Spatial Downscaling of Gridded Soil Moisture Products Using Optical and Thermal Satellite Data: Effect of Using Different Vegetation Indices

Satellite remote sensing offers global-scale soil moisture (SM) estimation to assess water and energy cycles. However, the coarse resolution of SM products from microwave remote sensing is unsuitable for fine-scale analysis. This study explored spatial downscaling methods to refine the 0.25° ESA CCI SM product to a 1-km resolution, utilizing optical and thermal remote sensing data, including the normalized difference vegetation index (NDVI), enhanced vegetation index (EVI), kernel NDVI (kNDVI), and plant phenology index (PPI), together with land surface temperature from MODIS products over two study areas in Europe. The vegetation temperature condition index based approach was used for downscaling, in which the wet and dry edges of the triangular feature space were determined by fitting a line to the maximum and minimum temperatures, respectively, for each vegetation index. The PPI-based downscaling showed consistent results between the two study areas, having a good correlation coefficient and unbiased root-mean-square deviation (ubRMSD) against the in-situ measurements. The NDVI-based downscaling had poor performance overall in terms of ubRMSD and correlation. Results from the EVI- and kNDVI-based methods varied in the two study areas. Compared with the original coarse SM product, spatially downscaled SM products exhibited inferior performance against in-situ SM measurements in terms of evaluation metrics.