Predictive Modeling of Prehistoric Site Distribution in the Southeastern Sulawesi Karst Landscape Using MaxEnt

Abstract

This study applies a predictive modeling approach using the MaxEnt (Maximum Entropy) algorithm to estimate the distribution of prehistoric archaeological sites in the karst landscape of the southeastern arm of Sulawesi Island. This region, which includes the karst zones of Matarombeo, Mekongga, Tangkelemboke, and Sombori, holds high archaeological potential but remains largely unexplored. By integrating presence data from 64 previously recorded archaeological sites with six environmental variables—elevation, slope, aspect, hydrology, and geology the MaxEnt model produced an indicative map of potential archaeological site locations. The model’s output was validated through a field survey in North Kolaka Regency in December 2024, which successfully identified several new sites across various probability classes. The resulting data and maps offer significant potential for cultural heritage preservation, spatial planning, and further research in similar karst environments.