High-Resolution Soil Moisture Mapping using MSAVI-LST based Triangle Method

Abstract. High-resolution soil moisture information is an essential parameter for hydrological, agricultural, and climatic processes. However, existing satellite products, such as SMAP, provide soil moisture at coarse resolution (9-36 km), limiting their regional utility. To address this, various studies have introduced downscaling techniques using optical and thermal data, but accounting for vegetation modulation remains a challenge. Therefore, in this study, an improvised downscaling approach based on the ‘Triangle method’ is used to estimate satellite soil moisture at 1 km spatial resolution from 9 km through the SMAP soil moisture product. Traditionally, this method uses the normalized difference vegetation index (NDVI) and Land Surface Temperature (LST) as an input for the downscaling approach. Here, the Modified Soil Adjusted Vegetation Index (MSAVI) is integrated as an alternative vegetation index to improve the vegetation sensitivity. The Triangle method has been used with first and second order of polynomial regression (1-1,1- 2,2-1,2-2) relations between normalized vegetation index (VI*) and land surface temperature (LST*) to generate the soil moisture trapezoid for both NDVI and MSAVI. The validation of this methodology is carried out across different regions of Varanasi, on twelve dates of (2019-20,2023,2024,2025). Results indicate that the 1-1 polynomial order with MSAVI significantly outperforms the other combinations as well as the traditional NDVI-based approach. The highest correlation coefficient (R) of 0.76 with a minimum RMSE of 0.032 m³/m³ using MSAVI provides valuable insights for improving the downscaling technique of soil moisture in heterogeneous landscapes.