PARAFAC和FRI解析ISI中DOM分布

利用平行因子分析（PARAFAC）和荧光区域积分（FRI）的方法解析三维荧光光谱（3D-EEMs），结合主成分分析、相关性分析、聚类分析和多元线性回归分析，研究了曝气预处理改良土壤渗滤系统（ISI）处理生活污水时溶解性有机物（DOM）的垂直分布特征.根据FRI分析，ISI中DOM可以分为5个荧光区域，包括3个类蛋白物质区域（I、Ⅱ、IV）和2个类腐殖质物质区域（Ⅲ、V）.沿着垂直方向向下，ISI中DOM有溶出的现象，导致总荧光区域积分体积（TOT）与TN、TP、NH₄⁺-N、COD、TOC等都呈现显著负相关的关系，而与EC呈现显著正相关关系，其中荧光区域V与NO₃^--N浓度呈显著正相关的关系，氮素去除与DOM组成之间关系密切.通过进一步做PARAFAC分析表明，可以从DOM中识别出四种荧光组分，分别为C1类富里酸类物质和C2、C3、C4类蛋白类物质.荧光组分浓度得分值F_max表明，ISI对物质降解由易到难依次为C2 > C4 > C1、C3，即类酪氨酸最易降解，其次为类色氨酸类物质和类富里酸类物质.根据多元线性回归分析，可以用F_max间接表征TN、TP和COD等水质指标的浓度.

The distribution of DOM in aeration pretreatment improved soil infiltration system based on FRI and PARAFAC

It was studied direct distribution characteristics of dissolved organic matter (DOM) in the treatment of domestic sewage by aeration pretreatment of improved soil infiltration system using three-dimensional fluorescence spectroscopy (3D-EEMs) based on parallel factor analysis (PARAFAC) and fluorescence regional integration (FRI), combined with principal component analysis, correlation analysis, cluster analysis and multivariate linear regression analysis. According to FRI, DOM in soil infiltration system could be divided into five fluorescent regions, including three protein-like regions (I, Ⅱ, IV) and two humus-like regions (Ⅲ, V). Along the vertical direction downward, DOM dissolution in soil infiltration system leaded to a significant negative correlation between TOT and TN, TP, NH₄⁺-N, COD, TOC and a significant positive correlation with EC, in which fluorescence region V had a significant positive correlation with NO₃^--N concentration and nitrogen removal. It was closely related to the composition of DOM. Four fluorescent components could be extracted from DOM species by PARAFAC, namely C1type fulvic acid and C2, C3, C4 type protein. F_max showed that the degradation order of soil infiltration system was C2 > C4 > C1, C3. It meant that tyrosine was the most easily degraded, followed by tryptophan-like substances and fulvic acid-like substances. Results from our multivariate linear regression analysis suggested that the concentration of water quality indicators such as TN, TP and COD could be indirectly expressed by F_max.