Our objective was to model mean annual number of zero-flow days (days per year) for small streams in the Upper Colorado River Basin under historic hydrologic conditions on small, ungaged streams in the Upper Colorado River Basin. Modeling streamflows is an important tool for understanding landscape-scale drivers of flow and estimating flows where there are no gaged records. We focused our study in the Upper Colorado River Basin, a region that is not only critical for water resources but also projected to experience large future climate shifts toward a drier climate. We used a random forest modeling approach to model the relation between zero-flow days per year on gaged streams (115 gages) and environmental variables. We then projected zero-flow days per year to ungaged reaches in the Upper Colorad River Basin using environmental variables for each raster stream cell in the basin. This data layer shows modeled values for zero-flow days per year of each stream cell.