The Soil Moisture and Ocean Salinity (SMOS) mission has the potential to improve the predictive skill of land surface models through the assimilation of its observations. Several alternate products can be distinguished: the observed brightness temperature (TB) data at coarse scale, indirect estimates of soil moisture (SM) through the inversion of the coarse-scale TB observations, and fine-scale soil moisture through the a priori downscaling of coarse-scale soil moisture. The SMOS TB products include observations over a large range of incidence angles at both H- and V-polarizations, which allows the merit of assimilating the full set of multi-angular/polarization observations, as opposed to specific sub-sets of observations, to be assessed. This study investigates the performance of various observation scenarios with respect to soil moisture and streamflow predictions in the Murray Darling Basin. The observations are assimilated into the Variable Infiltration Capacity (VIC) model, coupled to the Community Microwave Emission Modeling (CMEM) platform, using the Ensemble Kalman filter. The assimilation of these various observation products is assessed under similar realistic assimilation settings, without optimization, and validated by comparison of the modeled soil moisture and streamflow to in situ measurements across the basin. The best results are achieved from assimilation of the coarse-scale SM observations. The reduced improvement using downscaled SM is probably due to a lower number of observations, as a result of cloud cover effects on the downscaling method. The assimilation of TB was found to be a promising alternative, which led to improvements in soil moisture prediction approaching those of the coarse-scale SM assimilation.
All Science Journal Classification (ASJC) codes
- Soil Science
- Computers in Earth Sciences