Satellite remote sensing of climate-driven changes in terrestrial ecosystems continues to improve, yet interpreting and rigorously validating these changes requires extensive ground-truthed data. Satellite measurements of vegetation indices, such as the Normalized Difference Vegetation Index (NDVI, or vegetation greenness), indicate widespread vegetation change in the Arctic that is associated with rapid warming. Plot-based studies have indicated greater vegetation greenness generally corresponds to greater plant biomass and deciduous shrub cover. However, the spatial scale of traditional plot-based sampling is much smaller than the resolution of most satellite imagery and thus does not fully describe how plant characteristics such as structure and taxonomic composition relate to satellite measurements of greenness. To improve interpretation of Landsat measurements of vegetation greenness in the Arctic, we developed and implemented a method that links satellite measurements with ground-based vegetation classifications. Here we describe data collected across the central Brooks Range of Alaska by field sampling hundreds of Landsat pixels per day, with a field campaign total of 23,213 pixels (30 m). Our example dataset shows that vegetation with the greatest Landsat greenness was taller than 1m, woody, and deciduous; vegetation with lower greenness tended to be shorter, evergreen, or non-woody. We also show that understory vegetation influences Landsat greenness. Our methods advance efforts to inform satellite data with ground-based vegetation observations using field samples at spatial scales more closely matched to the resolution of remotely sensed imagery.