MICROWAVE REMOTE SENSING BASED EVALUATION OF WATERSHED MANAGEMENT IMPACT ON SOIL MOISTURE IN WUTAME MICRO WATERSHE

Abstract

The effectiveness of watershed management practices on changing hydrological variables like soil moisture content is not well explored in the scientific literature. Remote sensing is an alternative data source to fill the gap of conventional methods for measuring soil moisture. However, remote sensing products and techniques also need to be evaluated with ground measurements. In this study, the impact of conservation measures on soil moisture content was evaluated by integrating remote sensing with in situ soil moisture content measurements. The approach started by measuring the soil moisture content using both direct and indirect methods. Fifteen soil moisture content samples were collected from the downstream, middle stream and upstream parts of the micro watershed. Additionally, the corresponding measurements were conducted using the Delta T HH2 moisture meter. Then the gravimetric method was used to calibrate the Delta T HH2 moisture meter. A uniform sampling technique was used to collect soil moisture content measurements, used to characterize its distribution by conventional methods. Therefore, this sampling technique brought 57 sampling sites over the micro watershed. Next, these 57 samples were used to spatially interpolate. An empirical (Multiple Regression Analysis) relation was established for remote sensing-based soil moisture using 45 measurements for calibration and 15 measurements for validation. The soil moisture content was served as a dependent variable, while the backscattering coefficient and local incidence angle of Sentinel 1A were served as independent variables. The performance of the developed model showed different accuracy at downstream, middle stream and upstream parts of the micro watershed with R 2 = (0.32, 0.80 and 0.82), root mean square error (8.88 %, 6.26 % and 7.09%) and mean absolute error (6.44 %, 5.0 % and 5.48 %), respectively. The result of this study showed that the Sentinel 1A satellite can be used to estimate soil moisture content. Therefore, the use of Sentinel 1A satellite product for soil moisture estimation is promising and hence future planning, development and decisions can be based on it. The result from both direct and indirect methods showed that the spatiotemporal dynamics of the soil moisture content is highly affected by the rainfall event and topographic conditions. Moreover, the result showed that there is a significant change in soil moisture due to the sustainable land management interventions in Wutame micro watershed.