I am honored to be the first author of the manuscript, "Stability Amidst Change: Ten-Year Insights into Sundarbans Mangrove Dynamics," co-authored with Md. Enamul Hoque (Neel), Associate Professor at the University of Chittagong. The article is currently under review at the International Journal of Big Data Mining for Global Warming.

The work applies trend detection, spatial pattern decomposition, and forecasting to produce decision-ready evidence for conservation and climate adaptation planning.

• Designed a multi-stage analysis pipeline in Earth Engine and Python combining NDVI trend analysis, Empirical Orthogonal Functions (EOF), Granger causality, and SARIMA forecasting to assess mangrove phenology and resilience under climatic pressure.

• Demonstrated that long-term NDVI is stable rather than monotonically improving, with climate-linked shifts in seasonal productivity (cooler months), directly relevant to climate-risk and adaptation assessments in coastal forests.

• Quantified spatial modes of variability (basin-wide synchrony, north–south salinity gradient, local disturbance mode), providing a spatially explicit basis for targeted management and monitoring in a high‑value blue carbon ecosystem.

Below is the abstract of the manuscript:

Abstract

The Sundarbans, the world's largest contiguous mangrove forest, represents a critical yet vulnerable ecosystem facing escalating climatic and anthropogenic pressures. This study provides a comprehensive decadal assessment (2013–2024) of the Bangladeshi Sundarbans by integrating remote sensing with a multi-stage statistical framework encompassing trend analysis, spatial pattern decomposition, causal inference, and forecasting. Contrary to narratives of widespread degradation, the analysis of vegetation health (NDVI) reveals a fundamentally stable ecosystem, where an apparent long-term greening trend is attributable solely to a strong seasonal cycle rather than sustained improvement. However, significant phenological restructuring was detected, characterized by asymmetric greening concentrated in the cooler transitional months (January, April, May, December), suggesting an extension of the productive season. Climatic driver analysis uncovered a complex relationship with rainfall, featuring a significant immediate suppressive effect likely from cloud cover, followed by delayed beneficial impacts at 3- and 5-month lags. Temperature exhibited no significant short-term influence on vegetation anomalies. Crucially, neither rainfall nor temperature demonstrated a predictive, Granger-causal relationship with future NDVI, underscoring the system's stability. Empirical Orthogonal Function (EOF) analysis identified the dominant spatial structure of variability, revealing a hierarchical system comprising: (1) a dominant regional synchrony mode (33–39% of variance), reflecting uniform responses to large-scale climate forcing; (2) a north–south gradient mode, highlighting differential vulnerability along salinity and freshwater continua; and (3) a patch-scale heterogeneity mode, capturing localized disturbance and recovery. A highly accurate Seasonal ARIMA (SARIMA) model (MAPE: 3.99%) was developed and forecasts stable vegetation health through 2027, with all projected values remaining within the bounds of historical variability. The study concludes that the Sundarbans exhibits resilience through adaptive phenology and spatially organized responses rather than systematic degradation. This integrated spatiotemporal assessment provides a robust quantitative baseline and forecasting tool essential for effective, spatially targeted conservation, adaptive management, and climate adaptation planning.