<![CDATA[Floods and droughts in Sumbawa Regency have intensified in frequency and impact over recent decades, driven by complex climatic interactions and anthropogenic activities. Accurate rainfall forecasting is critical for effective disaster risk reduction and water resource planning. This study develops a Long Short-Term Memory (LSTM) model that integrates satellite-derived rainfall with global climate indicators (Sea Surface Temperature (SST) and Southern Oscillation Index (SOI)) to enhance monthly rainfall prediction. Compared to statistical baselines (SARIMAX: RMSE ≈ 92 mm) and machine learning baselines (Random Forest: RMSE ≈ 84 mm), the multivariate LSTM achieves superior performance with RMSE = 65.2 mm (R = 0.82), reducing forecast error by ~25%. The 12-month forecast for June 2025–May 2026 indicates an extended dry season (June–September) followed by intense rainfall peaking at 266 mm in February 2026, highlighting risks of hydrometeorological extremes. By pioneering the fusion of satellite data and LSTM in Indonesia, this research provides actionable insights for early warning systems and supports climate adaptation strategies in water management, agriculture, and disaster preparedness. The model offers a scalable framework for operational rainfall prediction in climate-vulnerable tropical regions.]]>
Copyrights © 2025
