Topoclimate versus macroclimate: how does climate mapping methodology affect species distribution models and climate change projections?

Abstract

Abstract Aim We analyse how and why ‘topoclimate’ mapping methodologies improve on macroclimatic variables in modelling the distribution of biodiversity. Further, we consider the implications for climate change projections. Location Greater Hunter Valley region ( c . 60,000 km 2 ), New South Wales, Australia. Methods We fitted generalised linear models to 295 species of grasses and ferns at fine resolutions (< 50 m 2 ) using (a) macroclimatic variables, interpolated from weather station data using altitude and location only, (b) topoclimatic variables, interpolated from field measurements using additional climate‐forcing factors such as topography and canopy cover, and (c) both topoclimatic and macroclimatic variables. We conducted community‐level analyses and examined the reasons for differences through single‐species analyses. We projected species distributions under 0–3° warming, comparing biodiversity loss predicted by topoclimate and macroclimate variables. Results At the community level, the topoclimatic variables explained significant variation ( p < 0.002) in the distribution of both ferns and grasses not explained by macroclimatic variables, resulting in increases of 0.036–0.061 in the pseudo R ‐squared. Topoclimate performed better (as determined by AIC) than macroclimate for grass species living in cold extremes under topoclimate and most fern species. Models using topoclimatic temperature variables projected different locations of biodiversity loss/retention and in general projected substantially fewer species becoming critically endangered in the study region than models using macroclimatic temperature variables – in one scenario, topoclimate projected 10% of species becoming critically endangered where macroclimate projected 28%. Main Conclusions How climate variables are constructed has a significant effect on species distribution models and any subsequent climate change predictions. Misleading conclusions may result from models based on fine‐resolution climate data if climate‐forcing factors such as cold air drainage, topography and habitat have not been addressed in the climate mapping methodology.