Composition and sources of carbonaceous aerosol in the European Arctic at Zeppelin Observatory, Svalbard

Abstract. Our current understanding of Arctic carbonaceous aerosol (CA) is rudimentary and there is a lack of long-term observations for many components, such as organic aerosol (OA), exceptions to this include equivalent black carbon (eBC) and methane sulfonic acid (MSA). 

To address this, we analyzed long-term measurements of organic carbon (OC), elemental carbon (EC), and source-specific organic tracers from 2017 to 2020 to constrain CA sources in the rapidly changing Arctic. We also used absorption photometer (aethalometer) measurements to constrain equivalent BC from biomass burning (eBC_BB) and fossil fuel combustion (eBC_FF) using Positive Matrix Factorization (PMF). 

Our analysis showed that organic tracers are essential to understand Arctic CA sources. For 2017 to 2020, levoglucosan had a bimodal seasonality, with a signal from residential wood combustion (RWC) in the heating season (H-season; November to May) and from wildfires (WF) in the non-heating season (NH-season; June to October), demonstrating a pronounced inter-annual variability in the WF influence. Biogenic secondary organic aerosol (BSOA) species (2-methyltetrols) from isoprene oxidation appeared only in the NH-season, peaking in July to August. Intrusions of warm air masses from Siberia in summer caused three- and ninefold increases in 2-methyltetrols compared to 2017 to 2018, in 2019 and 2020, respectively, warranting investigation of the local vs. the long-range atmospheric transport (LRT) contribution, as certain Arctic vegetation has highly temperature sensitive biogenic volatile organic compounds (BVOC) emission rates. Primary biological aerosol particles (PBAP) tracers (various sugars and sugar-alcohols) were elevated in the NH-season but evolved differently, whereas cellulose was completely decoupled from the other PBAP tracers. Peak levels of most PBAP tracers and of 2-methyltetrols were associated with WF emissions, demonstrating the importance of measuring a broad spectrum of source specific tracers to understand sources and dynamics of CA. Finally, CA seasonality is heavily influenced by long-range atmospheric transport (LRT) episodes, since background levels are extremely low. E.g., we find the OA peak in the NH-season is as strongly influenced by LRT as is EC during Arctic Haze (AH). 

Source apportionment of CA by Latin Hypercube Sampling (LHS) showed a mixed contribution from RWC (46 %), fossil fuel (FF) sources (27 %), and BSOA (25 %) in the H-season, whereas BSOA (56 %) prevailed over WF (26 %) and FF (15 %) in the NH-season. Source apportionment of eBC by PMF showed that FF combustion dominated eBC (70 ± 2.7 %), whereas RWC (22 ± 2.7 %) was more abundant than WF (8.0 ± 2.9 %). Modeled BC concentrations from FLEXPART attributed an almost equal share to FF (51 ± 3.1 %) and BB. Both FLEXPART and the PMF analysis concluded that RWC is a more important source than WF. However, with a modeled RWC of 30 ± 4.1 % and WF of 19 ± 2.8 %, FLEXPART suggests relatively higher contributions to eBC from these sources.

We find that OA (281 ± 106 ng m⁻³) is a significant fraction of the Arctic PM₁₀ aerosol particle mass, though less than sea salt aerosol (SSA) (682 ± 46.9 ng m⁻³) and mineral dust (MD) (613 ± 368 ng m⁻³) as well as typically non-sea-salt sulfate (nssSO₄²⁻) (314 ± 62.6 ng m⁻³), originating mainly from anthropogenic sources in winter and from natural sources in summer. FF combustion was the prevailing source of eBC, whereas RWC made a larger contribution to eBC_BB than WF.