Plant species cover and richness data were collected in experimental plots at three sites: one at Badlands National Park (BADL), South Dakota, one at Scotts Bluff National Monument (SCBL), Nebraska, and one at Wind Cave National Park (WICA), South Dakota. At the Badlands and Scotts Bluff sites, data were collected in the summers of 2015 and 2017-2020 in 50 x 50-m plots with one of five treatments: unburned for the entire time of data collection, burned only in fall 2016, burned only in fall 2018, burned in both fall 2016 and fall 2018, or burned then treated with the herbicide imazapic in fall 2016. At the Wind Cave site, data were collected in the summers of 2017-2020 in 10 x 10-m plots in which the herbicide indaziflam was applied either in fall 2017 or spring 2018, with or without previous removal of standing vegetation (mowing and raking). Data were collected using a point-intercept method. At the Badlands and Scotts Bluff sites, this method used 100 points per plot; at the Wind Cave site, this method used 64 points per plot. Species-level data were used to calculate attributes (relative percent cover, diversity) used in the Annual Brome Adaptive Management (ABAM) model. This model is described in the parent publication.