Statewide Mapping of California's Forest Stands using Synthetic Tree Data

Globally, the frequency and intensity of wildfires have increased as a result of climate change. Forests in California have been particularly vulnerable as a result of past fire-exclusion practices that have led to denser forests and increased fuel loads. Monitoring and modeling the state and dynamics of forests to help support fuel management and reduce wildfire risk typically requires detailed forest inventory data at the scale of individual trees as inputs. These data, however, are either restricted to field-based plots, limiting the spatial scope of an analysis using these as inputs, or are almost entirely based on remote sensing, limiting the types of metrics and accuracy of the inputs. Here, we evaluated a method by which we could generate "synthetic forest stands" across all of California that estimate individual tree positions and metrics (diameter-at-breast-height, height, and species) and their uncertainties to support forest and fuels modeling. To accomplish this, we developed a spatial point process model using topoclimate and mesoscale satellite remote sensing as spatially continuous covariates, calibrated against field forest inventory data. At each position where the covariates were present, we generated a set of synthetic forest stands, thus allowing us to estimate the range of potential conditions at a given location. Finally, we investigated the degree to which the addition of small footprint aerial LiDAR could reduce uncertainties in these estimations. This approach has the potential to bridge the gap between forest inventories and remote sensing data in a more effective way than has been done previously.