This dataset contains carbon dioxide dry air mole fractions measured during the year-long MOSAiC expedition from October 2019 to September 2020. The measurements were performed in the Swiss container on the D-deck of Research Vessel Polarstern. Data were collected by cavity ring-down spectroscopy using a commercial Picarro instrument (model G2401). The minute-averaged dry air mole fractions were adjusted after cross-evaluation against discrete whole air samples collected for post-cruise analysis at the National Oceanic and Atmospheric Administration (NOAA) Global Monitoring Laboratory. Pollution spikes due to local anthropogenic pollution sources (e.g., exhaust by the vessel's engine and vents, skidoos, helicopters, on-ice diesel generators) were identified and flagged as follows. For each data point, the gradient (time derivative) was calculated (Beck et al., 2022). Data points corresponding to an abnormally high gradient (> 1.5 times the interquartile range) and neighboring points were discarded. The function “despike” from R package oce (version 1.3-0) was then applied to the time-series to remove any remaining local pollution spikes. Briefly, this function first linearly interpolates across any gaps (missing values). Then, it calculates a running median spanning k elements. The result of these two steps is the “reference” time-series. The standard deviation of the difference between values and the reference is then calculated. Values that differ from the reference by more than n times this standard deviation are considered to be spikes and eliminated. The function was applied once with n = 1 and k = 61 (~ 1 hour). The data columns include the Date and Time in Coordinated Universal Time (UTC), the latitude and longitude of Research Vessel Polarstern, the MOSAiC event label, the original carbon dioxide dry air mole fraction in µmol/mol, the adjusted carbon dioxide dry air mole fraction in µmol/mol after cross-evaluation, and a pollution flag where 'yes' means that local pollution was detected.

We used 1 sec time resolution positional data from the following datasets: PS122/1: Rex, M (2020), PS122/2: Haas, C (2020), PS122/3: Kanzow, T (2020), PS122/4: Rex, M (2021a), PS122/5: Rex, M (2021b)