Accurate and efficient prediction of typhoon-induced surge is an important task for coastal disaster prevention. The purpose of this study is to develop a long-lead-time prediction model for storm surge using effective controlling factors and artificial neural network. Effective controlling parameters will be carefully investigated to establish the artificial neural network model, including the maximum wind speed, the center air pressure, the radius of the storm, the relative location between the station and the center typhoon (i.e., distance and angle). The combined effective information can not only reduce the input dimension but also improve the model's learning and forecasting capability.
Coastal regions of a country are highly developed owing to rich natural resources and economic potentials. To date, more than one billion people are living and working in the coastal areas over the world. The continuously increasing population might approach 2 to 5 billion by 2080 (IPCC, 2007). However, coastal environment system, affected by land/river, ocean and atmosphere, would face the threats by various disasters. One of them is the typhoon-induced surges.
Storm surge has always been regarded as a very important topic in science or practice due to its severe impacts. For example, storm surge struck the North Sea in Europe and caused a serious flooding in the Netherlands, Belgium and the United Kingdom at midnight on the of Feb 1, 1953. A total of 2,167 people died in the disaster. The worst storm surge over 12 meters was recorded in Bangladesh in 1970. Within 20 minutes, the number of deaths was as high as 300,000. At the beginning of this century, Hurricane Katrina swept through the Gulf of Mexico in the south of the United States in 2005. Thousands were killed and New Orleans almost destroyed. In 2012, Typhoon Sandy hit New York in the east coast of the United States. In Asia, Cyclone Nargis struck Myanmar in 2008, causing 130,000 deaths. In 2013, typhoon Haiyan hit the Philippines, leading to 6,000 deaths and $ 14 billion in economic losses. In addition, under the influence of the climate change and global warming, the frequency and magnitude of super typhoons would surpass historical records (Webster et al., 2005; Emanuel, 2005), resulting in more destructive storm surge. A deeper understanding of storm surge (i.e., characteristics and related physical processes), and accurate/efficient predictions are still the subject of concern.