2001年7月台湾で発生した土砂災害

Abstract

2001年7月30日, 台風0108号(桃芝台風)がもたらした豪雨により, 台湾の花蓮県および南投県を中心とした各地で土砂災害が発生した。これにより, 214名の死者・行方不明者を出すとともに, 家屋や公共土木施設等に多大の被害が生じた。少なくとも121名以上が土砂災害による犠牲者であり, この災害では洪水による被害よりも土砂災害による被害が中心であった。本研究では, 現地調査で得た情報をもとに土砂災害の概要を示すとともに, 花蓮県光復郷大興村清水渓で発生した土石流によって死者・行方不明者が41名にものぼった災害をとりあげ, 発生土石流の規模推定およびその氾濫・堆積過程の再現を試みた。その結果, 推定流出土砂量は約80万m3程度と見積もられた。空中写真から判読された土砂量は約150万m3であり, 計算結果はかなり過小評価となった。この原因については, 計算で考慮した崩壊地が実際にはもう少し多い可能性があること, 空中写真からの見積もりが少し過大である可能性があること等が考えられる。

The Typhoon 0108 (Typhoon Toraji) attacked the eastern and central Taiwan on 30 July, 2001.This typhoon brought about a heavy rainfall. In the central area of Hualien-ken and the southern area of Nantou-ken, the hourly rainfall more than 100mm lasted three hours, and the total rainfall amounted to more than 400 mm, resulting in the occurrence of disasterous debris flow in many places in both Hulien-ken and Nantou-ken.The severe inundation flows also occurred due to both overtopping flows and bank breaches in Chosui River basin in Nantou-ken. By these river floods and sediment disasters, 214 person were killed or missing and huge damages spread out all over the country. This report summarizesthe findings of these disasters obtained through the field survey. Moreover, estimation of debris flow hydrographs of the Tsingsui River in Taiko village, Hualien-ken has been done by using one-dimensional simulation model, in which the grain size distribution of the riverbed materials and the landslides are taken into account. Flooding and deposition processes on the fan of the river are reproduced using two-dimensional model, where calculated results obtained by one-dimensional model such as flow discharge, sediment discharge, and sediment diameters are used as inflow boundary conditions on to the fan area.Calculated sediment volume outfiowed was about 850, 000 m3 and estimated one by using aerial photos was about 1, 500, 000 m3. The reasons of this difference are thought that the landslide areas in the calculation might be too small or the accuracy of analysis of aerial photos might be not so high.