The population of sagebrush sparrow in Washington is low. The sparrow is a species of concern because large expanses of big sagebrush (Artemisia tridentata), its preferred habitat, have been lost or degraded.

Sagebrush Sparrow (Artemisiospiza nevadensis) modeled potential distribution in Washington:minimum of 4 model agreement. Shrubsteppe passerine distribution models were developed from presence/absence data collected at 324 sites in eastern Washington, 2014-2019 (Vander Haegen et al.2021). Models of potential distribution were built using R code and an ensemble modeling approach that included 4 statistical models (General Linear, General Additive, Random forest, and Boosted Regression Tree). Spatial data layers used as explanatory variables in the model included fractional vegetation cover (sagebrush, herbaceous vegetation, annual herbaceous vegetation, litter, and bare ground [Young 2017]), elevation, soil depth, and 2 landscape variables: percent unconverted shrubsteppe within a 3km radius, and percent Conservation Reserve Program land within a 3km radius (estimated from the WSDA 2018 Cropland layer). Raster layers of species potential distribution were created using results from all 4 models concurrently, with each cell coded with the number (0-4) of models predicting species presence.
Data for each species are supplied as 4 raster layers, each layer representing concurrence of from 1 to 4 models. More specifically, layers indicating 1 model agreement depict all raster cells where at least 1 of the 4 models predicted presence of that species. These layers are the least conservative of the set and represent the broadest depiction of potential distribution from the modeled data. Layers indicating 2 model agreement depict all raster cells where ≤ 2 models predicted presence of that species. These layers are slightly more conservative than the 1 model layers. Layers indicating 3 model agreement depict all raster cells where ≤ 3 models predicted presence of that species. These layers are slightly more conservative than the 2 model layers. Layers indicating 4 model agreement depict only raster cells where all 4 models predicted presence of that species. These layers are the most conservative of the set and represent the narrowest depiction of potential distribution from the modeled data.
The different model layers, ranging from 1 to 4, can be thought of as representing increasing evidence of habitat suitability for each species as estimated by the models. The 1 and 2 model agreement layers include raster cells with less evidence but are less likely to exclude potential habitat. Conversely, the 3 and 4 model layers include only raster cells with substantial evidence and are less likely to include non-habitat. Users should consider the goal of their analysis or presentation when selecting which layer to use.
Cited Sources
Vander Haegen et al.2021. Shrubsteppe songbird occurrence and distribution in Washington State: a community-science based study. Final report (in preparation)
Young, S. M. 2017. U.S. Geological survey shrub/grass products provide new approach to shrubland monitoring. Report 2017-3084. U.S. Geological Survey, Reston, VA. https://doi.org/10.3133/fs20173084
Natureserve. https://www.natureserve.org/conservation-tools/terrestrial-ecological-systems-united-states
Spatial Extent
Each layer has the same spatial extent determined by the Washington state boundary and by the extent of the fractional vegetation layers used in modeling. The fractional vegetation layer extent appears to include most, if not all, of the land categorized as shrubsteppe or steppe ecological systems in Washington’s Columbia River Basin by Natureserve.
Spatial Data Information
Data type: Raster
Datum: North American 1983 HARN
Projection: NAD83 HARN Washington State Plane South FIPS 4602 Feet
Raster cell size: 180m

Additional 2km buffered sagebrush sparrow location data was supplied to Conservation Biology Insitute by the Washington Department of Fish and Wildife. This location data was merged with the 4 model agreement raster by Conservation Biology Institute.