Mapping flood susceptibility and settlement exposure using random forest for spatial planning in Barru Regency

Abstract

This study aims to analyze flood susceptibility and evaluate the spatial exposure of settlement areas in Barru regency, Indonesia, using a geographic information system (GIS) and a random forest approach.Environmental variables, including elevation, slope, soil type, rainfall, land cover, and river buffer, were integrated into the model.Flood susceptibility was expressed as probability values ranging from 0 to 1 and subsequently classified into low, moderate, and high categories for interpretation.The results show that low susceptibility dominates the study area (67.51%), followed by moderate (15.47%) and high susceptibility (17.02%).Model evaluation indicates stable performance, with an accuracy of approximately 0.86 and consistent results across validation schemes.Variable importance analysis identifies elevation as the most influential factor, followed by slope and soil type, highlighting the dominant role of topographic characteristics.Spatial overlay analysis reveals that approximately 72.63% of settlement areas are located within high susceptibility zones, indicating substantial exposure.These findings emphasize the importance of integrating flood susceptibility assessment into spatial planning to support susceptibility-informed and more sustainable settlement development in hazard-prone areas.

Access to Document

Other files and links

• Link to publication in Scopus
• Open Access Version Available

Fingerprint