Divergent metabolism estimates from dissolved oxygen and inorganic carbon: Implications for river carbon cycling

Rivers efficiently collect, process, and transport terrestrial-derived carbon. River ecosystem metabolism is the primary mechanism for processing carbon. Diel cycles of dissolved oxygen (DO) have been used for decades to infer river ecosystem metabolic rates, which are routinely used to predict metabolism of carbon dioxide (CO₂) with uncertainties of the assumed stoichiometry ranging by a factor of 4. Dissolved inorganic carbon (DIC) has been less used to directly infer metabolism because it is more difficult to quantify, involves the complexity of inorganic carbon speciation, and as shown in this study, likely requires a two-station approach. Here, we developed DIC metabolism models using single- and two-station approaches. We compared metabolism estimates based on simultaneous DO and DIC monitoring in the Upper Clark Fork River (USA), which also allowed us to estimate ecosystem-level photosynthetic and respiratory quotients (PQ_E and RQ_E). We observed that metabolism estimates from DIC varied more between single- and two-station approaches than estimates from DO. Due to carbonate buffering, CO₂ is slower to equilibrate with the atmosphere compared to DO, likely incorporating a longer distance of upstream heterogeneity. Reach-averaged PQ_E ranged from 1.5 to 2.0, while RQ_E ranged from 0.8 to 1.5. Gross primary production from DO was larger than that from DIC, as was net ecosystem production by $100\text{mmol}{m}^{-2}{d}^{-1}$. The river was autotrophic based on DO but heterotrophic based on DIC, complicating our understanding of how metabolism regulated CO₂ production. We suggest future studies simultaneously model metabolism from DO and DIC to understand carbon processing in rivers.