Variations in the summer oceanic <i>p</i>CO<sub>2</sub> and carbon sink in Prydz Bay using the self-organizing map analysis approach

Abstract. This study applies a neural network technique to produce maps of oceanic surface pCO2 in Prydz Bay in the Southern Ocean on a weekly 0.1∘ longitude × 0.1∘ latitude grid based on in situ measurements obtained during the 31st CHINARE cruise from February to early March 2015. This study area was divided into three regions, namely, the “open-ocean” region, “sea-ice” region and “shelf” region. The distribution of oceanic pCO2 was mainly affected by physical processes in the open-ocean region, where mixing and upwelling were the main controls. In the sea-ice region, oceanic pCO2 changed sharply due to the strong change in seasonal ice. In the shelf region, biological factors were the main control. The weekly oceanic pCO2 was estimated using a self-organizing map (SOM) with four proxy parameters (sea surface temperature, chlorophyll a concentration, mixed Layer Depth and sea surface salinity) to overcome the complex relationship between the biogeochemical and physical conditions in the Prydz Bay region. The reconstructed oceanic pCO2 data coincide well with the in situ pCO2 data from SOCAT, with a root mean square error of 22.14 µatm. Prydz Bay was mainly a strong CO2 sink in February 2015, with a monthly averaged uptake of 23.57±6.36 TgC. The oceanic CO2 sink is pronounced in the shelf region due to its low oceanic pCO2 values and peak biological production.