Development and Testing of Improved Techniques for Modeling the Hydrologic Cycle in a Mesoscale Weather Prediction System

Soil-water content is the single most important land-surface variable in atmospheric prediction models. Sophisticated surface physics-soil hydrology parameterization schemes are beginning to be used in mesoscale weather prediction models however, soil-water content is not measured over large enough areas on a regular basis where it could provide suitable initial conditions for those models. Therefore, the initialization of the soil-water content profile has to depend on a knowledge of the hydrological balance of the soil in the area represented by each mesoscale-model grid point. In turn, this information must be obtained from a knowledge of the precipitation, evaporation, and substrate recharge from the water table. We have undertaken the task to develop a systematic means for providing initial values of the soil-water-content profile for the Penn StateNCAR Mesoscale Model Anthes and Warner, 1978. This task is composed of three phases. The first phase is to develop an off line, one- dimensional hydrological model that is driven by conventional meteorological, soil and vegetation data. The second phase is to develop the database to drive the hydrological model in a form that is compatible with the surface physics- soil hydrology parameterization scheme utilized in the mesoscale model i.e., the Biosphere-Atmosphere Transfer Scheme BATS, described by Dickinson et al. 1986. The last phase consists of generating an automated update of the soil- water content profile at each of the mesoscale-model grid points.