Research On Big-data-based Urban Flood Defense Decision SupportSystem

As Urbanization’s progress promoting social civilization, it also endangered big challenges; urban flood is one of them. The rapid progress of urbanization has a significant impact on urban original hydrological environment, thus led the urban flood control into a new situation. With the development of Internet of Things and sensing technologies, there has a variety of big data produced in urban spaces. Using big data to predict urban flood effectively will help to make effective flood mitigation measures.Based on the depth research of the historical hydrological data in Hangzhou, we proposed a hybrid water level prediction model under the consideration of temporal correlation and spatial correlation. The model can provide effective forecast for 1-6 hours of water level, thus providing information reference for urban flood control when flood season is coming.Flood defense decision support system is an important non-structural measures for flood control and disaster reduction. After in depth study of big data and cloud computing technology, we established a big-data-based urban flood defense decision support system which utilize big data and cloud computing technology. As the system uses distributed database HBase, it can store massive amount of historical hydrological data and handle the rapid growth of data. Combined with the hybrid water level prediction model, the system could provide useful information for urban flood defense work. Using the big data analytics engine-Impala-to process the massive historical hydrological data in real-time, and applying the data visualization technology to find the intrinsic value of historical data, it can provide information to guide the urban flood defense.Big-data-based urban flood defense decision support system consists of real-time water level query module, water level prediction and early warning module, and historical hydrological data analytics module. Real-time water level prediction module provides water level information of the whole city in an intuitive way. Water level prediction and early warning module forecast water level of the next 1-6 hours, and giving warning message on station which may exceed limit. Historical hydrological data analytics module utilizes big data analytics tools to find the hidden value in massive historical hydrological data, and provides effective information to urban flood defense.