Hydrologic variables such as evapotranspiration (ET) and soil water storage are difficult to observe across spatial scales in complex terrain. Streamflow and lidar‐derived snow observations provide information about distributed hydrologic processes such as snowmelt, infiltration, and storage. We use a distributed streamflow data set across eight basins in the upper Tuolumne River region of Yosemite National Park in the Sierra Nevada mountain range, and the NASA Airborne Snow Observatory (ASO) lidar‐derived snow data set over 3 years (2013–2015) during a prolonged drought in California, to estimate basin‐scale water balance components. We compare snowmelt and cumulative precipitation over periods from the ASO flight to the end of the water year against cumulative streamflow observations. The basin water balance residual term (snow melt plus precipitation minus streamflow) is calculated for each basin and year. Using soil moisture observations and hydrologic model simulations, we show that the residual term represents short‐term changes in basin water storage over the snowmelt season, but that over the period from peak snow water equivalent (SWE) to the end of summer, it represents cumulative basin‐mean ET. Warm‐season ET estimated from this approach is 168 (85–252 at 95% confidence), 162 (0–326) and 191 (48–334) mm averaged across the basins in 2013, 2014, and 2015, respectively. These values are lower than previous full‐year and point ET estimates in the Sierra Nevada, potentially reflecting reduced ET during drought, the effects of spatial variability, and the part‐year time period. Using streamflow and ASO snow observations, we quantify spatially‐distributed hydrologic processes otherwise difficult to observe.