Relative contributions of different sources to DOM in Erhai Lake as revealed by PLS-PM

Dissolved organic matter (DOM) is a complex mixture of organic molecules that plays an important role in freshwater lake ecosystems. Excitation emission matrix-parallel factor analysis (EEM-PARAFAC) is an important tool for the identification and source analysis of DOM but is still inadequate for studying the differences and quantifying the contributions of different sources. Here, based on the maximum fluorescence intensities (Fmax) of the four fluorescent components (e.g., protein-like component C1, and the humic-like components C2, C3, C4) identified by EEM-PARAFAC, combined with large-scale data obtained from observations at Erhai Lake, we used partial least squares path modeling (PLS-PM) to improve the understanding of the migration and transformation mechanisms of DOM. The results showed that the phytoplankton and trophic state had greater impacts on DOM, while the impacts of sediment and inflow rivers were less significant. The results of the models that used nitrogen nutrients (N) and phosphorus nutrients (P) to separately indicate the trophic state suggested that the driving force of P on DOM was stronger than that of N in Erhai Lake. Among the four fluorescent components, the protein-like component with the largest relative proportion (41.98%) was mainly affected by phytoplankton, which was consistent with the autogenic characteristics obtained through spontaneous source index (BIX). The duality of the humic-like components was consistent with the duality of DOM through fluorescent index (FI). C3 had a higher sensitivity to the trophic state than the other components, and C2 received the greatest positive contribution from the rivers entering the lake. These results provide an improved insight into the influence of different sources on the behavior of DOM and demonstrate the potential of using PLS-PM to study the complex driving mechanism of aquatic biogeochemical parameters.