A hybrid (physical-statistical) scheme is developed to resolve the finescale distribution of daily precipitation over complex terrain. The scheme generates precipitation by combining information from the upper-air conditions and from sparsely distributed station measurements; thus, it proceeds in two steps. First, an initial estimate of the precipitation is made using a simplified orographic precipitation model. It is a steady-state, multilayer, and two-dimensional model following the concepts of Rhea. The model is driven by the 2.5?? ?? 2.5?? gridded National Oceanic and Atmospheric Administration-National Centers for Environmental Prediction upper-air profiles, and its parameters are tuned using the observed precipitation structure of the region. Precipitation is generated assuming a forced lifting of the air parcels as they cross the mountain barrier following a straight trajectory. Second, the precipitation is adjusted using errors between derived precipitation and observations from nearby sites. The study area covers the northern half of California, including coastal mountains, central valley, and the Sierra Nevada. The model is run for a 5-km rendition of terrain for days of January-March over the period of 1988-95. A jackknife analysis demonstrates the validity of the approach. The spatial and temporal distributions of the simulated precipitation field agree well with the observed precipitation. Further, a mapping of model performance indices (correlation coefficients, model bias, root-mean-square error, and threat scores) from an array of stations from the region indicates that the model performs satisfactorily in resolving daily precipitation at 5-km resolution.