Land use and land cover change detection and prediction in the cross-Sanaga-Bioko coastal forest for sustainable forest management

Land use and land cover (LULC) change in forested regions is a key indicator of environmental transformation, often leading to deforestation, degradation, and biodiversity loss, with consequences for both ecological stability and human well-being. Sustainable forest management (SFM) plays a critical role in mitigating these impacts by balancing development with conservation. This study assesses forest dynamics in the Cross-Sanaga-Bioko (CSB) region of West Africa between 2000 and 2021 and projects future changes to support SFM efforts. Using Google Earth Engine, supervised classification was performed via the CART algorithm, with post-classification change detection used to quantify gains and losses. The Multi-Layer Perceptron and Cellular Automata-Markov chain models forecast land cover transitions to 2063. The classification achieved an overall accuracy of 90% and Kappa statistics averaging 89%. The 2063 prediction model yielded a validated K-standard of 88.43%, indicating high reliability. Findings reveal a 21.03% projected decline in forested areas, with corresponding increases in agricultural land (19.69%) and built-up areas (10.88%). These trends emphasize the urgency of implementing sustainable land-use practices to harmonize forest conservation with agricultural expansion and urban growth.