Terrestrial or marine? – Indications towards the origin of Ice Nucleating Particles during melt season in the European Arctic up to 83.7°N

Ice nucleating particles (INPs) initiate the primary ice formation in clouds at temperatures above ca. -38°C and have an impact on precipitation formation, cloud optical properties and cloud persistence. Despite their roles in both weather and climate, INPs are not well characterized, especially in remote regions such as the Arctic. We present results from a ship-based campaign to the European Arctic in May to July 2017. We deployed a filter sampler and a continuous flow diffusion chamber for offand online INP analysis, respectively. We also investigated the ice nucleation properties of samples from different 5 environmental compartments, i.e., the sea surface microlayer (SML), the bulk seawater (BSW), and fog water. Concentrations of INP (N INP) in the air vary between two to three orders of magnitudes at any particular temperature and are, except for the temperatures above -10°C and below -32°C, lower than in mid-latitudes. In these temperature ranges INP concentrations are the same or even higher than in the mid-latitudes. Heating of the filter samples to 95°C for 1 hour we found a significant reduction in ice nucleation activity, i.e., indications that the INPs active at warmer temperatures are biogenic. At colder temperatures the 10 INP population was likely dominated by mineral dust. The SML was found to be enriched in INP compared to the BSW in almost all samples. The enrichment factor (EF) varied mostly between 1 and 10, but EFs as high as 94.97 were also observed. Filtration of the seawater samples with 0.2μm syringe filters lead to a significant reduction in ice activity, indicating the INPs are larger, and/or are associated with particles larger than 0.2μm. A closure study showed that aerosolization of SML and/or seawater alone cannot explain the observed air-borne N INP unless significant enrichment of INP by a factor of 105 takes place 15 during the transfer from the ocean surface to the atmosphere. In the fog water samples with -3.47°C we observed the highest freezing onset of any sample. A closure study connecting N INP in fog water and the ambient N INP derived from the filter samples shows good agreement of the concentrations in both compartments, which indicates that INPs in the air are likely all activated into fog droplets during fog events. In a case study we considered a situation during which the ship was located in the marginal sea ice zone and N INP in air and the SML were highest in the temperature range above -10°C. Chlorophyll-a 20

microphysically unstable nature. These clouds show a lower degree of glaciation in comparison to clouds at similar altitudes in other parts of the globe (Costa et al., 2017) which might be due to a lack of ice nucleating particles (INPs). INPs are the 55 catalyst needed for the primary ice formation at temperatures relevant for mixed-phase clouds and thus essential to induce the freezing of supercooled liquid cloud droplets.
As INPs can directly affect the phase state of the cloud, their abundance and efficiency to initiate freezing also affects precipitation, life time and the radiative effects of clouds (e.g. Loewe et al., 2017;Prenni et al., 2007;Ovchinnikov et al., 2014;Solomon et al., 2015). Solomon et al. (2018)  Several previous studies have reported that marine as well as terrestrial sources contribute to Arctic INPs ice active at temperatures above approximately -15°C. For the marine environment it was found that especially the sea surface microlayer (SML) can be highly ice active (Alpert et al., 2011a, b;Bigg, 1996;Bigg and Leck, 2008;Irish et al., 2017Irish et al., , 2019bKnopf et al., 2011;Leck and Bigg, 2005;Schnell and Vali, 1976;Wilson et al., 2015;Zeppenfeld et al., 2019). Especially marine microorganism such as bacteria and algae as well as their exudates are thought to be the source for the INPs. Connections to biological driven processes like plankton blooms have been made (Creamean et al., 2019). Another recent publication by Kirpes et al. (2019) found locally produced open leads to be the dominant aerosol source in winter. The emitted sea spray aerosol particles were found to possess organic coatings, consisting of marine saccharides, amino acids, fatty acids, and di-70 valent cations. These substances are known from exopolymeric secretions produced by sea ice algae and bacteria, which, as mentioned before are thought to be responsible for the ice activity in seawater. Studies on INPs at coastal sites tend to find influences from marine and terrestrial sources, often with a contribution of biological INPs and seasonal changes Šantl-Temkiv et al., 2019;Wex et al., 2019). For terrestrial sources, mineral dust itself is known to be relevant for lower temperatures (Sanchez-Marroquin et al., 2020), but Tobo et al. (2019) showed for glacial outwash material that dust can 75 be the carrier for biological material, which is more ice active than the dust alone. Highly ice active biological INPs have also been found Arctic ice cores from up to 500 years ago . Also millennia old permafrost soil was found to contain biological INPs that can be mobilized into the atmosphere, lakes, rivers, and the ocean when the Permafrost thaws (Creamean et al., 2020). This highlights the importance of biological INPs especially in the changing Arctic environment.
Despite past significant efforts and increase in knowledge, we still lack quantitative insights concerning the abundance, the 80 properties, and sources of Arctic INPs. Especially concerning the latter, the relative importance of marine vs. terrestrial sources is still debated. Therefore, open questions addressed in this paper are -What is the abundance of Arctic INPs and in what temperature range can they nucleate ice?
-What is the nature of Arctic INPs (biogenic material vs. mineral dust)?

