In 2010, the Pacific Region of the U.S. Fish and Wildlife Service (Service) initiated a native bee sampling program to document the bee fauna at 15 National Wildlife Refuges (NWR), 1 National Monument, and 1 Bureau of Land Management site in Washington, Oregon, and Idaho (hereafter referred to as Refuges or Sites). This program was comprised of six sampling components, utilizing two methodologies depending on the available staff, site objectives, and accessibility. For purposes of data publication, these components are summarized as follows. Component 1 (Array Sampling) was designed to collect baseline data within shrub-steppe habitats along a north-south gradient in Washington and Oregon. A secondary objective of this sampling was to determine if there were genera and/or species differences along this latitudinal gradient. As there were no wildlife refuges in northeastern Oregon, the Bureau of Land Management’s Oregon Trail Interpretive Center (OTIC) was selected to fill this gap. Refuges included in this project include Turnbull NWR in NE Washington, McNary NWR in central Washington, and Malheur NWR in SE Oregon; the OTIC site is located in northeastern Oregon. All sites were sampled during 2011 and 2012, using a 9-cup glycol array (see Methodology). This was the only component in this program to use glycol arrays. Sampling sites were situated in easily accessible shrub-steppe habitats typical of the specific NWR and OTIC. Component 2 (Baseline Sampling) was designed to collect baseline data at a suite of NWRs and was the primary objective and sampling scheme used from 2011-2016. Refuges within this component used a variation of the standard 30-bowl (pan trap) transect, instead deploying 2-15 bowl transects set adjacent to different habitats to increase the prospects of increasing catch diversity and attaining a better representation of bee fauna on the refuge (see Methodology). Depending on the site, sampling occurred: for two consecutive seasons using the same transect location; 1 season of sampling only; incidental sampling for 1 or 2 seasons but without fixed locations between sampling events; and habitat-based plots to assess variability in succession.