Is greenery in Sri Lanka fading away? Statistical analysis

As a tropical island, Sri Lanka’s rich vegetation growth is strongly influenced by climatic factors, teleconnection indices, and the Indian Ocean Dipole (IOD). The impact of climate drivers, temperature, precipitation, Niño1+2, Niño3, Niño3.4, Niño4, and IOD, on vegetation can be investigated with the help of the Normalized Vegetation Index (NDVI). This study focuses on the variability of NDVI and the dependence of variability on climate drivers over a time period in Sri Lanka. The preliminary results on trend analysis of the NDVI dataset along with the meteorological data for the period 1984-2021, with respect to province-wise vegetation, showed; there is a downward trend in the average NDVI in the Eastern province, an insignificant long-term change in the North Central province, while the other provinces showed upward trends. The impact of climate drivers on vegetation difference, the LASSO regression model, and the Vector Error Correlation Model (VECM) were fitted, while Autoregressive Integrated Moving Average (ARIMA) models were fitted to investigate the behavior of average NDVI over time. Based on the Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and Mean Absolute Percent Error (MAPE), model performance varied by region. LASSO regression provided the most accurate forecasts in the Eastern and North Central provinces, while VECM performed better in the Southern and Western provinces. Temperature was identified as the dominant climatic factor influencing NDVI across the entire country, whereas precipitation, Niño1+2, Niño3, and IOD affected NDVI only in specific regions. However, Niño3.4 and Niño4 showed significant impact on vegetation dynamics when forecasted using VECM. The findings of this study are important for policymaking, land-use planning in agriculture and construction, climatic factor analysis, ecological assessments, and conservation strategies.