Figure 2. Surface mixed layer INP concentrations throughout the campaign. a) The number of INPs per litre of air sampled was calculated using data from offline droplet freezing experiments, conducted within hours of the samples being taken. The spectra shown in blues represent samples that were heated to close to 100 °C for 30 min. Background values were subtracted from the data. Sampling times varied from 6 h to 3 days and were taken using a heated whole air inlet on the 4th deck (25 m above mean sea level) of the Oden Icebreaker. Temperature uncertainties (not shown) for the droplet freezing experiments were estimated to be ±0.4 °C. The format of the key is YYMMDD_hhmm; these correspond to the start time of the filter sample in the 24-hour time format; the start and end times are in Table S1. b) The data from this study are presented alongside literature data for ground, ship and aircraft-based campaigns around the Arctic [Bigg, 1996; Bigg and Leck, 2001; Borys, 1989; Conenet al. , 2016; Creamean et al. , 2019; Creamean et al. , 2018; DeMott et al. , 2016; Flyger and Heidam, 1978; Hartmann et al. , 2020b; Hartmann et al. , 2021; Irishet al. , 2019; Mason et al. , 2016; Porter et al. , 2020; Prenni et al. , 2007; Rogers et al. , 2001; Sanchez-Marroquin et al. , 2020; Si et al. , 2019; Wexet al. , 2019] and a compilation derived from precipitation samples [Petters and Wright, 2015].
It is is remarkable to note that the highest concentrations measured here at the North Pole are as high as the highest INP concentration reported in INP rich environments such as the mid-latitude terrestrial environment [O’Sullivan et al. , 2018; Petters and Wright, 2015], despite having much lower aerosol concentrations. Overall, our INP measurements indicate that the INP concentration spectra within the high Arctic surface mixed boundary layer can be extremely variable, perhaps far more variable than anywhere else on Earth.
In order to test for the presence of proteinaceous biological ice-nucleating material, we heated the most active sample suspensions to close to 100 °C [Daily et al. , 2021]. The activity of these samples was always reduced, with all of the activity above −20 °C being removed (Figure 2a and Figure S1). Hence, it appears that the most active INPs sampled close to the North Pole were most likely of biological origin. Atmospheric INP at lower latitudes were also found to be heat sensitive [Hartmann et al. , 2021]. Together, this indicates that proteinaceous biological INP are important in the Artic.
The time series in Figure 3 shows the temperature at which a concentration of 0.1 INP L-1 was measured (T [INP]=0.1), and highlights the variability of INP concentrations at the North Pole throughout August and September of 2018. The first peak in ice-nucleating activity was observed during a period in which the ship was breaking ice prior to being moored to an ice-floe (i.e. prior to 16th August). It is reasonable to question whether the very high INP concentrations observed during the ice-breaking period resulted from the ice-breaking itself. Ice-breaking involved frequent backward and forward motions, hence there is the potential for sampling ship emissions (such as ship stack emissions, detailed in the methodology) and aerosol resulting from ice-breaking and the disruption of the sea surface. We manually stopped sampling if there were activities planned that would affect sampling (such as helicopter flights) in addition to using a pollution control system that stopped the flow through the filters when aerosol concentrations increased rapidly in a manner associated with sampling a ship plume, and also when the wind was not from the correct direction (from forward of the ship’s superstructure). Despite the precautions taken to eliminate these sources of contamination, we cannot completely exclude the possibility of contamination. However, there was a pause in ice-breaking when a clean air station was established (10th August, we sampled for 6 hours) that coincided with high INP concentrations. At this clean air station the ship was moored facing into the wind, with ship-based aerosol sources aft of the aerosol inlets. Hence, we were confident that sampling of ship pollution and ice-breaking aerosol were eliminated, increasing confidence that these high values in this period were indeed representative of the central Arctic Ocean. In addition, there was also a period of very high ice-nucleating activity a few days after the ice-floe station had been established and the ship was pointing into the wind, demonstrating that there were very active INPs that were not related to ice-breaking or ship emissions. The time series also highlights two distinct periods. The period up to the 23rd August was characterised by variable but often very high INP concentrations, whereas the period after this was characterised by much lower INP concentrations. We examine the back trajectories associated with these different periods later in the paper.