New debris flow critical rainfall line setting via cluster analysis and support vector machine after the Chi-Chi huge earthquake

The Chi-Chi huge earthquake occurred in Taiwan in 1999 and it caused landslides and debris flows, which brought about considerable loss of life and property. To help prevent damage by debris flows, it is necessary to establish a new critical rainfall line of debris flow. Following the Chi-Chi huge earthquake, the rainfall threshold of debris flow streams in Taiwan was ostensibly reduced. To comprehend these changes, this study used a four-year (1999–2003) dataset of 79 debris flow events, and it adopted the family competition genetic algorithm as a clustering method to merge rainfall data of streams based on similar characteristics. In addition, 8 predisposing features for debris flows were used to cluster 377 streams with similar predisposing features into 7 groups, and this study used a family competition genetic algorithm. Then, this study apply the support vector machine (SVM) for establishing new critical rainfall lines. Experimental results confirmed that the FCGA and the SVM method performed well in setting a new critical rainfall line.