Inconsistency between ship- and Argo float-based pCO2 at the intense upwelling region of the Drake Passage, Southern Ocean

Abstract

The Southern Ocean absorbs a quarter of anthropogenic carbon dioxide (CO2) from the atmosphere to modulate the climate system. However, less attention has been paid to the CO2 outgassing phenomenon at the Antarctic Circumpolar Current (ACC) region of the Southern Ocean due to strong upwelling. Recent studies using autonomous biogeochemical-Argo float revealed a greater winter CO2 outgassing than previously estimated at ACC zone of the Southern Ocean, which, however, remains controversial and urgently needs to be validated. Here we take the Drake Passage as a case study to present new insights into the Southern Ocean carbon cycle and examine the validity of float-based CO2 outgassing. Upon integrating the ship-based data over the past two decades, we investigate the spatiotemporal variability of sea surface CO2 partial pressure (pCO2) in Drake Passage. We show that Drake Passage is acting as a year-round weak CO2 sink, although some CO2 uptake is counteracted by winter CO2 outgassing. The float-based pCO2 values are overall higher than ship-based values in winter, by 6 to 20 µatm (averaged 14 µatm) at the most intensive upwelling region. We then develop a surface carbon balance calculation (considering mixing between surface, subsurface, and upwelled waters) to estimate the potential of surface pCO2 increase due to upwelling, and we find that upwelling of CO2-rich subsurface waters in Drake Passage cannot support an excess ΔpCO2 of 14 µatm as suggested by float detections. We further compare our results to previous study and find that, although we used same datasets and obtained comparable results, the way to conclude the bias in float-based pCO2 would cause significant difference: an uncertainty of ±2.7% (i.e., ± 11 µatm) in float-based pCO2 estimated by other study seems acceptable, however, it is five times larger than the typical ship-based pCO2 uncertainty ( ± 2 µatm), and would cause ~180% bias in CO2 flux estimates. Going forward, there is special need for caution when interpreting the float-based CO2 flux; meanwhile, further comparisons and corrections between float- and ship-based pCO2 are clearly warranted.