资源环境科技发展态势分析平台

Climate change may reduce forest growth and increase forest mortality, which is connected to high carbon costs through reductions in gross primary production and net ecosystem exchange. Yet, the spatiotemporal patterns of vulnerability to both short-term extreme events and gradual environmental changes are quite uncertain across the species' limits of tolerance to dryness. Such information is fundamental for defining ecologically relevant upper limits of species tolerance to drought and, hence, to predict the risk of increased forest mortality and shifts in species composition. We investigate here to what extent the impact of short- and long-term environmental changes determines vulnerability to climate change of three evergreen conifers (Scots pine, silver fir, Norway spruce) and two deciduous hardwoods (European beech, sessile oak) tree species at their southernmost limits of distribution in the Mediterranean Basin. Finally, we simulated future forest growth under RCP 2.6 and 8.5 emission scenarios using a multispecies generalized linear mixed model. Our analysis provides four key insights into the patterns of species' vulnerability to climate change. First, site climatic marginality was significantly linked to the growth trends: increasing growth was related to less climatically limited sites. Second, estimated species-specific vulnerability did not match their a priori rank in drought tolerance: Scots pine and beech seem to be the most vulnerable species among those studied despite their contrasting physiologies. Third, adaptation to site conditions prevails over species-specific determinism in forest response to climate change. And fourth, regional differences in forests vulnerability to climate change across the Mediterranean Basin are linked to the influence of summer atmospheric circulation patterns, which are not correctly represented in global climate models. Thus, projections of forest performance should reconsider the traditional classification of tree species in functional types and critically evaluate the fine-scale limitations of the climate data generated by global climate models.