【作者】 张岩；

【导师】 崔丽娟；

【作者基本信息】 中国林业科学研究院 ， 生态学， 2014， 博士

【摘要】 目前，河道、湖泊等天然水体的自净能力逐渐降低，大量未经处理的工农业及生活污水的过度排放导致水体富营养化严重。作为传统污水处理技术的一种有效替代方案，人工湿地是在天然湿地基础上发展起来的污水处理技术，因具有投资少、耗能低、运行和维护简单等优点而被用于多种水体的净化过程。复合人工湿地通过整合不同类型人工湿地生态功能，充分发挥其功能特点，改善自然景观的同时能够更有效地去除污染物，且更具稳定性和耐冲击能力。目前针对人工湿地不同学者开展了理论和基础性研究，然而，污染物在人工湿地中的动力学过程尚未得到全面、细致的了解和掌握，人工湿地的设计和运行多基于经验数据，对长期运行效果缺乏准确可靠的预测方式，从而使得人工湿地的生态功能无法最大限度发挥，人工湿地技术的应用和发展受到局限。磷是水体富营养化的主要限制因子，本研究针对间歇式进水自由表面流-水平潜流复合人工湿地，通过探究研究区复合人工湿地基质中磷的主要存在形态及分布规律，从污染物迁移规律出发，分析了复合人工湿地基质对磷的吸附/解吸特性；从水力学角度出发，通过示踪实验研究了水力停留时间及其分布特征，分析了复合人工湿地的水力效率；从反应器流动特性和污染物降解一级动力学模型角度研究了污染物降解规律，估算了适用于研究区间歇式进水复合人工湿地的一级动力学参数；以污染物迁移转化规律和STELLA动力学软件为基础，通过对水位、植被、悬浮颗粒物浓度变化进行模拟和整合，构建了磷去除的动力学过程模型，量化了磷在复合人工湿地内的动力学过程。通过对表层（0~10cm）和次表层（10~20cm）基质中不同形态磷的提取和测定，分析了磷在复合人工湿地中的形态分布特征，探究了不同形态磷与粒径和有机质之间的相关关系。结果表明：复合人工湿地基质中总磷含量为139.73~242.03mg/kg，各处理单元表层和次表层基质中总磷含量无显著差异（p>0.05），分别为198.15±32.45mg/kg和160.44±15.62mg/kg，无机磷是基质中磷的主要存在形态，占总磷的71.83~97.28%，其中钙磷是无机磷的主要组成部分（65.17~98.71%）；基质粒径差异对钙磷的含量影响不大，可吸附态磷和有机磷在粒径小于2μm的基质中的含量较大，铁磷、有机磷和有机质之间存在较强的正相关性，钙磷、可吸附态磷和有机质之间呈负相关性。通过室内静态实验，以复合人工湿地基质为研究对象，对磷在基质-上覆水界面的动力学过程进行了模拟，分析了基质对磷的吸附/解吸特性。结果表明：基质对磷存在净吸附过程，吸附平衡浓度为4.23±1.77~22.52±2.71mg/L；各处理单元基质对磷存在化学吸附作用，且在吸附反应开始后的20min内吸附速率最快，并在24h内逐渐达到动态平衡状态，继续延长水力停留时间不会增加吸附量；各处理单元基质对磷具有较强的吸附能力（k：250.23~403.28L/mg），最大吸附量为286.82~604.63mg/kg。通过野外示踪实验分析了复合人工湿地的水力停留时间及其分布特征。结果表明：通过在进水口设置阻隔以抬升进水水位，可有效提高示踪剂的扩散度，从而能够提高示踪剂的回收率。采用修正的Gauss模型能够较好地拟合复合人工湿地不同运行工况下水力停留时间的分布特征，且拟合参数因植被类型和进水方式不同而有所差异。与其它处理单元相比，浮水植物处理单元B和F内存在较多的短流区，而潜流和挺水植物处理单元中死区的数量对运行效果无显著影响（ε：0.73~0.79），各处理单元具有较强的水流混合程度（N：1.6~2.1），水力效率维持在23~43%范围内。通过室内模拟和野外监测相结合，采用k-C*模型对污染物降解的一级动力学特征参数进行了拟合。结果表明：进水浓度与背景浓度和降解速率常数之间呈正相关关系，各处理单元中磷背景浓度为0.004±0.003~0.019±0.015mg/L，在估计背景浓度下污染物降解速率常数大于零背景值时降解速率常数，平均降解速率常数分别为0.69±0.48m/d和0.55±0.37m/d。随系统运行时间的增加，污染物降解速率常数逐渐增加。各处理单元降解速率常数随温度增加而降低，不同处理单元θk无显著差异（0.936~0.957），在20℃下的污染物降解速率常数为0.366~0.925m/d，且RE>CR>CA>CT。背景浓度随温度变化波动较小，温度与处理单元CA、CR和CT的背景浓度之间存在一定的负相关性，而与处理单元RE之间呈正相关，各处理单元中θc*为0.935~1.029，且RE>CR>CA>CT。采用动力学模型软件STELLA对复合人工湿地的水位、植被、悬浮颗粒物的动力学过程进行了描述，通过构建和整合子模型对不同组分之间磷的迁移转化过程进行了量化。结果表明：研究区复合人工湿地能够有效去除进水中89.1%的总磷，不同处理单元对总磷的平均去除率存在差异（0.25~2.58g/m2/d）。该复合人工湿地主要通过理化沉降方式除磷，而进水中有近1/4的磷通过有机质分解和再悬浮过程被重新释放到水体中，平均释放速率为0.3~6.9g/m2/d。水生植物从基质中吸收的磷大部分会通过自身代谢过程重新沉积于基质中，通过收割植物，复合人工湿地能够以0.3~7.9g/m2/d的速率去除进水中43.1%的总磷。本研究中所构建的磷去除动力学过程模型能够基本反应研究区磷动力学变化过程，模型对出水浓度变化的拟合值与观测值基本一致，然而由于模型的构建和校正基于研究区定期监测数据，且系统稳态的假设与实际运行条件存在差异，模型的应用具有一定地域性，模型的泛化能力有待于进一步研究。更多还原

【Abstract】 Currently, the self-purification ability of natural water bodies such as rivers and lakes is declining. The random discharge of great amount of agricultural and industrial wastewaters without any treatment for contaminants caused serious eutrophication problems. Phosphorus has been proven to be an important limiting factor. As an effective alternative for traditional treatment technologies, constructed wetland (CW) was developed gradually on the basis of ecological functions of natural wetlands and was used increasingly for the treatment of different kinds of wastewater regarding to the advantages of lower cost, lower energy consumption, convenient operating and maintaining conditions.Integrated constructed wetland was proposed as an integration of the landscape and ecological functions of different kinds of CWs. Combing the advantages of different CWs,integrated constructed wetland has a higher stability and anti-impact ability than the individual equipment. Currently, great effort has been applied to researches about CWs. However, it is still indistinct on the contaminant dynamics in the system. Moreover, the design and construction of CWs was