Project Overview:I led a comprehensive decadal (2013–2024) research project assessing the vegetation dynamics of the Sundarbans, the world’s largest mangrove forest. Using satellite-derived data and advanced statistical modeling, I evaluated ecosystem stability, phenological shifts, and climatic influences to provide evidence-based insights for conservation.

What Went Into the Project:

• Data Preparation: Acquired and processed monthly NDVI (MODIS), rainfall (CHIRPS), and temperature (TerraClimate) data in Google Earth Engine. Performed masking, gap-filling, resampling, and temporal aggregation to create analysis-ready time series for the Bangladeshi Sundarbans.

• Analysis: Conducted trend detection using Mann–Kendall and Sen’s slope tests; decomposed seasonal signals via STL; performed lagged correlation and Granger causality tests to explore climate–vegetation relationships; developed a SARIMA forecasting model.

• Journal:International Journal of Big Data Mining for Global Warming (under review).

Key Results and Their Interpretation:

1. Raw NDVI showed an increasing trend, but after deseasonalization, no long-term trend remained.Interpretation: The apparent greening was seasonal, not sustained. The ecosystem has been fundamentally stable over the past decade, contradicting degradation narratives.

2. Significant NDVI increases occurred only in January, April, May, and December.Interpretation: This indicates a phenological shift—extending the productive season into cooler months, likely due to milder temperatures.

3. Rainfall had an immediate negative correlation with NDVI but positive effects at 3- and 5-month lags.Interpretation: Immediate suppression is likely from cloud cover; delayed benefits suggest soil moisture recharge.

4. Temperature showed no significant short-term influence on NDVI.Interpretation: Short-term temperature fluctuations are not a primary driver of vegetation dynamics in this system.

5. No Granger-causal relationship was found from climate variables to NDVI.Interpretation: The ecosystem’s vegetation health is internally regulated and resilient; past climate alone does not predict future greenness.

6. SARIMA model achieved a MAPE of 3.99% and forecasts stable NDVI through 2027.Interpretation: The model provides a highly accurate quantitative baseline for monitoring, with no abrupt changes projected.