Due to the global climate changes, the scale and frequency of natural disasters are more difficult topredict and measure. Extreme rainfall often brings astonishing amount of water and causes very seriousdamage in the mountain areas. For instances, during typhoon Morakot in 2009, many roadways andbridges were destroyed by the rainfall in south Taiwan; furthermore, the typhoon Megi brought amazinghourly rainfall to damage the Highway No.9 in I-lan County at 2010. Facing the challenges from theextreme weather conditions, the development of an early warning system has become a critical issue.Therefore, in the proposed study, the authors collected the field data of landslides and debris flows inNo.18 Highway of middle Taiwan. And the authors collected the rainfall records of typhoons and stormsfrom 2008 to 2012. Pattern recognition analysis was conducted to identify typical precipitation patternsthat would cause slope failures and related debris flows. Threshold rainfall intensities and rainfallamounts that would possibly trigger the failures were chosen as two indices and their applicability basedon different theoretical or empirical approaches from the selected precipitation patterns. Correlationbetween triggering rainfall indices and occurrences of landslides is to be established through this analysis.
展开▼
