Get reliable rangeland science

The spread of invasive annual grasses, such as cheatgrass (Bromus tectorum), can contribute to increased size and frequency of wildfires in western rangelands, posing a significant threat to human use and sensitive wildlife species such as the Greater sage-grouse (Centrocercus urophasianus). Knowledge of the current distribution of cheatgrass is a key component in strategies to mitigate wildfire, reduce the spread of cheatgrass, and protect sage-grouse habitat. We assembled over 24,000 field measurements of cheatgrass cover from various sources to support development of a Generalized Additive Model to predict the spatial extent of cheatgrass across the historic range of sage-grouse in the U.S. (~288 million acres). We focused on identifying areas with >2 percent cheatgrass cover because a strong statistical separation was observed in the field data at this value. A total of 50 climatic, biophysical, and remote sensing variables were examined for statistical significance. Using a forward-stepping selection process, 13 variables were retained for inclusion in the predictive model, including a temporally disaggregated peak NDVI (Normalized Difference Vegetation Index) for years of above-average winter precipitation that was derived to account for regional differences in cheatgrass growth across the broad geographic area. The model correctly classified cheatgrass occurrence at 71% of field measurements reserved for validation. The spatial interpolation of the model indicates that approximately 60.3 million acres within the historic range of sage-grouse has been invaded by cheatgrass, with more extensive invasion in the Great Basin, Columbia Plateau, Snake River Plain, and Colorado Plateau ecoregions. The spatial data of cheatgrass occurrence produced in this study is intended to help inform landscape-scale management of cheatgrass within non-forested rangelands currently and formerly occupied by sage-grouse, and may be useful for guiding more detailed assessments of cheatgrass distribution and abundance in areas of management concern.