95
The first leg (PS106.1) started on 24 May in Bremerhaven (Germany) and ended 21 June in Longyearbyen (Svalbard) and featured a 10 day ice floe camp that was set up between 5 and 14 June 2017. The main area of investigation was the Arctic Ocean a few hundred kilometers northwest of Svalbard (see Fig. 1).
The expedition continued with its second leg (PS106.2) on 23 June from Lonyearbyen and ended 20 July in Tromsø (Norway).
In comparison to PS106.1, the second leg focused on the area northeast of Svalbard, went up to higher latitudes (up to 83.7°N) 100 and the vessel did not stop for extended stays at an ice floe.
As an overview about the meteorological situation during the campaign, Fig. S1 in the SI shows the frequency distributions for all meteorological parameters that were continuously measured on Polarstern. The mean and standard deviation of air temperature (T air ), relative humidity (RH) and atmospheric pressure (p) are given in the following: for the whole first leg are 105 T air = -0.01°C± 4.21°C, RH = 90.70 % ± 10.62 % and p = 1016.36 hPa ± 7.48 hPa, whereas the second leg the parameters were T air = 0.22°C± 2.71°C, RH = 94.82 % ± 6.09 % and p = 1006.84 hPa ± 5.12 hPa. During the time within the ice pack the averages of these parameters were as follows T air = -1,37°C± 1.50°C, RH = 94.35 % ± 4.54 % and p = 1011.27 hPa ± 8.52 hPa and out of the ice pack: T air = 4.75°C± 3.65°C, RH = 88.03 % ± 11.27 % and p = 1012.92 hPa ± 4.89 hPa.
For further details on the measurement strategy as well as the meteorological, sea ice, and cloud conditions during PASCAL 110 refer to Wendisch et al. (2019) and the PS106 cruise report by Macke and Flores (2018).

Sample collection
In order to gain a comprehensive insight into the abundance and nature of INPs in the Arctic during summer time, samples from different compartments were taken. These included atmospheric, bulk seawater (BSW), sea surface microlayer (SML), and fog samples. All samples were stored on the vessel directly after sampling in a cold room at -20°C and it was ensured that 115 the samples stayed below 0°C during the transport to the Leibniz Institute for Tropospheric Research (TROPOS), were they were stored at -24°C until they were analyzed.

