Abstract
Global forest ecosystems are experiencing significant degradation due to the combined effects of climate change and human activities, leading to a deterioration of forest health and increased ecological instability. In light of these global environmental challenges, it has become crucial to protect forest ecosystems, not only to maintain biodiversity but also to ensure the delivery of essential ecosystem services. This study introduces a comprehensive framework for evaluating forest health in the Sundarbans Biosphere Reserve (SBR), one of the largest and most ecologically vital mangrove ecosystems globally, through the integration of multi-temporal Landsat imagery, machine learning (ML) techniques, and multicriteria decision analysis (MCDA). Employing advanced ML algorithms, such as linear regression (LR), support vector machines (SVM), artificial neural networks (ANNs), and ensemble approaches, the research examines key structural and morphological vegetation characteristics. Additionally, a novel 2100 ARIMA predictive model is utilized to forecast long-term trends in forest health and fragmentation under various climate scenarios, offering a quantitative analysis of potential forest resilience outcomes. The results reveal that species diversity is the strongest indicator of forest health, while fragmentation emerges as the critical factor affecting both primary and secondary forest ecosystems. The integrated ML and MCDA framework achieved a commendable predictive accuracy of 89%, successfully classifying forest health across several metrics. Furthermore, ARIMA model projections indicate significant forest degradation by 2100, emphasizing the urgent need for adaptive management strategies. This study highlights the essential role of nature-based climate solutions (NbS) in bolstering ecosystem resilience. By correlating forest health with sustainability objectives, this research lays a scientific foundation for informed conservation planning and aligns with the United Nations Sustainable Development Goals (SDG 13: Climate Action, SDG 15: Life on Land), promoting a transition towards data-driven, sustainable forest management practices.
