艾利颱風於2004年8月24至25日通過北台灣海面時,帶來豐沛雨量,造成石門水庫集水區發生數量相當可觀之崩塌及土石流災害,大量土砂在水流運搬之下,除增加水庫集水區大量鬆動土石外,由於細顆粒泥砂導致之原水濁度過高,使得桃園地區停水長達十餘天。2005年8月4目至5目馬莎通過台灣東北方海面時,大量豪雨再度造成石門水庫原水混濁,桃園地區停水之災情。本研究為瞭解崩塌之原因,蒐集1986、1998、2003、2004及2005年之崩塌地,與不同之區位進行比較,以探討其中之關係。本研究並發現,由於不同之崩塌地判釋方法及精度,將產生不等之誤差,而造成判釋結果有低估之現象,以桃113線之研究結果為例,成功判釋出之崩塌面積約為實際崩塌面積之百分之二十二。
When the typhoon Acre passed through the sea north of Taiwan on August 24 and 25 in 2004, the extremely rainfall cause the huge landslides and debris-flow disasters in the Shihmen reservoir watershed. A large number of sand cause the reservoir to deposited and because the water muddied, there is no water to use for ten more days in Taoyuan. When the typhoon Matsa passed through the sea northeast of Taiwan on August 4 and 5 in 2005, the rainfall caused the water muddied and there is no water to use again. In order to understand the reason of the landslide, we compare with the landslide data included the years of 1986, 1998, 2003, 2004, 2005 and different location-specific in relationship. We found that different methods and accuracy on landslides interpretation caused different error, and the result will be underrated. For example in road ”Tao 113”, the area of landslides interpretation is about thirty-patent of the area of real landslides.