|摘要: ||竹溪位於台南市人口密集的東區，現階段主要功能做為都市區域性污水與雨水的主要排水渠道，排水區面積約7.44平方公里，總長度約10公里，直接流入台灣海峽。晴天時主要承受家庭污水，水質屬嚴重性污染，日平均排水流量約 23,000~44,000 m3/d。溪岸採混泥土施工，溪底坡度緩和，河深與水流速度單一化，加上水質惡化，因此水域環境之生態景觀條件相當貧乏。竹溪示範性人工溼地系統完成於2004年8月，由環保署提供經費，台南市政府執行建造。濕地系統建構於竹溪現有河岸之水泥河階地，進流部分的竹溪排水，主要設置目標為探討溼地生態工法現地淨化污染性都市排水的可行性及增進都市河川生態多樣性及景觀美化的功能，並由試驗結果提出竹溪污染整治的可行性與完整性的生態工法構想。竹溪濕地系統由三個單元串聯所組成，第一單元為表面下流動濕地(SSF，subsurface flow system，200 m2)，第二單元為表面流動式系統(FWS，free water surface system，600 m2)，第三單元為SSF濕地(200 m2)。濕地植物選擇兼具污染淨化及景觀美化的種類，如莎草、香浦、培地茅、風車草、美人焦。溼地進流水以沉浸式抽水機從竹溪取水，啟動階段(2004年8月至2005年3月)濕地系統進流流量平均控制於76 m3/day，水力負荷為0.076 m3/m2/day，水力停留時間平均為3.81天；之後(2005年3月至2006年1月)進流量提高到平均107 m3/day進行穩定階段的操作，此時水力負荷為0.107 m3/m2/day，水力停留時間約為2.66天。
啟動階段人工溼地系統對進流水主要污染物的平均去除效能分別為：總懸浮固體物(TSS) 75％ ( 2.31 g/m2/day)、生化需氧量(BOD5) 75％ ( 3.08 g/m2/day)、化學需氧量(COD) 55％ ( 4.22g /m2/day)、總氮(TN) 20% ( 0.45 g/m2/day)、總磷(TP) -6％( -0.0021 g/m2/day)。穩定操作階段的污染物的平均去除效能分別為：TSS 81％ (2.73 g/m2/day)、BOD5 79％ ( 4.17 g/m2/day)、COD 46％ (4.59 g/m2/day)、TN 24% (1.13g/m2/day)、TP 22％( 0.07 g/m2/day)。此結果顯示，人工溼地生態系統需經過一段適應期才能逐漸提升並表現穩定的污染物淨化效能，由其針對氮磷營養物的去除需更長的適應期。由一階柱塞流反應器模式(first order plug-flow reactor model)估算人工溼地系統啟動階段及穩定操作階段的BOD5去除速率常數平均分別為0.48 day-1(24℃)及0.75 day-1(25℃)，相當於或甚至高於一般河川水質自淨模式中的BOD5脫氧速率常數的0.232 day-1(20℃)。此結果顯示人工溼地生態系統的天然淨化功能可能高於河川的自淨功能，未來若將竹溪的水泥河階地全面性構築成人工溼地，並均散地分佈溪排水使流經溼地環境，預期將可有效的提昇竹溪的自淨功能，達到全面性水質淨化的目標。
Bamboo Creek, a typical urban river, is originated from the East District of Tainan City and offers an important function as a main drainage for both sewage and storm water runoff from urban areas. Its flow ends at Kun-Shen Bay, having a total length of about 10 km and drainage area of about 7.44 km2. Creek water is seriously polluted by the receiving sewage with a flow rate ranging from 23,000 to 44,000 m3/d when sunny days. The bank and bed of Bamboo Creek are mostly made of concrete; the slope of the creek bed is gentle, thus causing an unchanged water depth and velocity of water flow. Accordingly, ecological diversity and landscape esthetics along the creek are poor. The pilot-scale Bamboo Creek constructed wetland system was built in August 2004 by Tainan City Government under a project financially supported by Environmental Protection Administration. The treatment wetland system was constructed on the concrete bank of the creek and received part of the polluted creek water for purification. The goals of the wetland constructing project are to (1) investigate the performance of the pilot-scale constructed wetland for purification of the Bamboo Creek water, (2) examine the benefit of enhancement of biological diversity and landscape esthetics due to wetland construction, (3) propose feasible approaches for improvement in water quality of the whole Bamboo Creek. The Bamboo Creek wetland system has a channel configuration, which consists of a subsurface flow (SSF) unit (4 m × 50 m), a free water surface flow (FWS) unit (4 m × 150 m), and an SSF unit (4 m × 50 m) connecting in series. Macrophytes with functions of pollutants purification and landscape values, such as cattail, cyperus, vetiver grass, reed, and canna, were planted in wetlands. Creek water was pumped into the front part of the first wetland unit and then flowed through the system by gravity. The flow rate of influent was operated at 76 m3/day in average during the start-up phase from August 2004 to March 2005, and then increased to 107 m3/day in average afterward for a stable phase operation. This hydraulic condition represented an average hydraulic retention time (HRT) of the whole wetland system being 3.81 and 2.66 day for the start-up phase and the stable phase, respectively.
The monitoring results of influent-effluent water quality show that major pollutants in influent were significantly reduced by 75％ (2.31 g/m2/day) for total suspended solid (TSS), 75％ (3.08 g/m2/day) for biochemical oxygen demand (BOD5), 55％ (4.22 g/m2/day) for chemical oxygen demand (COD), and 20% for total nitrogen (TN) in the start-up phase; but, total phosphorous (TP) removal was insignificant. In the stable phase, pollutants removal of the wetland system increased to be 81％ (2.73 g/m2/day) for TSS, 79％ (4.17 g/m2/day) for BOD5, 46％ (4.59 g/m2/day) for COD, 24% (1.13 g/m2/day) for TN, 22％(0.07 g/m2/day) for TP. This founding suggests that constructed wetlands need an adaptation period, approximately 8 months in this study, to achieve a stable treatment performance particularly for nutrient removal. Using first order plug-flow reactor model, BOD5 removal rate constants of the whole wetland system were estimated to be 0.48 day-1 (at 24℃) and 0.75 day-1 (at 25℃) in the start-up and stable phase, respectively. These values were comparable to or even higher than the value of deoxygenation rate constant, 0.23 day-1 (at 20℃), that is normally used in the river water-quality model. This may indicate that a constructed wetland potentially provides higher performance in water quality improvement than a river does. Creation of more wetland zones using the existed creek bank areas is proposed to improve water quality of the whole Bamboo Creek. A cost-effective method to construct wetlands and distribute the creek water in wetlands by gravity is also proposed.