Development and application of a novel UHPLC-HRMS method for the analysis of organic wildfire tracers in ice cores.

Wildfires have an important role in affecting the Earth’s radiative balance. Biomass burning aerosols can scatter or absorb the incoming solar radiation, alter the ice and snow albedo and act as cloud condensation nuclei. Overall, their net contribution to the Earth’s radiative forcing is negative, however this estimate has large uncertainties. To better assess the impact of wildfires on climate (and vice versa), it is crucial to reconstruct their past regional and temporal variability on decadal and centennial timescales. Ice cores are excellent archives to perform such palaeofire reconstructions. Previous studies have reconstructed the occurrence of wildfires in ice cores using both inorganic (ammonium, potassium and black carbon) and organic proxies (levoglucosan, vanillic acid and p-hydroxybenzoic acid). However, a more comprehensive view that involved a broader suite of wildfire proxies was missing. Here, we present a new SPE-UHPLC-HRMS method for the determination of five organic biomass burning tracers (syringic acid, vanillic acid, vanillin, syringaldehyde and p-hydroxybenzoic acid) and pinic acid, as biogenic emission proxy, in ice core samples. This method showed average recoveries of 76% (58-88% range), excellent inter-day reproducibility, no significant matrix effects and fast analysis time (13 min per sample). Comparing the published concentration ranges of the selected species from different ice core regions (i.e. Alps, Greenland, Kamchatka, China and Svalbard Archipelago) with the procedural detection limits of this new methodology, we conclude that four of the six targeted compounds can be successfully detected in real ice and snow samples. Only for vanillin and syringaldehyde, no ice-core measurements have been reported in the scientific literature so far. The method development also involved the evaluation of common laboratory practices such as the melting and refreezing of ice samples before the analysis. We found that the melting and refreezing of the samples resulted in a mass loss for the majority of the investigated compounds, which was more evident at lower concentrations. We hypothesize that the reason of this phenomenon is the adsorption of the compounds on the walls of the glass vials used for this study. In light of this, we propose alternative sample storage strategies that can also be extended for the analysis of other compounds.

The method was successfully tested on nine ice core samples from the Colle Gnifetti (European Alps) and it will be applied on ice cores from the Alps and the Russian Altai, contributing to the better understanding of wildfire temporal evolution and their relations with climate.