GIS-Based Spatial ID3 Analysis for Cocoa Land Suitability

Abstract

The decline in suitable cocoa land in Indonesia has led to reduced productivity and increased reliance on imports. This study addresses the gap for cocoa land suitability analysis by utilizing the Spatial ID3 algorithm, a method not previously applied in this region, to assess cocoa cultivation potential. A key contribution of this paper is the development of an interactive, web-based geographic information system tool that integrates spatial ID3 classification with mapping capabilities, providing decision-makers with a valuable tool for sustainable cocoa production and land management. The model used eight predictor variables, including soil, topography, and environmental factors, to generate 53 rules, which were then converted into interactive graphics. The tool consists of three main elements: an interactive Leaflet map for cocoa land suitability, detailed suitability classifications at the village level, and recommendations for five cocoa genotypes best suited to local conditions. Results indicate that the most significant area (231,018 ha) is classified as S3 (marginally suitable), followed by S2 (moderately suitable) with 16,443 ha. The unsuitable (N) land covers 49,253 ha, and the unclassified land accounts for 1,363 ha. No S1 (highly suitable) land was found, suggesting that ideal conditions may be absent in the region. This system equips decision-makers with evidence-based tools to prioritize sustainable cocoa production strategies, optimize land management in S2 zones, and implement adaptive interventions in S3 areas. Ultimately, this methodology supports sustainable agricultural development and aids Indonesia in its goal of cocoa self-sufficiency.