Global pattern of forest disturbances and its shift under climate change

Forests are continuously altered by disturbances. Yet, knowledge of global pattern of forest disturbance agents, its drivers, and shifts under changing climate remain scarce. Here we present a meta-analysis of current and projected (+2 degrees and + 4 degrees C) distribution of forest disturbance agents causing immediate tree mortality (i.e., fire, pest outbreak, hydro-geomorphic, and wind) at country, continental, biome, and global scales. The model including combination of climatic (precipitation of driest quarter, actual evapotranspiration, and minimum temperature), geographical (distance to coast and topography complexity), and forest characteristics (tree density) performs better than any other model in explaining the distribution of disturbance agents (R-2 = 0.74). We provide global maps (0.5 degrees x 0.5 degrees) of current and potential future distribution of forest disturbance agents. Globally, the most frequent disturbance agent was fire (46.09 %), followed by pest outbreak (23.27 %), hydro-geomorphic disturbances (18.97 %), and wind (11.67 %). Our projections indicate spatially contrasting shifts in disturbance agents, with fire and wind risk increase between similar to 50 degrees S and similar to 40 degrees N under warming climate. In particular, the substantial increase in fire risk, exceeding 31 % in the most affected areas, is projected over Mediterranean, the western and southeast USA, African, Oceanian, and South American forests. On the other hand, pest outbreak and hydro-geomorphic disturbances are projected to increase in more southern (> similar to 50 degrees S) and northern (> similar to 40 degrees N) latitudes. Our findings are critical for understanding ongoing changes and developing mitigation strategies to maintain the ecological integrity and ecosystem services under shifts in forest disturbances. We suggest that projected shifts in the global distribution of forest disturbance agents needs to be considered to future models of vegetation or carbon sink dynamics under projected climate change.