A one-dimensional mass and energy balance model was developed to simulate rainfall interception in Sacramento County, California. The model describes tree interception processes: gross precipitation, leaf drip, stem flow, and evaporation. Kriging was used to extend existing meteorological point data over the region. Regional land use/ land cover and tree canopy cover were parameterized with data obtained by remote sensing and ground sampling. Annual interception was 1.1% for the entire county and 11.1% of precipitation falling on the urban forest canopy. Summer interception at the urban forest canopy level was 36% for an urban forest stand dominated by large, broadleaf evergreens and conifers (leaf area index = 6.1) and 18% for a stand dominated by medium-sized conifers and broadleaf deciduous trees (leaf area index = 3.7). For 5 precipitation events with return frequencies ranging from 2 to 200 years, interception was greatest for small storms and least for large storms. Because small storms are responsible for most pollutant washout, urban forests are likely to produce greater benefits through water quality protection than through flood control.