跳到主要內容 :::

年度	110	專案性質	非實驗性質
專案類別	研究專案	研究主題	自訂
申請機構	淡江大學學校財團法人淡江大學	申請系所	水環境資訊研究中心
專案主持人	王聖瑋	職等/職稱	助理教授
專案中文名稱	建構富砷地下水濃度變化預測模型與用水安全調適方案
中文關鍵字	地下水, 砷, 污染管理
專案英文名稱	Development of groundwater arsenic variation prediction model and adaptation plan for water safety usage
英文關鍵字	Groundwater, Arsenic, Contaminant management
執行金額	880,000元
執行期間	2021/3/19 至 2022/2/28
計畫中文摘要	近50年來，地下水砷污染問題已受到國際間高度關注，其主要影響地下水用水安全、農業灌溉與作物生長、人體健康危害等面向，而影響地下水砷濃度釋出過程與分布變化主要受到水文地質條件、生物地質化學反應、人為活動等因素的影響，臺灣地下水富砷問題由來已久，地下水富砷問題屬自然沉積環境還原反應伴隨產生之現象，並與地質沉積年代、質地、氧化還原條件等因素有關，富砷之鐵/錳氧化物為地層中砷的主要富集來源，臺灣西南部平原地區普遍且大量抽取地下水做為農業用途，已造成砷濃度累積於農地土壤。由於地下水砷釋出機制主要受到氧化還原條件影響，而自然條件之降雨入滲、河川補注，及人為活動之抽取地下水皆可能改變區域地下水氧化還原條件，目前雲林彰化地區地下水水井數量高達25萬口以上，農業灌溉期間的抽水致使地下水水位下降，且抽水過程導致地表氧氣進入淺層地下水，進而影響砷釋出至地下水之速率，因此，本計畫研究目的為探討抽用地下水造成砷濃度變化程度，建立地下水砷濃度變化預測模型，以強化水資源使用與管理之平衡關係，本計畫優先針對濁水溪沖積扇做為地下水砷濃度變化預測模型建立之研究區域，並收集地下水水位、地下水井資料、及降雨量做為模式基礎，由於地下水砷濃度僅在部分區域有明顯變化，故篩選9口目標監測井做為本計畫主要探討目標。由於各鄉鎮水井型態不同，導致抽水量無法以經驗公式進行轉換，故本計畫以用電量直接當成用水量做為模式模擬數據，主要研究方法為應用動態因子分析進行時序資料維度縮減與趨勢歸納，並找出監測井之最佳影響水井距離及名單，再利用隨機森林、線性迴歸、支持向量機、人工神經網路等機器學習方法，建立地下水砷濃度變化預測模型，再權重計算方式找出影響監測井砷濃度之重要特徵水井名單。本計畫執行成果提供評析各項調適與因應方案之可行性與效益，包括提供行政院環境保護署及農業委員會可能影響砷濃度變化之鄰近水井及特徵水井名單，提供做為定期監測、推動輔導作物轉作、抽測等之優先順序規劃參考名單，以確保未來土水環境品質與食品安全。
計畫英文摘要	In the recently 50 years, groundwater arsenic pollution has received great international attention, which affects groundwater safety, agricultural irrigation and crop growth, and human health risk. Primary mechanism of groundwater As release and mobilization is mainly affected by hydrogeological conditions, biogeochemical reactions, human activities and other factors. The release mechanisms of arsenic-contained groundwater is the reduction reaction of the hydrogeochemical environment including the age, texture, and redox conditions of the geological deposition. Arsenic-rich Fe/Mn oxides are the main source of arsenic in geological formation. However, amount of groundwater is commonly extracted for agricultural purposes in the southwestern plains of Taiwan, which has caused arsenic concentrations to accumulate in agricultural soil. Since the arsenic release mechanism from groundwater is mainly affected by redox conditions, hydrological conditions such as rainfall infiltration, river refills, and extraction of groundwater may cause the variation of groundwater redox conditions. At present, the extended private wells in Yunlin and Changhua counties are as high as 250,000. Irrigation pumping during agricultural demand causes the groundwater level decrease, and also the intrusion of oxygen to shallow aquifer, which affects the rate of arsenic release to the groundwater. Therefore, the purposes of this project are to investigate the changes in arsenic concentration caused by pumping groundwater, to establish groundwater arsenic prediction model, and to strengthen the relationship between water resources utilization and management. Chuoshui river alluvial fan is collected as the study area to establish the groundwater arsenic prediction model. Groundwater level, private wells parameters, and rainfall records are adopted as the modeling data. Arsenic concentration in groundwater only changed significantly in some areas, so selected 9 target monitoring wells as the main research target. Furthermore, the types of private wells were different in each township, so this project used the electricity consumption directly as the water consumption in the model simulation data. Dynamic factor analysis will then be applied to conduct dimension-reduction of time-series data and to extract the trends of groundwater As variation. Moreover, machine learning approaches, including random forest, linear regression, support vector machine, artificial neural network will be tested for As prediction. The accuracy of the groundwater arsenic concentration prediction model is established and verified by the water quality data. According to the groundwater As dynamic factor analysis and prediction model, the list of private wells with best distances and important characteristics affecting As concentrations can be selected. Adaptation strategies of groundwater usage and management will be proposed to Environmental Protection Administration and Council of Agriculture, and to ensure the quality of soil and water environment and food safety.