|摘要: ||近二十年來，由於密集農業的快速發展，化學肥料過量施用，造成氮化合物積存於自然環境中。其中有一部份以硝酸氮(NO3-N)形態溶入地面水及地下水中，並進而影響飲用水品質，造成飲水衛生上的警訊。本研究建立起一小型表面流人工溼地系統(microcosm)與一試驗規模(pilot-scale)人工溼地系統，以探討(1)種植不同植物之溼地組與對照組的NO3-N處理效能以及季節變化對NO3-N處理效能的影響，並藉由監測溼地底泥之各種特性，以探討小型溼地的NO3-N去除與底泥特性的關係；(2)表面流動式溼地(FWS)和表面下流動式溼地(SSF) 的NO3-N處理效能與其出流水質特性，以及操作不同水力負荷對NO3-N處理效能的影響與動
力學之表現。小型表面流人工溼地系統(每個溼地為長0.6 m、寬0.4 m、高0.45 m)以連續入流方式，水力負荷控制為0.035 m d-1，入流NO3-N與PO4-P目標濃度分別為為20 mg N L-1與5 mg P L-1之合成水。各溼地組分別種植水芙蓉、蘆葦、空心菜、水蠟燭、狼尾草，對照組分別為加上誘l以阻絕陽光之加遜儱茞梬P不加遜儱茞捸C除了水質的監測，為了了解溼地底泥的特性與水質的關係，本研究並監測溼地底泥之脫硝潛能、有機物含量、碳含量、pH、與氧化還原電位，以及底泥的氮含量與植物對於氮的攝取量。經由近兩年之實驗結果顯示，小型表面流人工溼地系統NO3-N之入流濃度為21.05±6.49 mg N L-1，種植植物(分別為水芙蓉、蘆葦、空心菜、水蠟燭、狼尾草)溼地組的NO3-N去除效能(72-98%)比無種植植物之對照組優越(1-33%)，溼地組NO3-N去除效能受季節變化的影響也不如對照組明顯。而各組底泥之NO3-N去除速率與脫硝潛能的大小呈正相關，而脫硝潛能與底泥底泥之STOC含量具明顯之線性正比例關係，脫硝潛能與底泥之ORP與pH值呈現負相關。以上結果顯示，能提供較還原態條件及較多的有機物作為脫硝碳源的溼地，可表現較高的NO3-N去除能力，而溼地中的植物對提供此脫硝環境是有助益的。另外，由氮質量平衡結果估算種植植物之溼地的底泥氮累積速率為0~0.163 g TN m-2 d-1，約佔溼地總氮去除的0~26.6%，而植物對於總氮的攝取速率為0.004~0.126 g TN m-2 d-1，佔該組溼地總氮去除量的0.8~26.8%。此結果顯示脫硝作用對於整個溼地的NO3-N去除相當重要。試驗規模人工溼地系統分別入流含NO3-N之地下水(目標濃度為20 mg N L-1)於FWS與SSF溼地(每座均為長5m、寬1m、高0.8m)，並調整不同水力負荷(範圍0.02-0.27 m d-1)，以及觀察溼地中植物的生長與底泥氮含量。經由一年之實驗結果顯示，當試驗規模人工溼地系統開始操作之後，SSF溼地的啟動適應期比FWS溼地短，並且在出流水有機物濃度及TSS濃度等表現均較FWS溼地低。當水力負荷不斷提升之後，FWS與SSF溼地的NO3-N去除速率也隨之升高，並於水力負荷控制為0.12 m d-1時達到最高(FWS與SSF分別為1.288與1.372 g N m-2d-1)，隨後水力負荷調整為0.25 m d-1時，FWS與SSF溼地的NO3-N去除速率反而下降。以零與一階柱塞流動力學模式描述NO3-N在人工溼地中的去除行為，結果發現一階柱塞流動力學模式與實驗結果的契合度較高，FWS 溼地一階面積去除速率常數(k1)介於0.018-0.093 md-1，SSF溼地介於0.027-0.135 m d-1。另外，於監測期間FWS溼地底泥的總氮含量並無明顯累積的趨勢(P=0.7211)，經由氮質量平衡估算發現，雖然植物的氮攝取對於溼地氮的去除具有部分貢獻，然而脫硝作用仍佔了溼地氮去除相當大的比例，顯示脫硝作用為溼地去除氮的主要機制。
Nitrate contamination of groundwater normally results from intensive use of fertilizers in agriculture. Such contamination can lead to health problems via drinking water contamination and is an environment concern in many areas. This study set up seven microcosm wetlands with free water surface flow (FWS) and tow pilot-scale wetlands (an FWS and a subsurface flow (SSF) type) to investigate nitrate removal from groundwater in constructed wetland. The aims of the study included: (1) determine nitrate removal of the microcosm wetlands, effect of seasonal variation, and the relationship between nitrate removal and sediment characteristic; (2) make a comparison of nitrate removal between FWS and SSF wetlands, and investigate their kinetic behaviors under various hydraulic loading rates. In the seven microcosm wetlands (each measured 0.6 m×0.4 m×0.45 m), five were planted with various native machrophytes (Pistia stratiotes, Phragmites australis, Ipomoea aquatica, Typha orientalis, and Pennisetum purpureum), whereas the others were unplanted (one with black plastic cover and another without cover). A simulated nitrate-contaminated groundwater containing around 20 mg N/l of nitrate and 3 mg P/l of phosphate were continuously fed into the microcosm wetlands at a hydraulic retention time (HRT) of around 4 d. Besides water sampling and analysis, the characteristics of wetland sediments including denitrification activity, organic matter content, carbon content, pH, and oxidation-reduction potential (ORP) were