Scale-dependent shifts in climatic and anthropogenic drivers of riverine dissolved organic matter dynamics

Dissolved organic matter (DOM) is central to the global carbon cycle, yet how urbanization interacts with climate to shape riverine DOM dynamics across scales remains poorly understood, limiting our ability to predict aquatic carbon feedbacks under global change. Here, we integrate a five-year intensive monitoring dataset from a representative peri‑urban watershed (Zhangxi River, eastern China) with a global meta-analysis encompassing 1844 data points across six continents, to disentangle the multiscale drivers of riverine DOM based on its optical properties (UV-Vis absorbance and fluorescence). We show that urbanization is the strongest driver in our regional dataset, whereas climate‑associated drivers and topography emerge as dominant drivers in the global meta‑analysis. This scale‑dependent pattern suggests a potential shift, which we propose as the Scale‑Dependent Driver Shift (SDDS) hypothesis requiring further independent testing. Urbanization consistently elevates DOC concentrations and shifts DOM composition toward protein‑rich, microbially derived components with lower aromaticity and higher bioavailability across watershed gradients. Globally, urban watersheds exhibit recurrent anthropogenic DOM signatures across temperate and subtropical regions characterized by enhanced lability and reduced compositional complexity compared to natural systems, though we caution that the global dataset is biased toward mid‑latitude regions, and extrapolation to tropical or wetland systems warrants further investigation. Random forest and variance partitioning analyses supported a scale‑dependent driver shift: human activities emerged as dominant predictors regionally, while climatic and topographic constraints structured global biogeographic patterns. Based on these findings, we propose the SDDS as a working hypothesis that offers a potential framework to reconcile long‑standing discrepancies between climate‑focused and human‑focused studies. Our results suggest that ongoing urbanization may decouple riverine carbon dynamics from natural climate‑associated drivers, with potential consequences for carbon mineralization, greenhouse gas emissions, and global carbon cycle feedbacks.