Filter sampling
Aerosol particles were sampled using a low volume filter sampler (LVS; DPA14 SEQ LVS, DIGITEL Elektronik AG, Volketswil, Switzerland) with a PM 10 inlet (DPM10/2.3/01, DIGITEL Elektronik AG, Volketswil, Switzerland). The sampler was located 120 on top of a measurement container placed on the starboard side of the monkey island (ca. 30 m above sea level). It was operated with an average volumetric flow of 27.9 L min -1 . It should be noted that our flow rate is lower than the standardized flow rate for PM 10 inlets, hence our cut-off diameter is higher than 10 µm (ca. 11.7 µm). The LVS was routinely operated with an 8 hours sampling period, which results in a total sampled air volume of 13.4 m 3 per filter sample. On four days the 8 hour cycle was replaced by a 2 hour cycle to study possible diurnal variation. The filter sampler features sealed storage cassettes and an 125 automated filter change that allows unsupervised sampling for multiple days. The samples were collected on polycarbonate pore filters (Nuclepore ® , Whatman™; 0.2 µm pore size, 47 mm diameter). Usually 12 filters were prepared and put in place inside the sampler. Two field blanks were taken on each leg and were used to define the lower limit of observable where the sample were taken near the ground (for details see Zeppenfeld et al., 2019). SML samples were collected with a glass plate sampler Van Pinxteren et al., 2017). The glass plate is dipped into the water body, slowly withdrawn and the surface film, which clings to the sides of the glass plate, is wiped off the plate into a sample container with a Teflon ® wiper.
The seawater sampling was conducted on a daily basis. The SML and bulk seawater samples were taken at the same time and 140 location with the only exceptions of shallow meltponds where no samples from one meter depth could be taken as well as days with harsh weather when no surface film could form. 42 SML samples and 42 bulk seawater samples were collected during the campaign. A further description of the seawater sampling, and a chemical and microbiological analysis of the samples can be found in Zeppenfeld et al. (2019). A list of the seawater samples can be found in Tab. S2 in the SI.

Fog sampling 145
Fog was collected with the Caltech Active Strand Cloud Collector Version 2 (CASCC2; described in Demoz et al., 1996). The CASCC2 is a non-selective sampler that catches hydrometeors by impaction on Teflon®strands (508 µm diameter). Droplets caught on the strands are gravitationally channeled into a Nalgene bottle. The instrument operates with a flow rate of approximately 5.3 m 3 min -1 resulting in a 50% lower cut-off size of approximately 3.5 µm. During daytime on leg 1 the sampler turned on every time the visibility decreased significantly and was running continuously during the night. On leg 2 the sampler was 150 running continuously and the sample bottle was changed whenever a significant amount of sample material was collected and the fog event was over. In all cases the sampler was rinsed with ultrapure water after a fog event was sampled and the sample bottle changed. During the entire campaign, 22 samples were collected, about two thirds of them on the second leg alone. A list of all fog samples can be found in Tab. S3 in the SI.

Sample preparation
Samples stored at -24°C were thawed only to perform the measurements. The measurements were performed on the same day as the thawing, and the remaining sample material was refrozen at the end of the day on which the measurements were completed.
The polycarbonate filters are put in a centrifuge tube along with 3 mL of ultrapure water ::::: (Type :: 1; :::::::: Direct-Q ® :: 3 ::::: Water :::::::::: Purification 160 :::::: System, :::::: Merck ::::::::: Millipore, ::::::::: Darmstadt, ::::::::: Germany) and are shaken in an oscillating shaker for 15 minutes in order to extract the particles from the filter and bring them into suspension. Then 100 µL of that suspension are removed for the analysis with the Leipzig Ice Nucleation Array (LINA; described section 2.3.3). For the analysis with the Ice Nucleation Droplet Array (INDA; described in the section 2.3.4), the remaining 2.9 µL of the suspension are made up to a total of 6 mL with ultrapure water and shaken again as before. The reason for this procedure is to use as little water as viable, i.e., to dilute the sample as little as 165 possible.
Sea and fog water samples do not require any preparation and can be directly measured with either setup.

Test for heat-labile INPs
After the initial measurement arbitrarily selected samples were chosen to test for the presence of heat-labile INPs in the samples. The sample solution was sealed in an centrifuge tube and placed in an oven. The sample was heated at 95°C for 1 h 170 and subsequently analysed with the LINA device (described section 2.3.3).

