Global surface soil moisture data from the Scanning Multichannel Microwave Radiometer (SMMR) for the period 1979-1987 are compared to soil moisture from two different land surface models, the MOSAIC model and the NASA Catchment Model, and assessed against available ground-based measurements. Satellite soil moisture data are retrieved from C-band passive microwaves. The land surface models are driven with a high-quality forcing data set based on re-analysis data and corrected with available observations of precipitation and radiation. Ground observations of soil moisture are only available in select locations in North America and Eurasia. The model, satellite, and ground-based data are entirely independent approaches to estimate soil moisture.

There are encouraging common features in the data sets that are evident in statistically significant correlations of the time series. In many locations, satellite and model soil moisture agree quite well in their seasonal cycles and anomalies. However, there are also many regions with no apparent correlation between the satellite and model soil moisture. Moreover, time average soil moisture and soil moisture variability are markedly different between the data sets in many locations. These discrepancies point to a need for bias estimation and correction or re-scaling before satellite soil moisture can meaningfully be assimilated into land surface models.