Investigation of predictability of cotton plant production area soil moisture and temperature values with SAR and optical satellite images

Abstract

In the study carried out in 8 villages and 27 cotton parcels in the Artuklu and Kızıltepe Districts of Mardin Province, data logger devices were installed on the lands. These devices are programmed to record soil temperature and humidity values every 6 hours. The data collected from the data loggers were compared with the Landsat-8 and Sentinel-1 images used by pre-processing in the Google Earth Engine (GEE) cloud environment, and the relationship between them was investigated by analyzing them. A significant and high correlation was found between soil moisture (TN) and Sentinel-1 values, VV (R² = 0.67), VV-VH (R² =0.65), and Landsat-8 SMI (R² = 0.85) values. A significant and high correlation was found between soil temperature (TS) and the Sentinel-1 values of VV (R² = 0.57), VV-VH (R² =0.54), and Landsat-8 SMI (R² = 0.75). In conclusion, it is recommended that the Sentinel-1 VV and VV-VH bands and the Landsat-8 SMI index could be used in soil moisture (TN) and soil temperature (TS) estimation studies, while the Landsat-8 LST band is recommended to be used in larger-scale land areas and regions