The predictability of the relationship between the land surface states and precipitation patterns over the continental U.S. is investigated within the context of the North American Monsoon System (NAMS). This relationship is explored using a 1/8° hourly resolution retrospective run from 1997-2001, which was generated using the North American Land Data Assimilation System (NLDAS)-Mosaic Land Surface Model (LSM). The precipitation data, used as forcing by the LSM, is derived from a merging of ETA model, hourly StageIV Doppler radar, and daily rain gauge precipitation products, which do include some observations from Mexico. Also, a 50-year retrospective Variable Infiltration Capacity (VIC) LSM run of the same NLDAS spatial resolution, developed at the University of Washington (Maurer et al., 2002), is used for additional statistical purposes and to build a baseline climatology to compare our shorter Mosaic simulation. Extending the correlation analyses over Mexico is done with Globlal LDAS (GLDAS)-Mosaic LSM simulations forced with atmospheric GCM and satellite datasets.

Based on differences between June and July precipitation climatologies, a significant shift in precipitation is found to occur over North America that affects the amount of soil moisture content in the Southwest and the central Great Plains, and thus latent heat and other surface fluxes during the rest of summer. To investigate this, lag correlation studies are performed for different timescales between soil moisture of both Arizona and New Mexico and precipitation over the contiguous U.S. Also, different spatial scales are compared to investigate more closely the role of topography on precipitation and soil moisture and how they behave locally (e.g., within the state of Arizona) and are influenced by remote moisture sources (e.g., Gulf of Mexico), and relating this to different temporal scales. Intraseasonal variability for both a wet and dry monsoon year is examined between Arizona and New Mexico soil moisture daily averages and Oklahoma and Texas precipitation. Both Arizona and New Mexico have negative correlations with Texas for the wet monsoon year, but weaker relationships are found in the dry year. On the interseasonal timescale, the land surface conditions that exist in the antecedent and subsequent months surrounding the NAMS months are found to relate well to anomalously wet and dry monsoon years. These and additional relationships associated with the different timescales will be presented.