Land use and land cover change (LULCC), predominantly driven by human endeavors such as urbanization and agricultural intensification, has become a significant global problem. The primary aim of this research was to evaluate the effects of LULCC on surface runoff in the Meki watershed. The Google Earth Engine platform’s Random Forest machine learning classifier was used to gather, process, validate, and analyze a variety of satellite photos in order to analyze the rate of LULC changes over four time references, beginning with 1990–2022. Before using the satellite images, preprocessing, classification, and accuracy assessment were performed sequentially. During the four periods from 1990 to 2022, the watershed’s six LULC classes, cultivated land, water, shrub land, grassland, forest, and bare land, were recognized. A significant rate change of LULC was observed in the watershed in each decade. Accordingly, the growth of agricultural land increased from 47.77 to 81.16%, followed by bare land 2.88% to 7.42%. In contrast, over the course of three decades, from 1990 to 2022, the percentages of forest cover, shrub land, and grassland declined sharply from 26.62 to 7.89%, 17.73% to 1.68, and 4.11% to 1.05%, respectively. In order to calculate surface runoff for the Meki watershed, the hydrological Soil and Water Assessment Tool (SWAT) model was set up and parameterized for flow and sediment load. Each LULC scenario’s model calibration and validation are carried out utilizing SWAT-CUP software’s SUFI-2. Model performance statistics, such as R2, NSE, RSR, and PBIAS, as well as model uncertainty metrics, such as p-factor and r-factor, were checked after the model was calibrated and validated. The mean annual surface runoff of the watershed is 117.15, 121.48, 133.93, and 158.84 mm. Accordingly, the Change in LULC from 1990 to 2001, 2001 to 2013, 2013 to 2022, and 1990 to 2022 resulted in an increment of 3.69%, 10.24%, 18.56%, and 35.58% in surface runoff, respectively.
Molla et al. (Tue,) studied this question.