measured. Throughout two years continuous operation, the influent-effluent data showed that planted wetlands reduced influent nitrate (21.05±6.49 mg N L-1) averagely by 72~98%. These nitrate removals were all significantly higher than those in two unplanted wetlands, in which the covered wetland was 1% and uncovered one was 33%. Seasonal effect on nitrate removal was more obvious in planted wetlands than unplanted wetlands. Additionally, the denitrification activities of sediment in planted wetlands (6.3~13.1 ug N/g/h) were also significantly higher than those in unplanted wetland (1.8~11.2 ug/g/h). Nitrate removal rate of microcosm increased proportionally with the increase of sediment denitrification activity; also, the denitrification activity increased with the increase of organic matter or carbon content of sediments. Furthermore, sediment denitrification activity correlated conversely to the sediment ORP and pH measured in situ. These results suggest that microcosms providing more reduced condition and greater organic carbon available for denitrification can perform higher nitrate removal ability. The presence of wetland macrophyte was advantageous to create denitrification conditions and support nitrate removal. The results of nitrogen balance in planted wetlands showed that nitrogen accumulation in sediment and nitrogen uptake by harvested plant accounted for 0~0.163 and 0.004~0.126 g TN m-2 d-1, respectively, which represented 0~26.6% and 0.8~26.8% of overall nitrogen removal. This implies that denitrification is the most essential part of nitrogen removal in wetlands. In pilot-scale wetlands study, FWS and SSF wetlands (each measured 5m××1m×0. 8m) received a simulating nitrate contaminated groundwater at a target concentration of 20 mg NO3-N L-1 under various hydraulic loading rates ranged from 0.02 to 0.27 m d-1. SSF wetland seemed to need shorter period to achieve stable performance than FWS wetland; also, the levels of total organic carbon ( TOC) and total suspended solid (TSS) in SSF wetland effluent were significantly higher than FWS wetland effluent. Nitrate removal rates of both wetlands increased with increasing hydraulic loading rate until a plateau was reached; the maximum removal rates, occurred at hydraulic loading rate of 0.12 m d-1, were 1.288 and 1.372 g N m-2 d-1 for FWS and SSF wetland, respectively. Afterward, removal rate decreased when hydraulic loading rate increased to 0. 25 m d-1. First-order nitrate removal kinetic was found to be more sufficient for fitting concentration decrease profile than zero-order kinetic. Removal rate constant values were higher for the SSF wetland (0.027-0.135 m d-1) than the FWS wetland (0.018-0.093 m d-1), probably because SSF wetland provided more surface area for attached growth of denitrifying bacteria than the FWS wetland. Nitrogen accumulation in FWS sediment was insignificant (P=0.7211). Nitrogen uptake by plant growth accounted for up to 12.5 and 9.2 % of total nitrogen removal in FWS and SSF wetland, respectively. Denitrification was still the major contribution for nitrate removal in the pilot-scale wetlands study.