本研究蒐集集集地震強地動測站紀錄與人員罹難屬性資料庫，利用「等震度」與「等人口母體數」的概念，透過GIS軟體的運算與分析，以建立不同強地動參數與人員罹難之易損曲線，並以集集地震為模擬地震，進行人員罹難之推估及驗證。 在建立的過程中，發現有三個項目可以使本研究概念更加的精進，分別為(1)強地動測站的數量、(2)區間內PGA的決定、(3)未定位罹難人數的補償，同時以等母體數為20萬人、10萬人、5萬人與2.5萬人四種尺度分別探討，研究結果顯示這三項的改進對於人員罹難易損曲線之結果都有一定程度的幫助，尤以對未定位罹難者進行補償最明顯。本研究也對於傳統的分析方法與等震度等母體數之概念進行比較，研究結果顯示等震度等母體數之概念所得之人員罹難易損曲線結果較為良好，趨勢也較為明顯。 本研究亦對易損曲線探討其信賴區間，並針對不同的建立模式對集集地震主要災區(中投地區)進行推估與驗證，由結果發現在許多地區的預測都相當良好，但是在建築物損壞率較高之地區都有明顯低估的情形，此現象說明建築物的耐震能力很直接的影響人員罹難機率的發生，即使是位於受震震度較小之地區。 In this research we collect the murdered people damage data and the strong motion data of the Chi-Chi earthquake, and propose concept of “coseismic” and “equal-population number”. We can construct fragility curves which were relation between strong ground motion index and mortality by means of GIS software. Further more we take the Chi-Chi earthquake as a study case to assess the mortality distribution. During the process of setting up, there are three project that can make this research concept progressing greatly further, is (1) numbers of the seismic station, (2) decision of PGA in each donut, (3) compensate for fatality population of non-orientation. The scales such as (1)200,000, (2)100,000, (3)50,000 and (4)25,000 death tolls are used as equal-population number to decide the analysis regions. The result of study shows that the improvement of these three items has certain help in fragility curve of mortality, especially in compensate for fatality population of non-orientation . In this study, we also compared the traditional method with concept of “coseismic” and “equal-population number”. The result of study shows that the fragility curve is better by using the concept of “coseismic” and “equal-population number”, and the trend is very obvious too. In this research we also find the confidence interval for fragility curve. Using different setting-up ways to estimate and prove in the main disaster area of the Chi-Chi earthquake. The research indicated that the estimation in most areas is quite good. In the area that the damage ratio of buildings is higher, it will underestimate obviously. This phenomenon state that anti-shaking ability of buildings directly influence mortality, even if the area in lower intensity.
