時空不確定性因素之選址決策分析

Abstract

掩埋場場址選擇須考量很多法規準則及處理大量空間資料，以避免設置於不適當的地方，且應選擇對環境衝擊較低的地點，以降低風險。然而；如何依法規及準則篩選場址及如何由眾多場址中挑選出環境衝擊較低的場址是相當具有挑戰性的工作，尤其是環境衝擊具有時間與空間之不確定性，令掩埋場選址工作更為複雜，本研究因而發展一些決策分析方法與工具來改善選址。 判斷場址環境衝擊大小時，若數據具有時序隨機性且變動太大，則不宜單以平均值作為評估場址環境衝擊的依據。本研究因而結合馬可夫鏈法計算出事件的穩定發生機率及以模糊理論降低不確定性的影響，且結合地理資訊系統的空間分析功能，建立模糊馬可夫鏈選址方法。利用此方法示範如何由案例區初選場址中篩選出環境衝擊較低的場址。選址除了受時序不確定所影響，空間上亦具有不確定性，例如掩埋場所排放污染物在空間分佈上會受流體的流速與流向所影響，使污染物在不同方向上產生不同的濃度分佈，加上人口分佈不均，在各方向上造成不同的曝露風險衝擊。本研究因而採用空氣品質模式結合風向風速資料，建立各初選場址在各方向的污染濃度圖層，進而與人口密度圖層疊圖分析出各方向的風險圖層。依據各場址在各方向的風險值選擇場址，並結合地理資訊系統的空間分析功能，建立方向性風險選址方法。此方法可改善只考慮優勢流向的污染濃度大小的缺點，改善選址決策品質。 依據案例研究結果顯示模糊馬可夫鏈方法可判識出較不容易轉換至高潛在風險的區域；而方向性風險分析則除了可判識出高濃度高風險區域，亦可判斷出較低濃度但具有較高潛在風險的區域。所發展的方法在面對時間及空間的隨機性因素上可以較明確的表達各場址對周圍環境的衝擊大小，提供較適當的選址資訊改善選址決策品質。

During landfill site selection, a significant amount of spatial data with respect to various regulations, criteria, and rules must be processed, in order to avoid a site being built at an inappropriate location. An appropriate site should not have significant impacts on or risks to the surrounding environment. However, determining how to locate a site with low environmental impacts and risks is still a challenging task, especially when the temporal and spatial uncertainties of the environmental impacts and risks are considered. This study was thus initiated to develop methods and tools to deal with the uncertainties in making landfill siting decisions. In evaluating a factor causing any environmental impact from a candidate landfill site, if the data for the factor exhibits significant temporal fluctuation or uncertainty, evaluating the factor based on its average value may be inappropriate and misleading. This study thus developed a method applying the Markov chain to evaluate the probability of occurrence and a fuzzy approach to reduce the effect of the uncertainty. The method was further integrated with the spatial analysis function provided by a geographical information system (GIS) for siting a landfill. This fuzzy-Markov-chain method was demonstrated by using it to select sites with low potential risks. In addition to the temporal uncertainty, spatial uncertainties also exist for some siting factors. For instance, the distribution of air pollutants emitted from a landfill is greatly influenced by wind directions and speeds, causing different impacts depending on the direction and location. Exposure risks are also different for areas with different population densities. Therefore, this study applied an air quality model to simulate the pollutant concentration distribution and created a pollutant concentration layer for each direction of the candidate site. Then, a directional risk layer for each candidate site was produced from the pollutant concentration layer and the population density layer, using the spatial analysis function provided by a GIS. This directional risk method is expected to improve the quality of a siting decision and to avoid the problems that may arise from implementing a siting analysis primarily based on the prevailing wind direction. The results obtained from a case study reveal that the fuzzy-Markov-chain method can identify sites with low potential risk. Furthermore, the directional risk method can identify both the areas with high concentration and high potential risk and the areas with low concentration but high potential risk. The proposed methods can deal with temporal and spatial uncertainties effectively and provide proper information for assessing the environmental impacts and risks posed by a candidate site. Consequently, the proposed methods are expected to improve the quality of a landfill siting decision.