Photochemistry of dissolved organic matter in water from the Pearl river (China): Seasonal patterns and predictive modelling.

ABSTRACT

Photochemical properties of dissolved organic matter (DOM) vary widely in natural and engineered water systems due to the different dominant compositions. However, seasonal patterns of DOM photochemical properties in urban rivers remain unclear. In this study, two seasons (wet and dry) of water samples were collected from eleven sites throughout the Pearl River (China) to investigate the spatiotemporal variability of DOM optical and photochemical properties. The optical properties of DOM in the Pearl River were characterized by UV-vis and fluorescence spectroscopies, which showed the substantial decrease in absorption coefficient and fluorescence intensity and increase in absorbence ratio (E2/E3) and specific absorption coefficient (SUVA) from the wet to dry season. The photochemical properties in terms of the apparent quantum yields of 3DOM*, 1O2 and ·OH from DOM (Φ3DOM*, Φ1O2 and Φ·OH, DOM) under illumination also displayed a significant decrease from the wet to the dry season. Spearman's rank correlation analysis revealed the strongest relationships between Φ3DOM*, Φ1O2 and Φ·OH, DOM and the relative abundance of microbial humic-like component (C2%) derived from parallel factor analysis (PARAFAC). Partial least squares regression (PLSR) modelling exhibited an excellent prediction strength for steady-state concentrations of 1O2 ([1O2]ss) and ·OH ([·OH]ss) with adjusted R2 values of 0.85 and 0.91, respectively, by using DOC concentration ([DOC]), optical properties, nitrate and nitrite concentrations as the response variables. In addition, the model identified that the Fmax of humic-like component C4 (Fmax-C4) was the most effective predictor amongst the used response variables. This study provides an approach to describe and predict the seasonal patterns of DOM photochemical properties in urbanized rivers, offering a good understanding of the formation mechanism of reactive species from river DOM.