Risk Assessment and Prediction of Floods and Droughts using Attention-LSTM

Floods and Droughts are among the most destructive natural disasters which cause extensive damage to infrastructure, displace populations and are threatening to life and property in a serious way. These disasters need timely and accurate prediction for the purpose of disaster management and mitigation efforts. This paper presents a new framework for the assessment and prediction of floods and droughts. The data integrated for this methodology include satellite imagery of Sentinel-1, Sentinel-2 and MODIS, weather data and historical disaster records. This paper makes a significant contribution to the field, as it presents a holistic solution for the assessment and prediction of floods and droughts. It integrates Attention Long Short Term Memory (LSTM) based model for prediction of disasters. It is the application of advanced technologies to build more resilient communities and enhance disaster preparedness. Focusing on floods and droughts, the system tries to solve two of the biggest problems in disaster management, and assists emergency responders, policymakers and communities make informed decisions and reduce the devastating consequences of these natural disasters.