| Summary: | Aquatic CO<sub>2</sub> emission is typically estimated (i.e., not measured) through a gas exchange balance. Several factors can affect the estimation, primarily flow velocity and wind speed, which can influence a key parameter, the gas exchange coefficient <i>K<sub>T</sub></i> in the balancing approach. However, our knowledge of the uncertainty of predictions using these factors is rather limited. In this study, we conducted a numeric assessment on the impact of river flow velocity and wind speed on <i>K<sub>T</sub></i> and the consequent CO<sub>2</sub> emission rate. As a case study, we utilized 3-year (2019–2021) measurements on the partial pressure of dissolved carbon dioxide <i>(p</i>CO<sub>2</sub>) in one of the world’s largest alluvial rivers, the lower Mississippi River, to determine the difference in CO<sub>2</sub> emission rate estimated through three approaches: velocity-based <i>K<sub>T</sub></i>, wind-based <i>K<sub>T</sub></i>, and a constant <i>K<sub>T</sub></i> (i.e., <i>K<sub>T</sub></i> = 4.3 m/day) that has been used for large rivers. Over the 3-year study period, river flow velocity varied from 0.75 ms<sup>−1</sup> to 1.8 ms<sup>−1</sup>, and wind speed above the water surface fluctuated from 0 ms<sup>−1</sup> to nearly 5 ms<sup>−1</sup>. Correspondingly, we obtained a velocity-based <i>K<sub>T</sub></i> value of 7.80–22.11 m/day and a wind-speed-based <i>K<sub>T</sub></i> of 0.77–8.40 m/day. Because of the wide variation in <i>K<sub>T</sub></i> values, the estimation of CO<sub>2</sub> emission using different approaches resulted in a substantially large difference. The velocity-based <i>K<sub>T</sub></i> method yielded an average CO<sub>2</sub> emission rate (FCO<sub>2</sub>) of 44.36 mmol m<sup>−2</sup> h<sup>−1</sup> for the lower Mississippi River over the 3-year study period, varying from 6.8 to 280 mmol m<sup>−2</sup> h<sup>−1</sup>. In contrast, the wind-based <i>K<sub>T</sub></i> method rendered an average FCO<sub>2</sub> of 10.05 mmol m<sup>−2</sup> h<sup>−1</sup> with a small range of fluctuation (1.32–53.40 mmol m<sup>−2</sup> h<sup>−1</sup>,), and the commonly used constant <i>K<sub>T</sub></i> method produced an average FCO<sub>2</sub> of 11.64 mmol m<sup>−2</sup> h<sup>−1</sup>, also in a small range of fluctuation (2.42–56.87 mmol m<sup>−2</sup> h<sup>−1</sup>). Based on the findings, we conclude that the effect of river channel geometry and flow velocity on CO<sub>2</sub> outgassing is still largely underestimated, and the current estimation of global river CO<sub>2</sub> emission may bear large uncertainty due to limited spatial coverage of flow conditions and the associated gas exchange variation.
|
|---|