<p>Air inclusions trapped in polar ice provide unique records of the past atmospheric composition ranging from key greenhouse gases to short-lived trace gases like ethane and propane. Provided the analyzed species concentrations and their isotopic fingerprints accurately reflect past atmospheric composition, valuable constraints can be put onto biogeochemical cycles. However, it is already known that not all drill sites or specific time intervals are equally suitable to derive artefact-free gas records; e.g., CO<sub>2</sub> data from Greenland ice is overprinted by CO<sub>2</sub> &#8216;in situ&#8217; production due to impurities in the ice, and only the cleaner Antarctic ice allows to reconstruct past atmospheric CO<sub>2</sub>.</p><p>Until recently, CH<sub>4</sub> artefacts in polar ice were only detected on melt affected samples or for short spikes related to exceptional impurity deposition events (Rhodes et al 2013). However, careful comparison of CH<sub>4</sub> records obtained using different extraction methods revealed disagreements among Greenland CH<sub>4</sub> records and initiated targeted experiments.</p><p>Here, we report experimental findings of CH<sub>4</sub> artefacts occurring in dust-rich sections of Greenland ice cores. The artefact production happens during the melt extraction step (&#8216;in extractu&#8217;) of the classic wet extraction technique and typically reaches 20 ppb in dusty stadial ice which causes erroneous reconstructions of the interhemispheric CH<sub>4</sub> difference and strongly affects the hydrogen isotopic signature of CH<sub>4</sub> (Lee et al. 2020). The measured CH4 excess is proportional to the amount of mineral dust in the ice. Knowing the empirical relation between produced CH4 and the dust concentration of a sample allows a first-order correction of existing CH4 data sets and to revise previous interpretations.</p><p>To shed light on the underlying mechanism, we analyzed samples for other short-chain alkanes ethane (C<sub>2</sub>H<sub>6</sub>) and propane (C<sub>3</sub>H<sub>8</sub>). The production of CH<sub>4</sub> was always tightly accompanied with C<sub>2</sub>H<sub>6</sub> and C<sub>3</sub>H<sub>8</sub> production at amounts exceeding the past atmospheric background levels derived from low-dust samples. Independent of the produced amounts, CH<sub>4</sub>, C<sub>2</sub>H<sub>6</sub>, and C<sub>3</sub>H<sub>8</sub> were produced in molar ratios of roughly 16:2:1, respectively. The simultaneous production at these ratios does not point to an anaerobic methanogenic origin which typically exhibits methane-to-ethane ratios of >>100. Such alkane patterns are indicative of abiotic degradation of organic matter as found in sediments.</p><p>We found this specific alkane pattern not only for dust-rich samples but also for samples that were affected by surface melting from the last interglacial (NEEM ice core) with low dust concentrations. This implies that the necessary precursor is an impurity also present in low-dust ice and the step leading to the production of the alkanes could then be activated when a sufficient boundary condition is met for the production, e.g. by melt/refreeze of surface snow.</p>
Atmospheric CO2, CH4 and N2O records over the past 60 000 years based on the comparison of different polar ice cores