mainly on the basis of empirical dataset, resultingin a decline of the modeling accuracy. As a result, the ecological function of CWs could not be exhibited and the application and development of CWs was limited.From the aspect of existence forms of phosphorus in the CW with intermittent feeding water, the distributions of phosphorus in the sediment were analyzed. Based on the discipline of transference, the adsorption and desorption characteristics at the sediment-water interface were analyzed and critical parameters were estimated based on datasetfrom incubation experiments. With the method of trace experiment in the field, the actual hydraulic retention time was estimated and its distributions were simulated by the revised Gauss model. The actual flow pattern in the CW was described and the hydraulic efficiency was calculated. Based on the theory of contaminant reactor and the degrad ation first order kinetics, relative first order removal parameters for the removal of phosphorus were estimated. Through the use of dynamic software package STELLA, four submodels including the hydrology submodel, the above-ground plant biomass submodel,the suspended solid submodel and the phosphorus submodel were developed and the phosphorus dynamics in the system was quantified.Surface (0~10cm) and subsurface (10~20cm) sediment samples were collected andforms of phosphorus were extracted and analyzed in the laboratory. Impacts of particlesize and organic matter on the distributions of phosphorus were analyzed. The results indicated that the amount of total phosphorus in the sediment ranged from139.73mg/kgto242.03mg/kg. No significant difference was found between the amount of phosphorus in the surface and subsurface sediments (p>0.05) and the amount of phosphorus in the surface and subsurface sediments was198.15±32.45mg/kg and160.44±15.62mg/kg respectively. Inorganic phosphorus comprising of71.83~97.28%in the total phosphorus, was found to be the main existence form in the sediment. The phosphorus extracted bycalcium (65.17~98.71%) contributes to the most of inorganic phosphorus. The particle size has no significant impact on the distributions of phosphorus extracted by calcium. The amount of absorbable and organic phosphorus in sediments with fine particle size (φ<2μm) was greater, while the amount of phosphorus extracted by iron was smaller compared with that in other mediate. Positive relationship was observed between the amount of iron and organic phosphorus, and the organic matter, while an opposite relationshipwas observed between the calcium extracted and absorbable phosphorus, and organicmatter.By using the sediment samples collected from the CW system, batches of incubation experiments were designed in the laboratory to analyze the phosphorus adsorption/desorption characteristics at the sediment/water interface. The results indicated that net adsorption processes exist for phosphorus. The equilibrium concentration was4.23±1.77~22.52±2.71mg/L. Besides of surface adsorption, a chemical adsorption mechanism exists. T he maximum adsorption velocity was found during the first20min. The experiments reached an equilibrium status after24h and proposes intended to prolong or shorten the hydraulic retention time contribute little to the improvement of phosphorus adsorption. The sediment in the CW system has a high phosphorus adsorption ability as indicated by the high k values of250.23~403.28L/mg. The maximum phosphorus adsorptions were286.82~604.63mg/kg.Field tracer experiments were performed in order to estimate