A Novel Optical Approach to Satellite-Based Remote Sensing of Soil Moisture.

The so-called triangle or trapezoid model is widely applied to estimate near-surface soil moisture from satellite remote sensing (RS) data. A scatter plot of vegetation index (e.g., NDVI) versus land surface temperature (LST) exhibiting a triangular or trapezoidal shape is used to estimate near surface soil moisture for each pixel by comparing its position to the edges of the triangle or trapezoid. A limitation of this approach is that it requires two different sensors (optical and thermal) and thus cannot be applied to satellites that do not supply LST data (e.g., Sentinel-2). In addition, the unique linear relationship between LST and soil moisture, as assumed in the model, is not always guaranteed. To overcome these limitations, we introduce a new optical trapezoid model, based on a recently-developed physically-based model (Sadeghi, M., S.B. Jones and W.D. Philpot. 2015. A Linear Physically-Based Model for Remote Sensing of Soil Moisture using Short Wave Infrared Bands. Remote Sensing Environment) for RS of surface soil moisture using shortwave infrared bands. The new model was applied for estimating near-surface (5 cm) soil moisture for the Walnut Gulch watershed in Arizona and Little Washita watershed in Oklahoma. Sentinel-2 and Landsat-8 data were used to evaluate the optical trapezoid model. Obtained results indicate significant potential of the proposed model for near-surface soil moisture estimation with RMSE of less than 4% for most cases when compared to ground truth data.