Mapping of Geomorphological Zones and Benthic Habitats Using PlanetScope Imagery with Google Earth Engine on Leti Island, Maluku Province

Abstract

The shallow marine ecosystem of Leti Island, Southwest Maluku, Indonesia, which is rich in biodiversity but lacks data, was mapped using 3-meter spatial resolution PlanetScope imagery processed in Google Earth Engine with the Random Forest algorithm for hierarchical classification of geomorphological zones (10 classes) and benthic habitats (6 classes), integrating radiometric correction, ArcGIS Pro mosaicking, and training samples from field data. The results show the dominance of coral/algae (621.25 ha, 38.27%) and seagrass (536.77 ha, 33.07%) habitats, followed by rock (389 ha, 23.96%), reflecting erosion-sedimentation dynamics and the potential for biodiversity and fisheries conservation, while the geomorphological zones are led by the terrestrial reef flat (705.40 ha, 40.03%) and outer reef flat (228.71 ha, 12.98%), with reef slopes and lagoons serving as adaptive transitions to local hydrodynamics and non-volcanic tectonics. The results of this study emphasize the ecological implications for resilience to climate change and human activities, recommending field validation, predictive spatial modeling via lidar/sonar, and the use of these thematic maps as a basis for the sustainable management of Indonesia's outermost marine conservation areas.