Estimation of salt affected soils distribution in paddy fields and its relationwith rice yields using Sentinel-2 images

- A case study in Mahasarakham province, Northeast Thailand -

Abstract

Spatiotemporal analysis using satellite imagery data is effective for extracting information about saline soils, which adversely affect the yield of agricultural products. Northeast Thailand has been suffering from the influence of saline soils for many years, and this has created difficulties in the sustainable cultivation of rice, which is a main agricultural product of this country. In the present study, we developed a method for extracting the saline soil areas using salts measurement data from paddy fields as well as time series satellite images taken by Sentinel-2, and we then analyzed the influence of salinity on the rice crop cultivation. In the soil investigation, we measured two items concerned with soil salinity, namely the electric conductivity (EC 1:5) and the hydrogen ion value (pH 1:5), to create a soil EC map and a soil pH map. The classification results showed a total accuracy of 83.67% for soil EC (Kappa coefficient: 0.47) and a total accuracy of 81.63% for soil pH (Kappa coefficient: 0.12). The saline soil areas were then extracted by overlaying both maps. We also investigated the correlation of the statistical yield data and the normalized difference vegetation index (NDVI) for the rice crop cultivation period, to create a rice crop yield map. Comparison of the generated soil salinity map with the rice crop yield map revealed a lower rice crop yield in the saline soil than non-saline soil, and the tendency was accepted remarkably, especially in the areas close to rivers or ponds.