Résumé
The role of modern climatic microrefugia is a neglected aspect in the study of biotic responses to anthropogenic climate change. Current projections of species redistribution at continental extent are based on climatic grids of coarse ( 1 km) resolutions that fail to capture spatiotemporal dynamics associated with climatic microrefugia. Here, we review recent methods to model the climatic component of potential microrefugia and highlight research gaps in accounting for the buffering capacity due to biophysical processes operating at very fine (< 1 m) resolutions (e.g. canopy cover) and the associated microclimatic stability over time (i.e. decoupling). To overcome this challenge, we propose a spatially hierarchical downscaling framework combining a free-air temperature grid at 1 km resolution, a digital elevation model at 25 m resolution and small-footprint light detection-and-ranging (LiDAR) data at 50 cm resolution with knowledge from the literature to mechanistically model sub-canopy temperatures and account for microclimatic decoupling. We applied this framework on a virtual sub-canopy species and simulated the impact of a warming scenario on its potential distribution. Modelling sub-canopy temperatures at 50 cm resolution and accounting for microclimatic stability over time enlarges the range of temperature conditions towards the cold end of the gradient, mitigates regional temperature changes and decreases extirpation risks. Incorporating these spatiotemporal dynamics into species redistribution models, being correlative, mechanistic or hybrid, will increase the probability of local persistence, which has important consequences in the understanding of the capacity of species to adapt. We finally provide a synthesis on additional ways that the field could move towards effectively considering potential climatic microrefugia for species redistribution.