High thematic resolution land use change models refine biodiversity scenarios: A case study with Belgian bumblebees

Aim: Projections of biodiversity scenarios often rely solely on climate change to inform species distribution shifts in the future. Land use projections are rarely used due to their unavailability and, when available, are often at coarse spatial and thematic resolutions, making them unsuitable to capture fine scale habitat suitability. This study aims to (a) show how coupled land use change (LUC) models of high thematic resolution (HTR) can be used in species distribution models (SDM), (b) compare the impacts of HTR and low thematic resolution (LTR) explanatory predictors on biodiversity scenarios and (c) assess the impact of species' present area of occupancy on the effect of thematic resolution in SDMs. Location: Belgium. Taxon: Bumblebees (Bombus). Methods: We compared species distribution models with 17 land use predictors (HTR) against models with 5 land use predictors (LTR). We modelled the distribution of 17 bumblebee species in Belgium projected until 2035. We examined how model performance, variable importance and projections of distribution change differed depending on the thematic resolution of the land use predictors. Results: Overall, HTR models performed better than LTR models. LTR models predicted greater extent per species. HTR projected a greater percentage of range gains, and both models projected similar losses of suitable habitat. However, the percentage loss and connectivity of suitable habitats varied differently for HTR and LTR models along a gradient of rare to common species. The HTR models projected greater loss of suitable areas for rare species and less loss for common species compared to LTR models. Main conclusions: These results illustrate the importance of using ecologically relevant explanatory variables in SDMs, particularly for rare and localized species with specific habitat requirements. The results also indicate the need for large-scale LUC projections to improve future biodiversity scenarios under climate change and to improve the ability of conservationists and policymakers to use SDM projections.