Normalized Difference Vegetation Index based Drought Prediction Model using Enhanced-Long Short-Term Memory Algorithm

One serious natural calamity that affects ecosystems, water supplies, and agriculture worldwide is drought. To lessen the consequences of drought, early detection of its circumstances is essential. By combining Normalized Difference Vegetation Index (NDVI) with Enhanced-Long Short-Term Memory (E-LSTM) algorithm for drought prediction model, this study offers an innovative approach for drought prediction. NDVI, a widely used satellite-derived vegetation index, is used to monitor vegetation health, which correlates with drought severity. The drought prediction model in this study uses E-LSTM's capacity to capture longrange dependencies to forecast drought conditions by examining temporal patterns of NDVI data over time with other related climate variables like temperature, rainfall and surface soil moisture, then results indicate its effectiveness in providing accurate early drought warnings. This research aims to fill the gap of issuing early warning for a slow onset type hazard like drought. The Root Mean Squared Error (RMSE) of 0.0571 is used to evaluate the model's performance.