921 地震之後，坡地災害發生之次數明顯的增加，使得坡地住民的受災風險因而提升。
這十年來政府雖積極進行許多整治工作，但由於氣候變異影響逐漸顯著，使得短延時且強降
雨的致災事件增多，例如：2009 年的莫拉克颱風夾帶超過兩百年頻率的超大豪雨，釀成全台
共 1690 處的坡地災害、130 處道路中斷、196 座橋梁損毀以及 699 人死亡等嚴重災情。過去
依據物理脆弱度因子 (有效集水面積、岩體之破碎程度、通過之斷層長度、崩塌面積等) 來進
行的預警工作，雖可針對山坡地的自然環境進行監控，但若要改善坡地災害造成的損失與傷
亡，亦不可忽略對山坡地社會脆弱因素的探討與分析，才能理解社會系統面對自然災害時的
抵抗能力，藉以有效降低災害衝擊。本文將坡地災害之社會脆弱度因素分類為四個取向，包
括：(1) 可能的最大損失 (保全人數、結構物損失、家俱家電、交通工具的損失)，(2) 環境建
設(土地使用、道路交通)，(3) 自保能力 (依賴人口、警消人力、避難所與受災次數等)，(4) 復
原與適應能力。同時為了使研究成果得以落實於現今防災工作，本文使用帕累托等級分析法
(Pareto ranking)，建立分項及綜合之評估指標。最後，透過 GIS 圖層的繪製，空間化不同地區
之社會脆弱度，使之易於結合坡地災害之區域分布特性。指標評估結果發現，屏東縣、台東
縣內高脆弱鄉鎮最多，依比例來說，嘉義市與桃園縣內的高脆弱鄉鎮市區比例最高，這四個
縣市是坡地災害衝擊下，高社會脆弱度區，此外，若套疊坡地歷史災點更可發現，南投縣信
義、水里、埔里鄉，新竹縣尖石鄉、苗栗縣泰安鄉等，是坡地災害的高風險地區。藉由指標
的評估結果有助於災害防救單位進行適切的災前減災規畫、災時應變評估與災後復原策略擬
訂等防災工作。

This study constructs a framework of social vulnerability index of slope-land disasters, and
assesses the idea of social vulnerability by analyzing a group of factors classifed by the framework
of SVIoL. The number of landslide disasters increased significantly after the Chi-chi earthquake,
putting residents of mountain areas at higher risk of debris fow or landslide. The government has
revised numerous engineering design over the past decade to reduce the risk of slope-land disasters.
However, weather events still cause enormous damage. For example, Typhoon Morakot caused the
heavy rainfall breaking the historical records, and resulted in 1,690 landslide/debris fow events, 130
broken roads, 196 damaged bridges and 699 deaths. This disaster revealed that disaster management
needs go beyond merely evaluating physical vulnerability or building engineering facilities, and
that social vulnerability index assessment is a potential means of improving disaster management
and catastrophe resistance. The framework of SVI has four important aspects attributed by previous
studies: 1) maximum loss of household property, 2) environmental engineering, 3) resistance to slope-
land disasters, and 4) self-recovery ability. Pareto ranking (PR) analysis was applied to integrate the
four aspects, and to rank the intensity of SV scores of each town, where higher ranks indicated greater
vulnerability. Pingtung and Taitung counties have the highest number of towns assigned to the highest
rank, while Chiayi City and Taoyuan County have the highest percentage of towns with the highest
PR. Furthermore, after overlapping the historical landslide hotspots with the SVI layer, the most risky
places are Shueili, Shini and Puli in Nantou County, Jianshi in Hsinchu County, and Taian in Miaoli
County. SVIoL can help not just local governments but also central government understanding the
disaster vulnerability of different places, and adjust their disaster prevention, response and recovery
strategies accordingly.