also by spatio-temporal variability in temperature. Access to refugia can mitigate the impact of thermal fluctuations amongst species. The effectiveness of refugia during periods of adverse warming scenarios, i.e., seasonal fluctuations, hotter-than-average summers, and warmer-than-average winters remains largely unexplored. Here, we study a bio-energetic consumer-resource model and identify the mixed success of refugia in maintaining species persistence and stability, depending on the amplitude of fluctuations, diverse warming scenarios, and species body size. Whilst refugia withhold otherwise inevitable extinction at high amplitude fluctuations in all the warming scenarios, at lower amplitudes, they may not provide similar benefits. This arises due to non-monotone thermal responses of their foraging efforts and monotonically increasing metabolic requirements. The qualitative difference among thermal responses leads to more energy losses rather than gains at low amplitudes. We find that refugia are most beneficial during hotter summers and least during typical seasonal fluctuations. Our results also suggest that refugia can be more favourable to species in temperate and Mediterranean regions, unlike those inhabiting tropical regions. We also consider an extreme heat wave event and observe that small-bodied species can counteract their negative effects by seeking refuge at low amplitudes. Overall, our work hints at selective adaptation to refugia - conditioned on the aggregated effect of thermal conditions of the local habitat and species body size - as a mechanism for biodiversity maintenance.

Early warning signals (EWSs) represent a potentially universal tool for identifying whether a system is approaching a tipping point, and have been applied in fields including ecology, epidemiology, economics, and physics. This potential universality has led to the development of a suite of computational approaches aimed at improving the reliability of these methods. Classic methods based on univariate data have a long history of use, but recent theoretical advances have expanded EWSs to multivariate datasets, particularly relevant given advancements in remote sensing. More recently, novel machine learning approaches have been developed but have not been made accessible in the R environment. Here, we present EWSmethods – an R package that provides a unified syntax and interpretation of the most popular and cutting edge EWSs methods applicable to both univariate and multivariate time series. EWSmethods provides two primary functions for univariate and multivariate systems respectively, with two forms of calculation available for each: classical rolling window time series analysis, and the more robust expanding window. It also provides an interface to the Python machine learning model EWSNet which predicts the probability of a sudden tipping point or a smooth transition, the first of its form available to R users. This note details the rationale for this open-source package and delivers an introduction to its functionality for assessing resilience. We have also provided vignettes and an external website to act as further tutorials and FAQs.