Precipitation over a given basin can be modulated by the interaction with land surface processes. In turn, soil moisture depends on precipitation and acts as a strong control of the partitioning between sensible heat flux and latent heat flux (Bowen ratio). The importance of these interactions, or feedbacks, has become increasingly evident in the last decades, when numerical simulations were employed to investigate the sensitivity of precipitation to land surface characteristics (including soil moisture and vegetation). Moreover, the link between surface states and the atmospheric hydrologic cycle involves the atmospheric boundary layer: it has been shown that the boundary layer variables, like equivalent potential temperature and specific humidity, depend at least in part on the underlying soil moisture. For example, wet surface conditions force larger equivalent potential temperature, and favor the potential for increased precipitation. Therefore, these interactions that involve complex feedbacks may lead to persistent wet or dry spells.

The National Centers for Environmental Prediction (NCEP) Eta model and its Eta Data Assimilation System (EDAS) have been successfully used for regional studies of the hydrologic cycle over the Mississippi basin. We have compiled a data set that exceeds seven years, and a climatology of the land surface-water cycle interactions was prepared. Because the Eta model is operational, its changes throughout the years may have affected the output. Nevertheless, we have not detected significant changes in the atmospheric water cycle, although variables that depend directly from model parameterizations, like evaporation, are more likely to be affected. Probably the most significant progress with the Eta model was achieved with the implementation of soil moisture cycling in mid-1998. Therefore our presentation will have two parts: the first one describes the atmospheric water cycle for the seven year period, its stability and variability. The particular effect of model changes will be addressed. The second part, based on approximately four years (1998-2002), will focus on the relationships between the land surface and the atmospheric water cycle.