This dataset contains a comprehensive collection of geospatial raster data (GeoTIFF format) detailing the spatio-temporal impacts of land cover change in the Paguyaman Watershed, Gorontalo, Indonesia, for the years 2016, 2020, and 2024. The data was generated as part of a doctoral dissertation at IPB University. The dataset includes annual maps for four key sustainability indicators:1. Erosion Potential: Estimated using the RUSLE model (tons/ha/year).2. Surface Runoff Potential: Estimated using the SCS-CN model for a 100 mm/24-hour rainfall scenario (runoff depth in mm).3. Land Productivity: Estimated using the Normalized Difference Vegetation Index (NDVI) from annual Landsat composites, focused on agricultural areas.4. Water Yield: Estimated using a water balance approach (Precipitation - Evapotranspiration) with CHIRPS and ERA5-Land data (water yield depth in mm). These datasets are based on a consistent, newly generated set of land cover maps derived from Google Dynamic World data, which are also included. This research was conducted using Google Earth Engine and desktop GIS software. The data is intended to support research in watershed management, hydrology, land use planning, and system dynamics modeling. For complete methodology and interpretation, please refer to the associated publications.