Mapping Cheatgrass Dieoff in the Northern Great Basin using Ecosystem Performance Modeling

The dataset contains a time series of cheatgrass dieoff maps covering the western and central areas of the northern Great Basin. This dataset maps annual cheatgrass performance in the northern Great Basin, USA, from 2000-2010, using the dynamic monitoring of ecosystem performance (DMEP) method. The power of DMEP is that it separates weather-related performance issues from disturbance- and management-related performance issues by integrating satellite data with land cover data, seasonal weather data, and a site potential dataset into a regression-tree model. The resulting model is called the expected cheatgrass performance (ECP) model, and this model performed well (R-square equaled 0.88). The EEP model results are statistically compared to a dataset that serves as proxy for actual cheatgrass performance (see the metadata for the cheatgrass cover dataset for a full explanation), and at the 80% confidence interval, normal performing cheatgrass areas are differentiated from both underperforming and overperforming cheatgrass areas. The underperforming cheatgrass areas represent potential cheatgrass dieoff. Dieoff areas concern land managers, scientists, and policy makers because they allow the invasion of new weed species, accelerate soil erosion, reduce early spring forage for livestock and wildlife, and present unknown recovery pathways.