Water availability, stand structure, and hydraulic trait diversity drive forest stability

Forest productivity, essential for climate change mitigation, is increasingly threatened by environmental fluctuations, particularly droughts. Temporal stability in forest productivity —its ability to remain consistent over time— is influenced by climate, forest structure, community maturity, and tree composition and diversity. However, while the effect of taxonomic composition and diversity on forest stability has been extensively studied, the role of composition and diversity in drought related trait has received less attention. This study explores how drought-tolerance traits, alongside climatic and structural factors affect forest stability between 2000 and 2018 across ∼40,000 forest plots in Spain. We used satellite-derived vegetation indices as proxies for productivity to assess stability, and forest inventory data and species traits to calculate community-level drought tolerance. Water availability was the strongest positive driver of stability, with temporal stability, soil water-holding capacity, and forest maturity and density also contributing positively, though to a lesser extent. Minimum temperature had a small and negative effect. Functional diversity in drought-tolerance traits related to hydraulic safety (P50 and Hydraulic Safety Margin) had a modest but statistically significant and consistent positive effect on stability. Community-weighted mean trait values showed mixed effects. While climate was the dominant factor shaping forest stability, but our findings highlight a relevant role of drought-tolerance diversity, providing a potential mechanism to explain previous findings that drought-tolerance diversity enhances long-term productivity in drought-prone regions. This suggests that fostering drought-related functional diversification could be considered as a potential management strategy to strengthen the stability of forest ecosystem services, especially amid rising climate variability and drought frequency.

R.G.-V. was supported by the project DivTraitFor (2024REGING-20327), JVR-D was supported by a “Juan de la Cierva” fellowship (IJCI-2019-038826-I) and EB is a Serra Húnter fellow. We also received funding from the projects TRACES (PID2021-127452NB-I00), BOMFORES (PID2021-126679OB-I00), RESONATE (European Union's H2020 Programme, grant no. 101000574), and MOTIVATE, funded by the European Commission (GA No. 101052342) and the Spanish PCI2023-2 (MCINN-24-PCI2023-146014-2).

Peer reviewed