Macroscale hydrological modeling using remotely sensed inputs: Applicationto the Ohio River basin

Predictions of water and energy budgets at the land surface are central to climate simulation and numerical weather prediction, as well as to water re sources planning and management. Macroscale hydrological models provide a n ew tool for simulating surface water and energy balances at the scale of la rge continental river basins, However, these models are limited by the scar city of in situ meteorological forcing data. Remote sensing data provide an alternative to in situ data, with observations that are, in some cases, at a higher spatial and temporal resolution than those available from traditi onal surface sources. Nonetheless, there remain important questions as to w hether the accuracy of remotely sensed surface variables is sufficient to s erve as forcings for surface hydrological models. This question is addresse d through comparison of hydrologic simulations for the Ohio River basin wit h the variable infiltration capacity (VIC) macroscale hydrology model, usin g in situ and remotely sensed data. In situ data consist of gridded (at 1/2 degree latitude-longitude spatial resolution) precipitation, temperature, and wind, with downward solar and longwave radiation inferred from the diur nal temperature range. Remotely sensed observations include incident solar radiation, air temperature, and vapor pressure deficit inferred from the Ge ostationary Operational Environmental Satellite (GOES), the advanced very h igh resolution radiometer (AVHRR) and the TIROS Operational Vertical Sounde r (TOVS), respectively. Precipitation, in all cases, is from gridded statio n data. The modeled streamflows and evapotranspiration rates are quite simi lar for the two cases. The largest differences in predicted surface hydrolo gy are associated with differences in modeled snow cover accumulation and s nowmelt, and result from a warm bias in the remotely sensed temperature dat a.