多种源解析方法在水体硝态氮污染溯源中的应用研究

Pollutant source tracking is to identify and apportion the sources and estimate the contributions of respective sources to the overall pollution, which is the key step to pollution control. The two most common ways for source tracking are (1) using the isotope signature of different sources of pollutants in conjunction with mathematic models to estimate the amount of pollutants from different sources; (2) using information other than the targeted pollutant itself to estimate the pollutant sources and their respective contributions. The latter such as Positive Matrix Factorization (PMF) model has been applied to soil and air pollution study, but has not been used in water pollution research. Through greenhouse experiment, this research is to investigate the nitrate isotope dynamics of the leachates from soil columns applied with known stable isotope ratio of chemical fertilizer, organic fertilizer, and natural soil under wetting-drying cycles, and use the models such as Isotope Mixing Model (IMM) and Stable Isotope Analysis in R (SIAR) to calculate the contributions of soil, chemical fertilizer and organic fertilizer to the nitrate in leachates. The data will then be used to verify the applicability of the PMF model in nitrate source tracking of polluted waters, and the tested PMF model will then be applied to identify the nitrogen sources in nitrate polluted waters from a watershed.

利用水体污染源解析方法追溯污染物的来源以及各来源对污染物的贡献率，是控制水污染的关键第一步。目前最常用的方法有两种：(1)利用不同污染来源中污染物质的同位素特征的差异来判断污染物来源，再通过数学模型对各种污染源的贡献率进行分摊；(2)利用随某种污染物质迁移的其他物质的信息来判断该污染物质的来源及其贡献率。前者用同位素法解析水体中NO3-的污染源已得到较多运用，而后者例如PMF模型虽然有其优点，且在土壤尤其是大气污染源解析中有了较多的应用，但在水体中的研究较少。本研究利用三种已知稳定同位素比值的氮素来源（土壤、化肥、有机肥）进行室内可控试验，分析干湿交替条件下三种氮源经土体后的氮氧稳定同位素比值的变化，并运用IMM和SIAR模型计算各氮素污染源的贡献率，用于验证PMF在水体污染源解析中的可行性。验证后的模型将与同位素法一起用于流域尺度的水体硝态氮污染源解析。

水污染问题在世界范围内引起极大重视。在我国，近几十年来随着人口的增长、经济的快速发展及城市化进程的加快，地表水、地下水污染问题变得尤为突出。了解水污染的总体状况、污染分布的时间和空间特点以及污染来源的解析是进行水污染治理的前提。本研究在流域尺度上对水污染问题进行研究。研究区是浙江省典型流域合溪水库流域，受氮素污染的饮用水源区。在合溪水库流域采集河流点、地下水点和水库点等水样进行水质分析，以单因子水质评价、综合水质评价以及污染指数法等水质评价方法对流域的污染现状进行评价，并对流域内的各个水质因子的污染状况进行时空分布特征分析。此外，利用主成分分析结合绝对主成分得分多元线性回归模型，以及正定矩阵因子分解模型研究区域的水质污染来源进行综合解析，同时利用氮氧双稳定同位素和基于贝叶斯算法的污染源解析模型SIAR对合溪水库的硝酸盐污染来源进行识别并计算各个污染源的贡献率。研究结果可为浙江省流域的水质治理源头控制提供理论基础。