Land Suitability Evaluation for Sustainable Tea Cultivation: A Machine Learning and AHP Integrated Approach

Abstract Land suitability evaluation is essential for achieving efficient and sustainable use of land resources. This study applied a multi-criteria decision analysis by integrating soil spatial information generated by using Random Forest (RF) model with climate, topography, land cover, and accessibility factors in the Analytical Hierarchy Process (AHP) framework to evaluate land suitability for tea cultivation in Ganyange Ward, Tarime District, Tanzania. Key soil properties relevant to tea growth were predicted from environmental covariates using RF model and integrated with climate, topography, land cover, and accessibility data in the suitability assessment process. AHP based on expert judgment was used to determine the relative priority weights of criteria in the suitability evaluation. The overall tea suitability map was generated using a weighted overlay approach within a geographic information system (GIS) framework. The AHP ranking identified climate as the most influential factor (0.37), followed by soil chemical properties (0.25), soil physical properties (0.18), topography (0.12), land cover characteristics (0.05), and accessibility (0.03). The overall suitability analysis indicated that 4.99% of the study area is highly suitable, 63.16% moderately suitable, 31.80% marginally suitable, and 0.06% not suitable for tea cultivation. To enhance suitability, integrated soil fertility management, and improvement of soil organic matter are recommended.