To the top or into the dark? Relationships between elevational and canopy cover distribution shifts in mountain forests

Numerous studies have reported that observed species shifts in mountain areas lag behind expectations under current warming trends, however, the mechanisms remain poorly understood. One important mechanism might be microclimatic heterogeneity causing migration of species to cooler conditions under closed forest canopies; however, evidence is scarce. We here compared the distributions of 710 species (11 taxonomic groups including fungi, plants, and animals) along an elevation gradient (287-1,419 m a.s.l.) in a temperate low mountain range between 2006-2008 and 2016-2017 to address this open question. We characterized each species’ distribution (peak and breadth) based on their abundance along two environmental gradients: elevation and canopy cover. We then analysed changes in species’ distribution peaks, asking whether shifts in canopy distribution and initial distribution characteristics explain variation in elevational distribution shifts. Across all taxa, the mean shift in elevational distribution peak was +35.3 metres (i.e., upslope). Species' baseline distribution peaks were strong predictors of elevational distribution shifts with stronger upslope shifts in low-elevation and open-forest species. Even though we observed considerable variation in the responses among species, canopy distribution shifts had a significant negative effect on elevational distribution shifts overall and in six taxonomic groups. We suggest that this is related to cooler microclimatic conditions under closed compared to open forest canopies. Shifts to closed-canopy forests may thus partly compensate for elevational distribution shifts, highlighting the conservation value of heterogeneous landscapes featuring microclimatic refugia. Yet, it is likely that other mechanisms, such as habitat limitation, are also at play. Future studies need to quantify the potential of microclimatic refugia under accelerating forest dynamics considering the interplay of canopy cover and other factors driving microclimate and to illuminate the complex climate change response mechanisms among species and taxonomic groups.

Methods
We used abundance data of 710 species (11 taxonomic groups including fungi, plants, and animals) collected during two surveys (2006-2008 and 2016-2017) in the Bavarian Forest National Park along an elevation and a canopy cover gradient using taxon-specific standard survey methods. The canopy cover of the upper two tree layers was visually estimated for all plots during both surveys and we used a probabilistic approach (Fischer 2015) accounting for the overlap between vegetation layers to obtain one combined value for canopy cover per plot and survey. We then calculated the species-specific abundance-weighted mean (= distribution peak) and standard deviation (= distribution width) of each gradient value during both surveys. We calculated each species' distribution shift in elevation and canopy cover as the change in distribution peak between the two surveys. Finally, we used linear (mixed) models to assess the relationship between distribution shifts along the elevational gradient with distributions shifts along the canopy cover gradient as well as species' baseline distribution peaks and breadths.

For details see the Methods section and Supporting information of the related article.