Google Earth Engine Developers

The methodology is too complicated for calculating soil moisture from SAR images. The simplest solution is having sparse soil moisture stations over the ROI. Then you can use SAR backscatter around the stations along with the ground measurements of soil moisture to train a machine learning model like random forest. It is important to have at least one station inside each crop type or land cover type over time. The RF will understand spatial and temporal trends per each class of land cover. Then the RF will do the soil moisture estimation for you.