The Mesoscale Analysis and Prediction System (MAPS) is a four-dimensional data assimilation system (implemented operationally at NCEP as the Rapid Update Cycle or RUC) running in real time. The forecast component of MAPS includes a multilevel soil/vegetation scheme for prediction of soil temperatures and volumetric water content. In January of 1997 a snow physics parameterization was added to this soil/vegetation model. This enhancement accounts for the presence of snow on the ground and its melting and subsequent runoff and infiltration of water into the soil.
Soil temperature, soil moisture and snow cover, as predicted
by the soil/vegetation/snow model, are carried forward from the previous forecast to initialize the next forecast in the MAPS cycle. This continuous cycling is desirable because it allows the effects
of abnormally wet or dry soil conditions to positively influence forecasts of the planetary boundary layer through surface fluxes. However, seemingly subtle aspects of the climatology of soil model, and especially the climatology of hydrological cycle components,
such as soil moisture, root zone drainage,surface runoff and snow cover, can dramatically affect atmospheric predictions. Further, in the cold season the existence of frozen moisture in soil considerably changes the processes of heat and moisture transfer inside soil, and also affects all components of the hydrological cycle. This motivated us to further improve the soil/vegetation model by searching for ways of incorporating the effects of frozen soil physics into our soil model. However, the non-linearity and complexity of frozen-soil
physics, particularly the coupling of thermal and hydrological processes, complicates including parameterizations of frozen soil physics into meteorological forecast models, especially those running operationally. We have therefore developed a computationally efficient and very simplified parameterization of processes in frozen soil for incorporation into MAPS. It was first extensivily tested in off-line one-dimensional setting, and subsequently implemented in the three-dimensional forecast scheme. The results from off-line one-dimensional experiments, and the evaluation of the impact on MAPS real-time forecasts, as well as a summary of the scheme itself, will be presented at the meeting.