Assessment of multiple outcomes of habitat models can significantly affect conservation decisions for threatened species.

There is a clear international commitment to protect old-growth forests for their biodiversity and ecological functions. However, it remains unclear which particular solutions would best mitigate overall biodiversity loss given uncertainties about the future, and how to assess that. As an approach, we modelled and assessed the current and likely future habitat distribution of a conspicuous deadwood-inhabiting fungal species, Anthoporia albobrunnea, across Europe. We used spatial distribution modelling (MaxEnt) to predict the potential future distributions for alternative climatic and habitat shifts, and we used two distinct ('cautious' and 'optimistic') interpretations of those products by also considering current habitat protection. We found that, although the higher radiative forcing climate scenario (SSP5-8.5) reduced habitat more than the conservative SSP2-4.5, an even larger difference was caused by using or neglecting a precautionary approach while generalising from multiple scenarios. This difference would further increase when habitat-turnover estimates (primarily due to timber harvesting) are added. Our approach shows that simply incorporating the best technical knowledge is insufficient for improved forest conservation, as the outcomes of decision-making critically depend on how uncertainties are assessed during the process.