Improving Carbonate Equilibria-Based Estimation of pCO2 in Anthropogenically Impacted River Systems

Abstract

Estimating riverine carbon dioxide (CO2) emissions has been constrained by lacking field measurements of the partial pressure of CO2 (pCO2) and inaccuracies in calculating pCO2 using carbonate equilibria-based models such as CO2SYS. To evaluate potential errors in applying the carbonate equilibria-based pCO2 calculation to river systems affected by monsoon rainfall and water pollution, we compared pCO2 values calculated using CO2SYS and those measured by headspace equilibration in five Asian rivers (Ganges, Mekong, Yangtze, Yellow, and Han rivers) undergoing various water pollution stages. Across the five rivers, calculated and measured pCO2 values exhibited larger discrepancies during the monsoon season, particularly in the low pH range, while in the Han River mismatches were also noticeable during the dry season. In the Han River, pH was negatively correlated with dissolved organic carbon (DOC) during the monsoon, indicating organic acids flushed from soils during rainfalls as a key factor for overestimated pCO2 at sites with low pH and alkalinity, whereas dry-season overestimation of pCO2 may be ascribed to non-carbonate alkalinity including organic acids and inorganic anions delivered by wastewater effluents or sporadic rainfalls. The four large rivers exhibited a positive correlation between pH and DOC in tributaries during the monsoon season, indicating that DOC flushed from soils may be diluted by monsoonal floods to such a degree as to exert little influence on pH and hence pCO2. Therefore, the monsoonal overestimation of pCO2 at sites with low pH and alkalinity warrants further investigation of other factors than non-carbonate alkalinity to explain the increased sensitivity of pCO2 to subtle changes in acidity and buffering. These results illustrate the importance of direct measurements of pCO2 in highly polluted rivers, especially during the monsoon season. For river systems lacking pCO2 measurements, we suggest that carbonate equilibria-based models be complemented with corrective measures: 1) presenting pCO2 values calculated from low pH values (pH < 6.5 for monsoon and pH < 6.3 for dry season) together with the pH range to warn potential overestimation; 2) using pre-established regressions between measured pCO2 and environmental variables to correct pCO2 values, particularly during wet periods when large changes in pH and acid buffering are expected.