Utilization of Landsat-8 Image Classification Results with the Random Forest Algorithm to Determine Groundwater Recharge Areas (Case Study: Singgahan and Montong Districts, Tuban Regency)

Groundwater is highly beneficial for society, and fulfills essential daily living needs and activities. With increasing population growth, groundwater use has increased, necessitating the creation of a map depicting the distribution of groundwater recharge areas. This map is expected to aid the government and communities to identify all the locations of groundwater distribution and ensure the sustainability of groundwater discharge in the area. The method employed involves using Landsat-8 image classification results to produce land cover maps of the study area. Land cover data processing was conducted using the open platform Google Earth Engine with a Random Forest algorithm. The classification of Landsat-8 images resulted in six land cover classes with an overall accuracy of 99.60% and a Kappa value of 0.994. Overlay and weighting were performed using parameters that determine the distribution of groundwater recharge areas. The parameters utilized included supporting data such as rainfall, soil type, and slope, which were obtained from DEMNAS data. The overlay and weighting results indicate that the potential recharge area covered 13,010.33 hectares (59.21%), the transition zone covered 8,935.27 hectares (40.66%), and the discharge area covered 28.64 hectares (0.13%).