Mechanisms Underlying Near-Universal Export Patterns of Dissolved Carbon From Land to Rivers

Land-to-river exports of dissolved carbon, characterized through concentration-discharge (CQ) relationships, follow strikingly consistent, near-universal patterns: dissolved organic carbon (DOC) typically exhibits a flushing pattern (C increases with Q), while dissolved inorganic carbon (DIC) exhibits a dilution pattern (C decreases with Q). Another, albeit sparsely documented, universal pattern is their contrasting vertical distributions in subsurface water—DOC concentrations generally decrease with subsurface depth, whereas DIC concentrations increase. These observations prompt intriguing questions: What mechanisms underlie these near-universal patterns, and how are they interconnected? Here, we address these questions by carrying out numerical experiments across a wide range of conditions using a data-grounded, catchment-scale reactive transport model (BioRT-HBV). Results reveal that biogeochemical reactions governing the production and consumption of dissolved carbon dictate the direction of export patterns (flushing or dilution) by establishing vertical concentration gradients—quantified as the concentration ratios between shallow and deep waters (C_ratio). When biogeochemical reactions lead to higher concentrations in shallow soil waters than in deep waters (C_ratio > 1), solutes exhibit flushing patterns, and vice versa. Meanwhile, the relative contributions of shallow and deep flow regulate stream concentration variability, where greater deep flow inputs dampen fluctuations and push CQ relationships toward near-zero slopes. These distinct roles of depth-dependent flow paths and biogeochemical processes indicate the importance of subsurface physical and biogeochemical structures in shaping export patterns of dissolved carbon from land to rivers. As climate extremes and human activities intensify, subsurface structure and processes can change, reshaping the future of carbon cycling, water quality, and ecosystem health.