SPATIAL ANALYSIS AND POPULATION DENSITY OF GREEN TURTLE (Chelonia mydas) IN INDONESIAN WATERS A SPECIES DISTRIBUTION MODELING APPROACH

Abstract

Indonesia is one of the world's centers of marine biodiversity, serving as an important habitat for the Green turtle (Chelonia mydas). Various anthropogenic pressures and environmental changes have led to a decline in habitat quality and turtle populations, necessitating large-scale spatial analysis to support conservation efforts. This study aims to analyze the spatial distribution patterns and population density of Green turtles in Indonesian waters using occurrence data obtained from the Ocean Biodiversity Information System (OBIS). The study applies a combination of Kernel Density Estimation (KDE) and Convolutional Neural Network (CNN) methods to model species distribution and identify potential habitats. Preprocessing stages include removing duplicate data, handling empty values, validating geographic coordinates, and spatial filtering within Indonesian waters. KDE is used to generate continuous density maps and identify population hotspots, while a 1° × 1° spatial grid is used as input for the CNN model. The CNN model is trained using the KDE-derived spatial attributes to classify potential and less potential habitats. The results show that the distribution of Green turtles is uneven and concentrated in several key hotspots, particularly in the Nusa Tenggara, Maluku, Banda Sea, and Papua regions. The CNN model is capable of recognizing spatial habitat patterns well and generating potential habitat predictions consistent with hotspots identified through the KDE. The integration of KDE and CNN has proven effective in mapping species distribution and identifying important Green turtle habitats at a national scale. The results of this study can be used as a scientific basis for planning marine conservation areas, protecting critical habitats, and managing sustainable marine resources in Indonesia.