Disaggregation of rainfall for one-way coupling of atmospheric and hydrological models in regions of complex terrain

The objective of this work is to incorporate orographic effects in the disaggregation of rainfall fields from atmospheric to hydrological models. For this purpose, a downscaling methodology based on the principles of fractal interpolation is proposed. The new orographic rainfall disaggregation scheme takes into account the spatial characteristics of topography and wind fields over the domain, and how these relate to the spatial variability of rainfall in the region. We illustrate the methodology for two simulated precipitation events in Pennsylvania, and the results are compared against observations at 200 raingauges within the region. Disaggregation of mesoscale meteorological model (MM5) rainfall fields from 4-km down to 1-km resolution using the orographic scheme resulted in average improvements of quantitative precipitation forecasts about 50% of total precipitation amounts. An overall improvement in the temporal correlation between observed and predicted hourly rainfall was also obtained (on the average, the correlation coefficient increased from 0.2 to 0.5), thus suggesting that the methodology may be used effectively to improve the timing of precipitation forecasts. The effect of spatial variability of precipitation on the hydrologic response of a large watershed was also assessed. In particular, MM5 output and disaggregated rainfall fields were used to force a distributed hydrological model of the watershed of the West Branch of the Susquehanna River (14,710 km2 areal extent). Improvements of 30% were obtained in the prediction of peak streamflow and runoff volumes just by including small-scale (1 km2) orographic effects on the space–time variability of rainfall. The intrinsic relationship between rainfall forcing and “optimal” hydrological parameters obtained through calibration is discussed.