PNNL

Abstract

Tropical cyclones are important disturbances in coastal tropical and subtropical forests. In the past three decades, Puerto Rico experienced five hurricanes that met or exceeded category three; these major hurricanes caused severe forest structural damage and elevated tree mortality. To assess how well pre- and post-hurricane forest function and structure are represented in Earth system models, we used in-situ forest measurements at the Bisley Experimental Watersheds in the Luquillo Experimental Forest of Northeast Puerto Rico to evaluate the Functionally Assembled Terrestrial Ecosystem Simulator, which was coupled with the Energy Exascale Earth System Model Land Model (ELM-FATES). By using observational metrics such as aboveground biomass (AGB) and stem density of tree size classes determined by diameter at breast height, as well as trait observations to provide the ranges of values, we constrain key parameters (e.g., Vcmax, wood density) of ELM-FATES. We then perform numerical experiments that simulate the damage of two hurricanes, Hugo and Georges, on the Bisley forests, including defoliation, sapwood and structural biomass damage, and hurricane-induced mortality. The results show that ELM-FATES can reasonably represent the AGB partitions between the two plant functional types – light-demanding and shade-tolerant – for both the pre-Hugo and post-Hugo conditions. The model also represents a reasonable size distribution of trees although underestimates total AGB. Compared to observations, the overestimates of mortality could explain the underestimation of AGB. This research underscores the importance of evaluating dynamic global vegetation demographic models in representing hurricane disturbance and recovery processes with a broad range of ecosystem measurements.