Remote sensing techniques for soil humidity monitoring in drought areas: case study of the Wadi Hallouf/Oum Zessar watershed (Tunisia)

This study aimed at monitoring the soil moisture in drought areas in southern Tunisia by combining multispectral and SAR images. A set of vegetation and humidity indices derived by multispectral images have been produced to check which of them had a better correspondence to rainfall events. The visual comparison shows that the NDWI and NMDI indices generally have a lower sensitivity to soil humidity changes, while NDMI and NDVI denote a significant increase when SMC increases, especially in presence of vegetation. On the other hand, the VSDI index has a more than good response equally distributed regardless of the type of land use. Results have been then organized to obtain statistical correlations between optical and combined SAR-multispectral indices and cumulative precipitation recorded in 10 meteorological stations located in the study area. The maps of these indices have then been analyzed to highlight which optical index had the best correlation with SMC values; a Combined (microwave + optical/infrared) Drought Severity Index (CDSl) was then calculated and correlations with SMC values were analyzed.