Analysis of drought sensitivity in the Pacific Northwest (Washington, Oregon, and Idaho) from 2000 through 2016

This data release includes data-processing scripts, data products, and associated metadata for a remote-sensing based approach to characterize vegetation sensitivity to droughts from 2000 through 2016 in the U.S. states of Washington, Oregon, and Idaho. Drought sensitivity analysis was conducted in minimally-disturbed (‘intact’) forest and shrub-steppe ecosystems, defined as 1-km pixels (i.e., grid cells) that had not experienced major recent insect mortality or fire. Drought conditions were assessed using the multi-scalar standardized precipitation evapotranspiration index (SPEI), for which positive values indicate wetter that average conditions and negative values indicate drier than average conditions for a given site (Vicente-Serrano and others, 2010). A multi-scalar drought sensitivity index (S’) was developed for two drought intensity levels (L): moderate drought (-1.5 < SPEI ≤ -1) and severe drought (SPEI ≤ -1.5). Vegetation response to droughts was quantified using remotely sensed Enhanced Vegetation Index (EVI) from the Moderate-resolution Imaging Spectroradiometer (MODIS) for summer months (June, July, and August) from 2000 through 2016. EVI is a vegetation index calculated from the blue, red, and near-infrared spectral bands representing atmospherically corrected surface reflectance and has advantages over other similar indices in its abilities to represent areas of dense vegetation (Huete and others, 2002). For each pixel, S’ represents the percent decrease in EVI under drought conditions relative to baseline (non-drought, non-pluvial) conditions. Relationships between S’ and a variety of landscape characteristics representing climatic water balance, topography, soil characteristics, and shallow groundwater availability were examined using Boosted Regression Tree (BRT) modeling, a machine-learning algorithm. For detailed descriptions of data-release components, including analysis methods and modeling, please consult the appropriate metadata documents that accompany the processing scripts and data products.