Soil moisture mapping in Indian tropical islands with C-band SAR and artificial neural network models.

This study aims at analyzing the patterns of soil moisture in the South Andaman district using an integrated approach that incorporates Sentinel-1A C-band synthetic aperture radar (SAR) data and other auxiliary data from Sentinel-2A and Landsat 8. A total of 60 surface soil samples (0-10 cm) were collected from four predominant land uses for 2020-2022 years to represent real-time soil moisture status. Soil moisture index (SMI) is assessed based on thermal remote sensing data besides, normalized difference vegetation index (NDVI) from red and infrared bands, and dielectric constants (ε) from soil textural analysis. Artificial neural network (ANN) models were developed along with multiple linear regression (MLR) to retrieve the soil moisture accurately using input parameters such as backscatter coefficients (σ°: VV and VH), NDVI, SMI, and ε. The performance of modelled soil moisture is evaluated using different statistical index-based criteria concerning field-based volumetric soil moisture measurements (SMCv). It is found that positive correlation among (σ°: VV + VH) and (SMCv: %) for all land uses and high R values for barren and vegetable fields. The vegetation interferes the backscatter signal and misinterprets the soil moisture estimation solely with only SAR data. However, consideration of NDVI and SMI improves the soil moisture estimation in case of vegetation abundance land uses. The comparative results showed that ANN models surpass MLR models in soil moisture estimation with high R (0.67-0.99) and η (62.6-99.9) and low RMSE (0.05-2.19%) and MAE (0.03-1.74%) values. By providing essential baseline data for hydrological modeling, this study supports the design of efficient irrigation systems.