We evaluate the ability of an augmented version of the Community Land Model (CLM) and a gridded version of the Catchment Land Surface Model (CLSM) to represent monthly changes in soil moisture and groundwater storage from 1996–2005. We drive the land-surface models offline on a 1/8º grid using North American Land Data Assimilation System meteorological forcing data. The version of CLM used here represents water-table dynamics using a one-dimensional, lumped, unconfined aquifer model; CLSM implicitly represents variation in water-table depth by tracking changes in the catchment water deficit.

Soil moisture and water table measurements obtained by the Illinois State Water Survey provide observations are used to assess the models' abilities to models simulate monthly changes in total soil-column water storage. We explore the degree to which each model realistically partitions the total change in terrestrial water storage between subsurface reservoirs. We draw preliminary conclusions regarding changes in model skill between seasons, interannually, and between soil locations. Results presented here are the first use of ground-based observations to evaluate the performance of high-resolution, regional gridded land-surface models that contain a representation of water-table dynamics.