Develope Daily Soil Moisture Maps for Binh Thuan Province Using Multi-Source Remote Sensing Data and Machine Learning Models

Soil moisture is a critical parameter influencing the survival of plants and soil organisms. It is a key climate variable for many hydrological processes as well as in carbon and energy cycles. Therefore, accurate remote monitoring of soil moisture is essential. This study proposes the utilization of multi-source remote sensing data, including CYGNSS, SMAP, and ancillary data from NOAA, to develop a daily soil moisture dataset. The experiment was conducted in Binh Thuan province during the period of 2020-2021. The standardized soil moisture index (SSMI) was calculated based on the obtained soil moisture data to create an agricultural drought risk map. The application of advanced machine learning algorithms enhances the spatial and temporal resolution and accuracy of soil moisture data. The results produced a series of daily soil moisture maps with a resolution of 250 meters, which serve as a foundation for forecasting agricultural drought risk in Binh Thuan province. This study highlights the potential of satellite-derived soil moisture data combined with machine learning techniques for effective soil moisture monitoring. The generated surface-level soil moisture maps can be utilized for daily monitoring, precise yield estimation, and the analysis of significant climate trends.