the actual hydraulic retention time and to describe its distributions. The tanks in series model was used to simulate the field datasets and analyze the flow characteristics in the CW system with intermittent feeding water. The results showed that a higher inlet elevation may favor the dispersion of tracer and lead to a higher tracer recovery. The revised Gauss model was efficient in simulating the distributions of hydraulic retention time under different operating conditions. Model parameters were different as the plant species and inlet design varies. Compared with other treatment ponds, B and F have more short flow areas.The death districts in the subsurface and macrophyte treatment ponds had no significanteffects on the removal of phosphorus (ε:0.73~0.79). A high mixture degree of flow existed in the system as indicated by the high N values of1.6~2.1. The hydraulic efficiency of the CW was23~43%.In the case of optimum operating parameters, the first-order kinetic model was used to fit data collected from the field and microcosm studies. The results showed that the influent concentration was positively correlated to the background concentration andthe removal rate constant. The averaged background concentrations for phosphorus in the CW were0.004±0.003~0.019±0.015mg/L. The removal rate constants were calculatedhigher at the estimated C*values than the values calculated at a zero background concentration. The averaged removal rate constants at the estimated C*values and zeros were0.69±0.48m/d and0.55±0.37m/d respectively. The removal rate constant declined as the system operated. The k values declined as the temperature increased during the micr ocosm studies. No significant difference was observed for the calculated θkvalues (0.936~0.957). The k20values for different treatment cells were0.366~0.925m/d and were ranked as RE>CR>CA>CT. The C*has little fluctuations as the temperature changed. TheC*in treatment cells CA, CR, CT were negatively correlated to the temperature. While the C*in treatment cell RE changed positively with the temperature variations. The averaged θc*values for different treatment cells were0.935~1.029and were ranked as RE>CR>CA>CT.The dynamic model software package STELLA was used to simulate the dynamicprocesses of water elevation, vegetation and suspended solid in the CW system. The transformation of phosphorus among different parts was quantified by integrating differentsubmodels. The results showed that the CW system in this present study has the capacity to sequester89.1%of the total phosphorus in the influent. The averaged areal phosphorus removal in different treatment ponds ranged from0.25g/m2/d to2.58g/m2/d. Themain mechanism for the removal of phosphorus in the system was sedimentation. However, the amount of phosphorus released through the processes of degradation and suspension of organic matter was about1/4. The averaged releasing velocity was0.3~6.9g/m2/d. Macrophytes play an important role in the removal of phosphorus. While most of the phosphorus absorbed by macrophytes were returned to the sediment environment during their metabolism. By harvesting the macrophyte, about43.1%of the total phosphorus could be removed at the absorption velocity of0.3~7.9g/m2/d. The phosphorus dynamic model developed in this study produced a reasonable match between the predictedand observed effluent concentration values. However, the generalization of this model should be studied further because of the discrepancy between the model hypothesizes and actual operating conditions.更多还原