Improving ecosystem modeling for realistic forest-management and wildfire simulation across California's Sierra Nevada

Mountain ecosystems typically serve as carbon sinks, yet, studies also suggest they could be carbon sources due to climate warming, drought, wildfires and management actions. To investigate the critical role of productive forests in carbon storage in a Mediterranean climate, we applied a process-based dynamic vegetation-ecosystem model (LPJ-GUESS) with parameterization constrained by local measurements, including realistic thinning (different size classes, fractions, and vegetation types), realistic fire (new LMFIRE module considered fire occurrence with additional wind speed, lightning and human ignition), plus projected CO2-fertilization effects and measured root-zone-accessible-water values. The model was calibrated and evaluated using MODIS-based GPP (PML-v2) and flux-tower ET, resulting in water-carbon feedbacks and underlying drivers being better represented. Results showed that improving LPJ-GUESS provided useful scenarios for guiding decision making. That is, modeling results go beyond providing process insights to successfully representing multi-year drought effects on both ecosystems' carbon and water fluxes (GPP and ET), and for high-severity wildfires (Starr Fire in 2001 and King Fire in 2014). Sensitivity analyses combined with thinning, accompanied by removing or adding dead ground fuel, showed that ecosystem carbon will recover in 10 and 30 years under low intensity (5 inches, 10% canopy cover removal) and medium intensity (10 inches, 40% canopy cover removal) management actions for fuels reduction. Possible doubling of the biomass-sequestration rate and halving the soil-organic-carbon release rate through various management actions aids in exploring mechanisms as well as vulnerability and resilience in Mediterranean mountain ecosystems. Improving LPJ-GUESS proved realistic simulations to help generalize findings across the Sierra Nevada range and explore management benefits in a real world.