Characterizing humic substances from a large-scale lake with irrigation return flows using 3DEEM-PARAFAC with CART and 2D-COS

The study was intended to determine whether the composition or distribution of humic substances from lake sediment could be characterized by spectroscopic data with chemometrics methods. The research materials were humic substances extracted from sediment samples collected in heavily polluted areas in Wuliangsuhai Lake. Humic substance extraction was performed using the method proposed by the International Humic Substances Society (IHSS). In extracted humic substances, the three-dimensional excitation–emission matrix (3DEEM) was determined, and parallel factor analysis (PARAFAC) with classification and regression trees (CART) and two-dimensional correlation spectra (2D-COS) was used to analyze spectroscopic data. Both the fulvic acid (FA) and the humic acid (HA) samples had five PARAFAC components (C1–C5), which contained Vis fulvic-like component (C1), Vis humic-like component (C2), UV humic-like component (C3), microbial fluorescent component (C4), and UV fulvic-like component (C5). This indicated the ubiquity of the humic substance components in the environment. For the CART analysis, C1 was the key factor used to differentiate FA and HA, while C1–C4 were the key parameters used to distinguish the humic substance samples in the eight units. Using 2D-COS and hetero 2D-COS, the PARAFAC components were formed with the increasing of sediment depth in the following order: C3 → C2 → C5 → C4 → C1, which suggested that the relatively labile structure appeared earlier than the relatively stable structure with the greater depth of the sediment. The Vis fulvic-like could be a key factor used to discriminate the FA and the HA fractions, while the latent factors of different depths of sediment were not identical. The order of the humic substance formations was Vis humic-like → UV humic-like → Vis fulvic-like → microbial fluorophore → UV fulvic-like.