Assessing drought-driven mortality trees with physiological process-based models
Résumé
This study describes how physiological process-based models can be used to assess the mortalityrisk of forest trees under global change. Using the CASTANEA model, we simulated the developmentover time of tree functioning with different ontogenetic and phenotypic characteristics (age, diameter,Leaf Area Index, leaf traits) and growing in different site conditions (elevation, soil water content). Basedon this set of simulations, we determined the carbon and hydraulic physiological thresholds associatedwith tree mortality that best reproduce the observed mortality rate. We tested this methodology ona long-lasting and patchy drought-induced mortality event of silver fir (Abies alba Mill.) in South-EasternFrance. (4) We found that tree mortality was not caused by a massive summer xylem embolism, butrather by depletion in carbon reserves probably associated with bark beetle attacks. Simulation outputsalso revealed that trees with high diameter and Leaf Area Index and growing on shallower soils weremore prone to die. (5) This study highlighted that physiological process-based models can be of highinterest to determine the factors predisposing and inducing tree death.