An extension of Sinclair's diagnostic model of orographic precipitation (“VDEL”) is developed for use in data-poor regions to enhance rainfall estimates. This extension (VDELB) combines a 2D linearized internal gravity wave calculation with the dot product of the terrain gradient and surface wind to approximate terrain-induced vertical velocity profiles. Slope, wind speed, and stability determine the velocity profile, with either sinusoidal or vertically decaying (evanescent) solutions possible. These velocity profiles replace the parameterized functions in the original VDEL, creating VDELB, a diagnostic accounting for buoyancy effects. A further extension (VDELB*) uses an on/off constraint derived from reanalysis precipitation fields. A validation study over 365 days in the Pacific Northwest suggests that VDELB* can best capture seasonal and geographic variations. A new statistical data-fusion technique is presented and is used to combine VDELB*, reanalysis, and satellite rainfall estimates in southern Africa. The technique, matched filter regression (MFR), sets the variance of the predictors equal to their squared correlation with observed gauge data and predicts rainfall based on the first principal component of the combined data. In the test presented here, mean absolute errors from the MFR technique were 35% lower than the satellite estimates alone. VDELB assumes a linear solution to the wave equations and a Boussinesq atmosphere, and it may give unrealistic responses under extreme conditions. Nonetheless, the results presented here suggest that diagnostic models, driven by reanalysis data, can be used to improve satellite rainfall estimates in data-sparse regions.