Reconstructed Historical Atmospheric Data by Dynamical Downscaling

Abstract

A 43-year database of reconstructed atmospheric data at 3-km spatial resolution and hourly temporal resolution was created using a nonhydrostatic regional atmospheric model and archived global, coarse-resolution atmospheric data. The archived atmospheric data at approximately 200-km spatial resolution and 6-h temporal resolution were used as initial and boundary conditions in the regional model to generate the reconstructed data. Data for seven variables were extracted from model results for the database: precipitation, atmospheric temperature, relative humidity, wind speed, incoming solar radiation, downward longwave radiation, and latent heat flux. In addition, dew point temperature was computed from simulation results and stored in the database. Model results for temperature and precipitation were compared to observed data collected at three sites in the basin over a 6-year period. Model-generated results match the seasonal, annual, and interannual fluctuations in the observed data quite well. Results indicate that the method for reconstructing atmospheric data for use in distributed parameter hydrologic and environmental models provides spatial detail unmatched by existing observed data networks and is a viable alternative to statistical manipulation of limited observed data.