Integrated approach for scenario-based flood simulation

This paper present recent achievements obtained within the context of the ANFAS project (Data fusion for flood simulation and decision support). The overal objective of ANFAS was to set up an integrated system for flood prevention and protection. Data fusion and models fusion techniques constitute the core of the scientific research work of the project. Data fusion refers to the use of multi-source data, such as remote sensing imagery, hydraulic data, meteorologic data, geographic data, etc. Models fusion enables to describe the mechanism and to simulate complex dynamic phenomena , from heavy rainfall to flood propagation in the plain. Models and data are integrated within the ANFAS system, thus enabling users to perform simulation with an easy-to-use interface. 1Introduction Economical loss resulting from floods count for about 70% of the total cost caused by natural disasters in China. The understanding of the mechanism and process leading to floods is of crucial importance to take appropriate long term preventive measures and to react efficiently during the crisis. Thought various models and tools have been developed around the world already, key scientific issues are still open and challenging. 2ANFAS project overview Context Although the occurrence of natural disasters cannot be avoided, the damage they cause can be reduced if early measures are taken by anticipation. ANFAS aimed at developing scientific tools that can provide decision makers with relevant and rigorous information, to help them to understand what early actions need to be taken, according to some fixed factors. The ANFAS project has been set up with a multi-disciplinary consortium a total of thirteen partners, from five countries (China –Chinese Academy of Sciences Institute of Automation, Institute of Atmospheric Physics, Institute of Remote Sensing Applications, Wuhan University, Yangtze River Scientific Research Institute---, Greece, France, Slovakia and United kingdom), including scientists, engineers and end-users. It was supported by the Chinese Ministry of Science and Technology 863 program, the European Commission 5 framework Information Society Technology program and the World Bank InfoDev program. Objective and pilot sites The objective of the project is to set a Decision Support System for Flood prevention and protection. The ANFAS system could be used: (i) to prepare possible early reaction plans in case of flood probable occurrence ; (ii) to develop the river banks. ANFAS System can perform flood simulation based on scenario defined by the user. It integrates heavy rainfall model, hydraulic and hydrologic modeling, scientific computing, computer vision, internet technology, geographic information systems, and soil sciences expertise. The output of the simulation provides information to the user for decision support. The system is under a distributed architecture, works on Internet based on Web interface. The application sites are on the Vah River (Slovakia), Loire River (France) and JingJiang reach on the Yangtze River, 80km downstream from the Three Gorges (China, Hubei Province). The three sites differ by their size --the site in JingJiang is several times larger than the site in Loire--; they all have dynamic economic activity and have recurrent heavy floods. Technical tasks The project is organised into five main technical tasks : • Preparation of a Geographical Information System database ; this includes hydraulic data, geographical maps, data required for the simulation models, socio-economic data ; • Scene modeling from remote sensing images ; image analysis and understanding techniques were developed in order to extract relevant information from multi-source images recently acquired; we thus can recreate a up-to-date description of the imaged scene; • Flood modeling and simulation ; this task relies on the selection or development of heavy rain models , hydraulic models and their adaptation to the pilot sites ; • Impact assessment and application ; it is the basis for decision support; the key point is to transpose the output of the simulation models into high level information; • System integration. Each partner is involved in one or several of these tasks, depending on each partner knowledge and experience.