Habitat associations drive species vulnerability to climate change in boreal forests

Species climate change vulnerability, their predisposition to be adversely affected, has been assessed for a limited portion of biodiversity. Our knowledge of climate change impacts is often based only on exposure, the magnitude of climatic variation in the area occupied by the species, even if species sensitivity, the species ability to tolerate climatic variations determined by traits, plays a key role in determining vulnerability. We analyse the role of species’ habitat associations, a proxy for sensitivity, in explaining vulnerability for two poorly-known but species-rich taxa in boreal forest, saproxylic beetles and fungi, using three IPCC emissions scenarios. Towards the end of the 21st century we projected an improvement in habitat quality associated with an increase of deadwood, an important resource for species, as a consequence of increased tree growth under high emissions scenarios. However, climate change will potentially reduce habitat suitability for ~9–43 % of the threatened deadwood-associated species. This loss is likely caused by future increase in timber extraction and decomposition rates causing higher deadwood turnover, which have a strong negative effect on boreal forest biodiversity. Our results are species- and scenario-specific. Diversified forest management and restoration ensuring deadwood resources in the landscape would allow the persistence of species whose capacity of delivering important supporting ecosystem services can be undermined by climate change.

consequence of increased tree growth under high emissions scenarios. However, climate change will potentially reduce habitat suitability for ~9-43 % of the threatened deadwood-associated species. This loss is likely caused by future increase in timber extraction and decomposition rates causing higher deadwood turnover, which have a strong negative effect on boreal forest biodiversity. Our results are species-and scenario-specific. Diversified forest management and restoration ensuring deadwood resources in the landscape would allow the persistence of species whose capacity of delivering important supporting ecosystem services can be undermined by climate change.
107 108 109 Original SCC values vary between 0 and 1. To put more emphasis on changes in values in 110 the middle of the gradient (SCC values around 0.5 having more biological importance), we 111 transformed CV values as follows:  (A1B) (Fig. 2).

182
With concern to decay stage associations, losers were more frequently associated with were associated with sunny microclimate.  Table 3).  , and microclimate [M]. The attribution of threatened species to each habitat association is based on the notes reported in Tikkanen et al. (2006). Species response: average climate vulnerability (CV) among the NFI sample plots. Species responses can be: winner (the species experiences an improvement in its habitat quality), loser (the species experiences a reduction in its habitat quality), or stable ( climate change (Mazziotta et al. 2014;Kellomäki et al. 2008).

213
The higher number of losers than winners for two emission scenarios, notwithstanding the 214 overall increase in habitat quality, highlights the importance of accounting for species-habitat 215 associations when evaluating vulnerability. This mismatch may be explained by the fact that, 216 even though global warming is expected to increase deadwood availability through increased 217 tree growth and mortality, the increased rate in deadwood turnover may ultimately limit increasing deadwood habitats. However this increased habitat availability is still partly deter-225 mined by the local landscape suitability.

226
Suitable habitat conditions will increase in the future for winners species. However, even if 227 habitat becomes available, many of these species may be unable to colonize this new space 228 because of limited dispersal ability (Menéndez et al. 2006;Devictor et al. 2008 (Urban et al. 2012). Many species may continue to persist at local scale as an effect of 235 extinction debt even after many decades of unfavorable environmental changes, ultimately 236 maintaining high local levels of species richness, but their populations might become extinct in 237 the long run (Hyvärinen et al. 2006;Berglund and Jonsson 2005). 238 We forecasted a positive trend in species associated with Scots pine and deciduous trees as a 239 consequence of the predicted enhancement in annual growth of these tree species with 240 increasing emissions (Mazziotta et al. 2014;Kellomäki et al. 2008). We also projected a 241 decline of species preferring well-decayed deadwood. This stems from the fact that with 242 climate change the retention time of the deadwood stock will be reduced by increased 243 decomposition rates (Tuomi et al. 2011) making their habitats more temporary. Climate change 244 also results in more frequent final harvest and subsequent harrowing (Kellomäki et al. 2008), 245 further shortening deadwood retention times (Rabinowitsch-Jokinen et al. 2010). On the other 246 hand, fresh deadwood will become more available, favoring species associated with this 247 resource.

248
Earlier research has suggested that the relative importance that climatic exposure and 249 ecological sensitivity have in determining vulnerability depends on the spatial scale.

250
Exposure has more importance than sensitivity at the landscape or regional scale, while the 251 opposite tends to be true at the local scale (Bradshaw et al. 2014;Garcia et al. 2014;Arribas Climatic Change Summers et al. 2012). In our regional-scale analysis, we projected that sensitivity, 253 i.e., habitat associations, accounted for a much larger proportion of the variance in vulnera-254 bility than exposure. Evidently, whether exposure or sensitivity is driving vulnerability varies 255 not only with spatial scale, but also among taxa and perhaps within their ecological niches.

256
Earlier work has shown that the abundance, diversity and community composition of wood-