Artemisia tridentata (Big Sagebrush) habitats that once dominated the North American intermountain west are increasingly rare due to many factors including altered disturbance regimes, increased development, and conversion to cultivated cropland. This species occupies ecological niches ranging from subalpine to semi-arid basins and creates habitat for many sagebrush-obligate species, including the Greater Sage-Grouse (Centrocercus urophasianus). Artemisia tridentata is resilient to the climatic extremes across this gradient of habitats. However, it is a long-lived, relatively slow growing species and, therefore, is less likely to reestablish following disturbances (e.g., increased fire return interval) or in the presence of fast-growing competitors (e.g. Cheatgrass (Bromus tectorum). Many studies have created models of A. tridentata for use in management and conservation of this imperiled species. However, most rely on coarse climatic data and presence/absence points to predict suitable habitat. We have leveraged a unique spatial dataset to predict potential continuous cover for A. tridentata for the western USA. Our results can be used to guide conservation efforts by highlighting regions and climate conditions most likely to support A. tridentataat varying levels of percent cover. Furthermore, this information will support management of Sage-Grouse and other sagebrush obligate species by allowing researchers to locate areas with adequate shrub cover.
Results/Conclusions
Percent cover for A. tridentata were collected from local and regional herbaria, vegetation surveys, and training data for the Regional Gap Analysis Program (REGAP). We compiled Daymet monthly climate data for current conditions (1980-2014) from which we created 19 bioclimatic predictor variables. Highly correlated bioclimatic predictors were removed and A. tridentata points were resampled to match the Daymet resolution (1 km2). We then used the Random Forest algorithm to create maps of A. tridentatacontinuous percent cover throughout the western USA.
We compiled percent cover data for over 16,000 A. tridentata points. The Random Forest model performed well at predicting continuous cover (R2 = 0.91). Within the 2.7 million km2 suitable for any A. tridentata cover, 375,000 km2 (14%) could support up to 10% cover, 1.4 million km2 (50%) could support 11-20% cover, 750,000 km2 (27%) could support 21-30% cover and the remaining 232,000 km2 (9%) could support >30% cover. Previous studies have shown that Sage-Grouse select habitats with intermediate sagebrush cover, so our results can inform prioritization of conservation areas.