Presented at the 2016 International Geoscience and Remote Sensing SymposiumState of the art soil moisture radar retrieval algorithms traditionally depend on substantial amounts of ancillary data, such as land cover and soil texture/composition maps, to parametrize complex electromagnetic models. In this work, we pursue an existing empirical approach as an alternative; it expresses radar backscatter of a vegetated scene as a linear function of soil moisture, thus reducing the dependence on ancillary data. We use 2.5 years of L-band Aquarius radar and radiometer derived soil moisture data to determine the two unknowns of the linear model function on a global scale. We investigate the impact of land cover type by utilizing the widely used IGBP land cover classification; it is found to be significant. We observe seasonal variation in the radar sensitivity to soil moisture, indicating and quantifying seasonally changing vegetation. Finally, we investigate the impact of vegetation heterogeneity within a radar pixel.