Two decades of flask observations of atmospheric (O2/N2), CO2, and APO at stations Lutjewad (the Netherlands) and Mace Head (Ireland), and 3 years from Halley station (Antarctica)

We present 20-year flask sample records of atmospheric CO2, (O2/N2), and atmospheric potential oxygen (APO) from the stations Lutjewad (the Netherlands) and Mace Head (Ireland), and a 3-year record from Halley station (Antarctica). We include details of our calibration procedures and the stability of our calibration scale over time, which we estimate to be 3 per meg over the 11 years of calibration, and our compatibility with the international Scripps O2 scale. The measurement records from Lutjewad and Mace Head show similar long-Term trends during the period 2002-2018 of 2.31g0.07g ppmg yr-1 for CO2 and-21.2g0.8 per megg yr-1 for (O2/N2) at Lutjewad, and 2.22g0.04g ppmg yr-1 for CO2 and-21.3g0.9 per megg yr-1 for (O2/N2) at Mace Head. They also show a similar (O2/N2) seasonal cycle with an amplitude of 54g4 per meg at Lutjewad and 61g5 per meg at Mace Head, while the CO2 seasonal amplitude at Lutjewad (16.8g0.5g ppm) is slightly higher than that at Mace Head (14.8g0.3g ppm). We show that the observed long-Term trends and seasonal cycles are in good agreement with the measurements from various other stations, especially the measurements from the Weybourne Atmospheric Observatory (United Kingdom). However, there are remarkable differences in the progression of annual trends between the Mace Head and Lutjewad records for (O2/N2) and APO, which might in part be caused by sampling differences, but also by environmental effects, such as North Atlantic Ocean oxygen ventilation changes to which Mace Head is more sensitive. The Halley record shows clear trends and seasonality in (O2/N2) and APO, the latter agreeing especially well with continuous measurements at the same location made by the University of East Anglia (UEA), while CO2 and (O2/N2) present slight disagreements, most likely caused by small leakages during sampling. From our 2002-2018 records, we find a good agreement with Global Carbon Budget 2021 (Friedlingstein et al. (2021) for the global ocean carbon sink: 2.1g0.8g yr-1, based on the Lutjewad record. The data presented in this work are available at 10.18160/qq7d-T060 (Nguyen et al., 2021).