Due to the lack of long-term consistent soil moisture analysis, numerical studies of the impact of soil moisture on medium to seasonal range forecasts are often based on extreme or idealized conditions. In this study, the atmospheric predictability at seasonal scale is investigated using soil moisture analyses that are more realistic and consistent with the forecast model. This is accomplished using the Air Force Weather Agency (AFWA) Agricultural Meteorology modeling system (AGRMET), an operational global database of land surface states and energy/water fluxes, and the Medium Range Forecast Model (MRF), a state-of-the-art general circulation model.

AFWA incorporated the NCEP community NOAH Land Surface Model (NOAH LSM) into AGRMET in late 1999. The soil hydrology physics are forced with analyses of shelter height temperature, relative humidity, and wind speed, short and longwave radition, and precipitation. As part of the efforts to unify land model in all NCEP global and regional models, NOAH LSM has been implemented into MRF in 2002. As AGRMET land states have spun up using same land physics that MRF executes, they provide ideal source of initial land states that are strictly self consistent with MRF land physics.

Two sets of ensemble integration of atmospheric model will be performed, one using climatological soil wetness derived from the NCEP/DOE Reanalysis (R-2) and the other using AGRMET soil wetness analysis as initial conditions. The geographical variations of the predictability of soil wetness, precipitation, and near surface temperature will be examined from this dataset.