Rangeland Fractional Cover

Rangeland fractional cover refers to the percentage of a pixel that is covered by plants that belong to different functional groups found in the arid and semi-arid ecosystems of the Columbia Basin. These fractions vary widely as a function of succession after disturbances like wildfire, and often have profound influences on habitat quality for shrubsteppe-associated species. TerrAdapt currently models fractional cover of four functional groups, including the percent of sagebrush cover, the percent of shrub cover, the percent of perennial grass cover, and the percent of invasive annual grass cover.

We used US Bureau of Land Management (BLM) Assessment, Inventory, and Monitoring (AIM) field data as empirical observations of fractional rangeland vegetation cover. Specifically, we extracted the values of sagebrush (AH_SagebrushCover_Live), shrub (AH_NonNoxShrubCover), perennial grass (AH_NonNoxPerenGrassCover), and invasive annual grass cover (AH_NoxAnnGrassCover) for all locations within our modeling boundary. In total, our training dataset contained field plot measurements from 932 locations spanning the years 2015-2020.

TerrAdapt's rangeland fractional cover models use a variety of covariates as environmental predictors. The data sources for these covariates are described in the DATA INPUTS section. All covariates were stored in Google Earth Engine (either in the public data catalog or TerrAdapt's private asset storage) and available for use in our dynamic workflow.

We used the random forest classifier (Breiman 2001) implemented in Google Earth Engine to model fractional cover of sagebrush, shrubs, perennial grasses, and annual grasses from the covariates at all training data locations. We trained 10 replicate models in regression mode per variable in a k-fold approach (Fushiki 2011) with a unique 'fold' of 1/10th of the data withheld for validation. The final model was calculated as the average of the 10 replicates.