Northern latitude forests account for about 30% of the terrestrial carbon sink, providing critical carbon storage to offset global fossil fuel emissions. Yet, increases in the frequency of high-severity wildfires and climate change may threaten the ability of northern forests to maintain high levels of productivity and postfire regeneration, and thus weaken carbon uptake. Dynamic global vegetation models (DGVMs) that simulate forest dynamics are often used to estimate terrestrial carbon sinks for the global carbon budget and make projections about future changes. Thus, uncertainty about the resiliency of northern forests requires further development of DGVMs to capture plant geography, biomass turnover, and forest regrowth to better estimate carbon dynamics.
Here we present development of LPJ-GUESS-LMfireCF, an ecosystem simulation model that combines the DGVM LPJ-GUESS with the fire module LMfire and a newly developed crown fire routine (CF). Historic simulations were run for the ecosystems of Yellowstone National Park (YNP) to evaluate model performance for simulating landscape biomass distribution, dominant plant cover, fire activity, and forest regeneration.
Results/Conclusions
Biomass estimates from the newly parameterized regional plant functional types (PFTs) and the LMfireCF module (mean 39.8 t ha-1) improved by 581% over simulations with the standard parameterization of LPJ-GUESS (mean 200.4 t ha-1), compared to satellite-based estimates (mean 27.6 t ha-1). LPJ-GUESS-LMfireCF realistically captured vegetational zones with elevation and climate gradients in YNP, compared to the National Park Service’s Yellowstone primary cover types. Simulated area burned and fire severity approximated satellite-derived observations. Importantly, LPJ-GUESS-LMfireCF simulated the large stand-replacing fires of 1988 in Yellowstone as emergent results without model initialization of vegetation cover or fire history. Simulated postfire regrowth was more rapid than field-based estimations, but could be improved by further calibration of the regional PFT parameters. In summary, LPJ-GUESS-LMfireCF serves as a useful tool for regional to global scale simulations that distinguishes crown and surface fires to more accurately simulate plant geography and biomass turnover. Applying LPJ-GUESS-LMfireCF to northern forests of the globe could provide valuable projections of fire regime changes and forest responses under future climate scenarios to predict changes to the global carbon cycle.