Abstract Hydrologic model input datasets such as climate, land use, elevation, soil, and geology information are available in a range of scales for use in water resources investigations. Smaller spatial and temporal scale input data allow groundwater recharge models to simulate more physically realistic processes and presumably result in more accurate estimates of groundwater recharge. Projected climate data are, therefore, often downscaled to smaller spatial and temporal scales for use in these models. It is unknown, however, if increasingly smaller-scale climate data produce substantially different simulated recharge results, either in magnitude or trend. Also, even if simulated recharge results are different at a higher space and time resolution, simulation at coarser resolution might be adequate to provide recharge information at decision scales (e.g., meeting Colorado River compact requirements on a ten-year moving average basis). Historical climate datasets at three spatial (∼800 m, ∼4 km, and ∼12 km) and two temporal (daily and monthly) scales were used in a Soil Water Balance (SWB) model of the upper Colorado River basin (UCRB) to simulate groundwater recharge over the water-year 1982–2014 time period. The magnitude of annual and moving ten-year annual average recharge results for daily climate data were within 5% and 7% of ∼4 km results for ∼800 m and ∼12 km climate data, respectively, with deviations from 1982 to 2014 means within 1% and 3% (median), respectively. Comparison of simulated recharge results using the coarsest spatial and temporal climate data with results from the finest scale data indicated similar small differences over ten-year moving annual averages, over water years, and during high recharge months. While differences in simulated groundwater recharge magnitude, which may be important for groundwater-flow simulations, were substantial during some seasonal comparisons, trends in recharge were almost identical across scales, leading to similar conclusions about change from “normal”. Considering the uncertainty inherent in projected climate data, coarser spatial and longer temporal scale input data may be sufficient for water resources managers who need to understand changes in trends in groundwater recharge over water-year or longer time periods.