River Flood Prediction using Time Series Model

With the passage of time the impacts of natural hazards continue to increase around the world. The globalization and growth of human societies and their escalating complexity and river flooding will further increase the risks of natural hazards. Flood prediction and control are one of the greatest challenges facing the world today, which have become more frequent and severe due to the effects of global climate change and human alterations of the natural environment. Therefore, it is important to protect people and their property from flooding, helping organizations. By giving warnings of possible floods so that people can make arrangements or move out of the area if is dangerous. Time series model that can predict river flooding using water level data over Ayeyarwady river in Myanmar is proposed. This model focus on the prediction of events and can capture the fact that time flows forward and is applied to model the spatial dependencies not only between the hydrological and hydro-meteorological variables but also between weather stations in Myanmar for river flood prediction. The system predicts water level by daily to daily forecast type for additive models in Time Series using Two-Period, Three-Period and Four-Period moving average. The system compares with the predicted results and actual weather station results. Finally, the best model will be shown for river flood prediction over Ayeyarwady River in Myanmar.