Figure 5. Backward trajectories over 7 days, starting at the ship location, for the INP samples taken throughout the campaign. Trajectories were launched every hour during the sampling period, and each point represents an hour in time along the back trajectory. The starting height for the trajectories was 32 m above mean sea level. Any points along the trajectories which were above the model boundary layer were removed, and any points along the trajectory that preceded precipitation events (>0.1 mm h-1) were removed. Hence, any potential sources of INP in the boundary layer are neglected if they occurred prior to a precipitation event (we assume precipitation removes INPs) a) The colour of the trajectories represents the temperature at which 0.1 INP L-1 was measured for that sampling period. b) The colour of the trajectories represents the wind speed for each point along the trajectory. The sea ice extent is from the NASA National Snow and Ice Data Center [Maslanik and Stroeve, 1999].
It is striking that the trajectories with the lowest INP concentrations spent most of the preceding seven days over the pack ice and to some extent over the MIZ. These results indicate that, during this campaign, open leads, sea ice and the MIZ were weak sources of INPs, in conflict with previous suggestions [Bigg and Leck, 2001; Hartmann et al. , 2020a].
The highest ice-nucleating activities from sampled aerosol originating along the Russian coast were also correlated with high wind speeds along the trajectories, i.e., in the region of the aerosol origin (Figure 5b). This, together with the heat tests and INP size information, point to a wind-driven marine biological source of INPs associated with organic-rich film droplet sea spray aerosol. There were trajectories with high wind speeds over the North American continent, the pack ice and the coast of Greenland, but the ice-nucleating activity for these was not greatly enhanced. Hence, our results are consistent with a strong source of highly active INPs in the coastal marine waters of northern Russia which were aerosolised during windy conditions. Marine waters elsewhere in the world are thought to produce aerosol with relatively low ice nucleating activities [Vergara-Temprado et al. , 2017] (see Figure 4); however, our results suggest that the shallow seas off the Russian coast might be relatively strong sources of highly active INPs. Composition analysis of the aerosol during the peak ice nucleation activity on the 11th August and the 19th August is consistent with a marine source for many of these aerosol particles (samples were rich in Na, Cl and sulfate; see Table S3). Hence, the question is why the aerosol from near the coast of Russia is so much more active than aerosol derived from other marine locations such as the North Atlantic and the Southern Ocean [McCluskey et al. , 2018a; McCluskey et al. , 2018b].
A major difference to the North Atlantic and Southern Ocean is that the shallow seas off the coast of northern Russia are strongly influenced by riverine input from Russia that is rich in organic material, silt and nutrients [Ahmed et al. , 2020; Juhls et al. , 2019]. In fact, much of the dissolved organic matter in the Arctic Ocean is derived from river input [Juhls et al. , 2019] and the discharge of these rivers (and the amount of dissolved organic carbon flowing in the sea via rivers) is increasing [Ahmed et al. , 2020; Juhls et al. , 2020]. It has been shown that melting permafrost, which is known to enter river water [Juhls et al. , 2020], harbors copious quantities of warm temperature INPs [Creameanet al. , 2020]. Hence, it is possible that the highest INP concentrations we detected at the North Pole were derived from marine waters rich in terrestrially derived ancient biological INPs. Alternatively, ocean biology fertilised by the nutrient rich waters on the continental shelf may produce more INPs than are present in remote marine locations. A measurement campaign to quantify the INP content of the waters off the coast of Russia is clearly required.
Some of the back trajectories that had the highest INP concentrations passed over islands in the Barents and Kara Seas, including Svalbard, Franz Josef Land, Novaya Zemlya and Severnaya Zemlya. Many of these locations have been identified as poorly defined dust sources [Bullardet al. , 2016] and dust from Svalbard has been shown to contain biological ice-nucleating materials [Tobo et al. , 2019]. However, in a further analysis of the back trajectory data (Figure 6), we find that there was little to no correlation with time spent over land, whereas the ice-nucleating activity increased with the time the air parcels spent over open ocean. This implies that the sources of INP were associated with the marine environment. Having said this, we cannot rule out relatively small island point sources being important sources of INPs.