Reactive oxygen species (ROS) are omnipresent in natural aquatic environments, and play an important role in biogeochemical cycles. One of the dominant sources of ROS in surface waters was thought to be from dissolved organic matter (DOM) interacting with photochemical process. The properties of DOM were different between the flood and drought periods in lakes; yet, information on how these variations influence ROS photoproduction is unknown. Through a three-year study, the photochemical properties of DOM and the resultant ROS photoproduction between the flood and drought period were determined in the largest freshwater lake in China (Lake Poyang). Results found that quantum yield coefficients of excited triplets (³CDOM^*), apparent quantum yields of singlet oxygen (¹O₂) and hydroxyl radicals (·OH) were holistically higher in the flood period than those in the drought period. The optical properties of DOM showed that DOM in the flood period featured an allochthonous input, accompanied by higher molecular size (E2/E3), aromatic content (SUVA₂₅₄), humification degree (HIX), while DOM in the drought period was mainly internal input. Fourier transform ion cyclotron resonance mass spectrometry (FI-ICR MS) further revealed that some refractory components, such as lignin-like and carboxyl-rich alicyclic molecules (CRAM) presented higher abundance in the flood period, and played the positive impacts on ROS production. Orthogonal partial least squares (OPLS) were used to build novel multivariate predictive models for indicating the spatio-temporal ROS production. Also, the relatively higher steady-state concentrations of ³CDOM^* and ¹O₂ in the flood period could significantly diminish the half-lives of acetochlor. Considering the photochemical activity of DOM varied considerably at different periods, this study provided a new method to predict ROS production and contributed to a new insight into stage-specific emerging contaminants removing in natural aquatic environments.

Na Song, Dinggui Wu, Huacheng Xu, Helong Jiang, Water Research, 225, 2022, 119136, doi.org/10.1016/j.watres.2022.119136.