The objectives for these data are to characterize seasonal and spatial patterns of drought propagation across a set of seven watersheds in the western United States. These watersheds vary in aridity, contain extensive elevation gradients, and support river systems with long-term gage data. We used a soil water balance model SOILWAT2 with soil physical structure from POLARIS and daily weather data from gridMET as well as DayMet as the inputs. Using gridMET inputs, we simulated daily timestep soil moisture and water balance for dozens to hundreds of 1/24th-degree pixels within each watershed. Using DayMet inputs, we simulated daily timestep soil moisture and water balance for hundreds of 1-km pixels within one watershed. We provide daily potential evapotranspiration, daily transpiration, daily total evaporation, daily diffuse recharge and runoff, daily available soil moisture at different soil depths, and long-term daily frozen soil conditions for each pixel.