Rising tree mortality in France is associated with distinct seasonal climate anomalies

Abstract

Abstract Tree mortality is rising across European forests, yet its dominant drivers remain difficult to disentangle. Analyzing over 500,000 trees from the French National Forest Inventory (2015-2023), we observe increasing mortality rates across major species. Using ensemble modelling and explainable machine learning, we identify tree size, competition, and seasonal climate anomalies as the primary mortality drivers. Warmer winters and springs are the most frequent and influential climate-related predictors, likely reflecting disturbed phenology and enhanced pest survival during milder cold seasons. Warmer summers are less frequently selected but similarly influential, possibly related to increased pest activity and drought stress. We further identify three drought-related anomaly patterns, each consistent with a distinct mortality pathway: long, intense droughts likely induce acute hydraulic failure; on-average drier summers might drive gradual carbon starvation; and insufficient post-drought rainfall possibly constrains recovery. Anomalously wet springs also frequently emerge as strong predictors of elevated mortality risk, particularly in tall temperate tree species. We hypothesize that this reflects the influence of structural overshoot, whereby wet spring conditions promote rapid canopy growth that may increase vulnerability to subsequent drought. Together, our findings indicate that tree mortality is shaped by interacting and potentially compounding seasonal climate anomalies rather than a single extreme event.