Towards hyper-resolution land-surface modeling of surface and root zone soil moisture

Abstract

The goal of this work is to estimate surface and root zone soil moisture at resolutions that are useful for decision making and water resources management. A 500-m atmospheric forcing dataset is developed from the 12.5-km NLDAS-2 (North America Land Data Assimilation System) products across Oklahoma, where high-quality observations are available for validation purposes. A land surface model is then forced with three combinations of input variables to simulate surface and root zone soil moisture: 1) NLDAS-2 atmospheric forcings at their original resolution; 2) downscaled NLDAS-2 atmospheric variables (i.e., near-surface air temperature and humidity, wind speed and direction, incident longwave and shortwave radiation, pressure) and original resolution NLDAS-2 precipitation; and 3) downscaled NLDAS-2 atmospheric variables and precipitation. Results show that the third simulation is able to bring modeled standard-normal deviates of both surface and root zone soil moisture closer to in-situ observations, whereas the second simulation only shows slight improvements with respect to one forced with original resolution NLDAS-2 data. This is particularly evident for negative values of standard-normal deviates, which correspond to drier than usual cases, due to the improved ability of the downscaled precipitation to detect missed events and no-rain cases. In summary, finer resolution forcings have the potential to improve simulations of soil moisture and the resolution of precipitation plays a critical role in improving time series of soil moisture standard-normal deviates.