부산대학교 도서관

Tropical cyclones are an important cause of forest disturbance, and major storms caused severe structural damage and elevated tree mortality in coastal tropical forests. Model capabilities that can be used to understand post‐hurricane forest recovery are still limited. We use a vegetation demography model, the Functionally Assembled Terrestrial Ecosystem Simulator, coupled with the Energy Exascale Earth System Model Land Model (ELM‐FATES) to study the processes and the key factors regulating post‐hurricane forest recovery. We implemented hurricane‐induced forest damage, including defoliation, structural biomass reduction, and tree mortality, performed ensemble model simulations, and used random forest feature importance. For the simulation in the Luquillo Experimental Forest, Puerto Rico, we identified factors controlling the post‐hurricane forest recovery, and quantified the sensitivity of key model parameters to the post‐hurricane forest recovery. The results indicate a tendency for the Bisley forests to shift toward the light demanding plant functional type (PFT) when the pre‐hurricane biomass between the light demanding and shade tolerant PFTs is nearly equal and forests experience hurricane disturbance with mortality >60% for both the two PFTs. Under more realistic conditions where the shade tolerant PFT is initially dominant, mortality >80% is required for a shift toward dominance of the light demanding PFT at Bisley. Hurricane mortality and background mortality are the two major factors regulating post‐hurricane forest recovery in simulations. This research improves understanding of the ELM‐FATES model behavior associated with hurricane disturbance and provides guidance for dynamic vegetation model development in representing hurricane induced forest damage with varied intensities. Plain Language Summary: To enhance the understanding of forest recovery after hurricanes, we implemented hurricane induced forest damage into the Functionally Assembled Terrestrial Ecosystem Simulator, coupled with the Energy Exascale Earth System Model Land Model (ELM‐FATES). We performed ensemble ELM‐FATES simulations with varied forest damage intensities in the Luquillo Experimental Forest, Puerto Rico, and used the output to identify factors controlling the post‐hurricane forest recovery, which was further evaluated with random forest feature importance (RFFI) that quantifies the sensitivity of the key model parameters to the post‐hurricane forest recovery. The results imply that hurricane mortality and background mortality are the major factors regulating post‐hurricane forest recovery. Changes to the intensity of simulated hurricanes could alter forest composition and structure during recovery, which modifies forest ecological processes and potentially shift the wet forests in Puerto Rico to states with increased vulnerability to tropical cyclones. This research enhances our understanding of the ELM‐FATES model behavior associated with hurricane disturbance and broadens the application of RFFI in quantifying the parameter sensitivity of a dynamic global vegetation model (DGVM). This research addresses the essential role of representing hurricane induced forest damage in DGVMs, an advanced tool for the future studies of tropical forest dynamics. Key Points: Post‐hurricane forest recovery favors the light demanding type when hurricane mortality rates are equal to or higher than 60% at BisleyHurricane mortality and background mortality are the key factors regulating the post‐hurricane forest recovery and biomass compositionELM‐FATES simulations at a Puerto Rico forest site can represent reasonable GPP and ET seasonality but the flux magnitudes are biased low [ABSTRACT FROM AUTHOR]