Document Type


Publication Date



Life and property losses because of disasters such as flood and landslide are getting dramatic increases for past years. Frequent extreme weather events have even worsened the damages globally. Given sufficient information in advance, disaster preparedness and management can be well-settled. The damages and losses can be mitigated and even prevented. An early disaster warning system with enough leadtimes is essential for decision-makers to take prevented measures and execute cautions. This study proposed a prototype of an early disaster warning system which issues warnings for rainfall related disasters such as flood and landslide. Using ensemble quantitative precipitation forecasts, the system could estimate the disaster risk and issue possible threat warnings up to 72 hours in advance. The system is a statistics-based, instead of physics-based model to meet the real time operational needs. The advantage is to provide the information and issue the warnings rapidly which helps the decision-makers have more time to deploy the rescue forces such as manpower and water pumps. For the flood forecasts, the proposed system integrated artificial intelligence with historical records, design capacity of storm sewer system, and threshold rainfall to identify the high risk areas in a township level. A simple physics-based model then was imposed to identify possible flooding areas downscale to the village level. The landslide susceptibility model was developed using the logistic regression that used rainfall, topography and geology as the geo-environment variables affecting slope stability. According to occurrence of disaster and available observed records, the proposed system used typhoon events from 2010 to 2012 in Pingtung County, Taiwan to evaluate the performance. The spatial and temporal comparisons with observations demonstrated promising potential as a valuable reference for better emergency response to alleviate the loss of lives and property.


Session S2-02, Special Session: Hydroinformatics Tools for Flood Resiliency in Urban Areas II



To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.