Figure 6. The time that air masses spent over water, ice and land derived using back trajectory analysis. The colour represents the temperature at which an INP concentration of 0.1 INP L-1 was reached. The black dots correspond to individual trajectories. The same filters as in Figure 5 were applied.
Overall, the evidence indicates that there is a strong source of biogenic INPs in the Barents, Kara and Laptev Seas off the Russian coast that can be sporadically transported to the central Arctic Ocean. There was high wind along the trajectories off the Russian coast which would be consistent with both the production of INPs in sea spray from these organic-rich seas or dust combined with terrestrial biogenic material from the various islands in this region. We also note that a recent study found that very active INP were produced in the Chukchi sea (near Alaska) under high wave conditions [Inoue et al. , 2021]. Furthermore, we note that these potential sources are sensitive to a changing climate, with river discharge, permafrost melt and organic matter input into the ocean from the major Russian rivers increasing in a warmer world [Jahn et al. , 2020]. In addition, the removal of ice from the Arctic could also expose marine and terrestrial sources of ice-nucleating aerosol around the Arctic Ocean, where they can then be aerosolised by the action of wind [Schmale et al. , 2021].
3.5 Implications for ice production in boundary layer central Arctic mixed-phase clouds
In this section, we assess whether the measured INP concentrations are high enough to initiate a transition from liquid-dominated clouds to ice-dominated clouds. Model simulations indicate that on the order of one ice crystal per litre of air is required to remove the bulk of liquid water from an Arctic cloud, whereas lower concentrations still reduce the liquid water path [Stevens et al. , 2018; Vergara-Temprado et al. , 2018b]. Hence, we use our INP measurements, both at ship level in the surface mixed layer and those from the SHARK in the cloud mixed layer, to estimate the concentration of INPs that become active at the ambient temperature of the atmosphere ([INP]ambient).
The quantity [INP]ambient combines the atmospheric temperature profiles from radiosonde measurements with the corresponding INP spectra as a rough indicator of primary ice crystal production. This is a crude analysis and a full cloud model would be required to represent ice crystal formation and sedimentation as well as INP recycling and latent heat release, to predict ice crystal concentration given an INP spectrum, but it does give an indication of what the measured INP spectra might mean for primary ice production in clouds. The highest [INP]ambient will be at the cold points, i.e. the top of the surface mixed layer and the top of the boundary layer (top of the cloud mixed layer), hence this is where we focus this analysis. We assume that ship level measurements are representative of the [INP]ambient at the top of the surface mixed layer and those from SHARK are representative of the top of the boundary layer, the assumption being that these respective layers are individually well-mixed.
The temperature minima within the main boundary layer were determined from radiosonde profiles, which were made every 6 hours throughout the entire cruise [VĂ¼llers et al. , 2021]. The calculated [INP]ambient for the duration of the cruise are shown in Figure 7, where the minimum temperature at the top of the cloud mixed layer (main boundary layer) and the top of the surface mixed layer is indicated for each filter period. In many cases, the temperature of the atmosphere was higher than the highest INP measurement, hence we are only able to report upper limits to [INP]ambient in these cases.