都市防災生活圈服務潛力之區位問題研究

Location Problem in Regard to the Service Potential of Urban Disaster Prevention Living District

基於地震防災的考量，必須能有效掌握居民避難設施選擇類型與避難設施配置之區位，亦即進一步思考於滿足居民避難設施選擇偏好，同時也能有效檢討避難設施配置規劃之適當性。本研究以921重建地區之南投都市計畫區為例，透過問卷調查方式，應用多項羅吉特模型，以「設施型」設施為比較基準要項，學校型設施之避難路線選擇、規劃設計、硬體設施；廣場型設施之硬體設施，係為影響居民面對災害時，決定避難設施類型選擇之重要變數。藉由區位配置準則的影響（公共設施可及性、交通可及性、可容納避難人數、開放空間比、可服務人口數、淹水潛勢、斷層帶影響範圍），應用Weighted Voronoi圖形劃設防災生活圈其邊界線大部份為曲線，比環域分析以距離為覆蓋半徑，更適合避難設施防災生活圈範圍之劃設。 The earthquake evacuation behavior of residents relates to geographical characteristics. The contingent plan considers proper disaster prevention living area for prompt reaction assumed the least nearest distances to safety shelter. This study considered the willingness of residents' evacuation and selected the reconstruction area after the 921 earthquake in Nan-tou urban planning district as a survey area. A questionnaire survey and Multinomial Logit Model, the facility type shelters were selected as a comparative variable. The results explained the school shelter choice significantly affected variables were routes choice, planning and design of school, hardware. However the square type shelter hardware availability was significant in evacuation choice. Weighted Voronoi diagram was further used to adjust the edge of the dominant area of each service shelter according to the location allocation criteria (public service accessibility, transportation, accommodation of shelter, open space ratio, service range population, fault line impact range). Then spatial econometrics was applied to reflect the spatial correlation of evacuation behavior.

The earthquake evacuation behavior of residents relates to geographical characteristics. The contingent plan considers proper disaster prevention living area for prompt reaction assumed the least nearest distances to safety shelter. This study considered the willingness of residents' evacuation and selected the reconstruction area after the 921 earthquake in Nan-tou urban planning district as a survey area. A questionnaire survey and Multinomial Logit Model, the facility type shelters were selected as a comparative variable. The results explained the school shelter choice significantly affected variables were routes choice, planning and design of school, hardware. However the square type shelter hardware availability was significant in evacuation choice. Weighted Voronoi diagram was further used to adjust the edge of the dominant area of each service shelter according to the location allocation criteria (public service accessibility, transportation, accommodation of shelter, open space ratio, service range population, fault line impact range). Then spatial econometrics was applied to reflect the spatial correlation of evacuation behavior.