Exploring vegetation dynamics, driving factors, and future vegetation coverage in the Chittagong Hill Tracts of Bangladesh

ABSTRACT This study investigated the historical and possible future alterations of vegetative cover in the Chittagong Hill Tracts (CHT) using geospatial methods. The study utilized five cloud-free Landsat images for the years 1989, 1998, 2008, 2018, and 2024 to analyze vegetation dynamics using NDVI differencing and a combinatorial matrix approach, integrated with spatial regression modelling. The ANN-CA model was applied to forecast the future vegetation transitions up to 2030. Additionally, spatial regression analysis (Ordinary Least Squares, Global Moran’s I of OLS residuals, and Geographically Weighted Regression) was employed to quantify spatial dependence and identify key environmental and anthropogenic drivers of vegetation change. Results indicate a substantial increase in sparse vegetation from 15.83% in 1989 to 28.17% in 2024, driven largely by anthropogenic pressure. Moderate vegetation remained the dominant land cover class throughout the study period, showing signs of degradation from 77.75% in 1989 to 65.29% in 2024. Non-Vegetation and Dense vegetation showed a minimal portion of the CHT. From the transition matrix, it was identified that a notable portion of moderate vegetation is converted to sparse vegetation. This trend is predicted to continue with non-vegetation accounting for 4.87%, sparse vegetation at 29.71%, moderate vegetation at 64.39%, and dense at 1.03% of the total area of CHT by 2030. The analysis further reveals that slope, elevation, NDMI, population, and settlement expansion significantly influence vegetation dynamics. The findings highlight accelerating vegetation degradation in the CHT and emphasize the need for spatially informed and evidence-based land management strategies for biodiversity conservation and climate resilience.