Leipzig Ice Nucleation Array (LINA)
LINA is a Droplet Freezing Assay (DFA), the design of which is based on a DFA called BINARY by Budke and Koop (2015).
An array of 90 droplets with a typical volume of 1 µL of the sample suspension is placed onto a hydrophobic glass slide (40 mm diameter). Each droplet is within its individual compartment made from a perforated, anodized aluminum plate and covered 175 with another glass slide. In this way it can be ensured that droplets do not interact during the freezing process, e.g., via ice seeding by frost splintering or the Bergeron-Wegener-Findeisen process. Furthermore, droplet evaporation is minimized. At a cooling rate of 1°C min -1 the sample droplets are cooled by a 40 x 40 mm 2 Peltier element inside a freezing stage (LTS120, Linkam Scientific Instruments, Waterfield, UK). The freezing stage is coupled with a cryogenic water circulator (F25-HL, Julabo, Seelbach, Germany) in order to achieve temperatures below -25°C down to the temperature at which homogeneous 180 freezing occurs naturally, i.e. -38°C. A thin layer of squalene oil thermally connects the Peltier element and the glass slide with the droplets on top. The freezing stage itself consists of a gas tight aluminium housing, which is purged with dry, particlefree air during the measurement. A LED dome lighting (SDL-10-WT, MBJ-Imaging GmbH, Hamburg, Germany) is used for shadow-free illumination of the droplets. A charge-coupled device camera is mounted at the apex of the dome and takes images every 6 s which corresponds to a temperature resolution of 0.1°C if cooled with 1°C min -1 . An aperture below the dome blocks 185 the light partially and creates a ring-shaped reflection in each droplet. This is used as detectable feature that vanishes upon freezing of the droplet. A custom Python algorithm then evaluates each image in terms of the number of frozen droplets, N f , in each individual image. As every image corresponds to a certain temperature the frozen fraction at the respective temperature, f ice (T), can be easily derived. :::: LINA :::: was :::: used :: to ::::::: evaluate ::: all :::: filter ::::::: samples :: as :::: well :: as ::: all :::: SML :::: and ::::: BSW ::::::: samples. :

190
The basic design of the INDA device is inspired by Conen et al. (2012), but as suggested in Hill et al. (2016), Polymerase Chain Reaction (PCR) plates instead of individual tubes were used. In each of the 96 wells of the PCR plate 50 µL sample material is filled. Then the PCR plate is sealed with a transparent cover foil and immersed in the bath of a cryostat (FP45-HL, Julabo, Seelbach, Germany) in a way that the wells itself are surrounded by refrigerant (ethanol), but not so deep that the PCR plate would be completely submerged. The PCR plate is illuminated from below which makes the phase change of the sample 195 suspension visible as darkening of the respective well. The temperature of the refrigerant lowered with a rate of ca. 1°C min -1 , while simultaneously the temperature is recorded and a on-top mounted camera takes pictures at 0.1°C intervals. The images are then again evaluated with a custom Python algorithm for N f in order to derive f ice (T). ::::: INDA :::: was :::: used ::: for :::::::::::: measurements :: of :::: SML :::: and ::::: BSW ::::::: samples ::: as :::: well :: as ::: for ::: fog ::::: water. :

200
Cumulative number concentrations of INPs per volume of sample as a function of temperature were calculated for each experiment utilizing the equation given in Vali (1971): where N total is the number of droplets, and N frozen (T) the number of frozen droplets at temperature T . With the given number of droplets (N total = 90) and volume (V drop =1 µL), the upper and lower limits of the detectable range 205 of LINA are 1.12·10 4 and 4.5·10 6 L -1 (water), whereas 2.1·10 2 and 9.1·10 4 L -1 (water) are the limits for INDA (N total = 96; V drop = 50 µL). The temperature values of the seawater samples were corrected for freezing point depression due to the salt content as described in Koop and Zobrist (2009).
In case of the atmospheric filter samples in order to derive atmospheric N INP , the denominator in Eq.1 needs to be modified so that it represents the volume of air distributed into each droplet: where V air is the air volume sampled onto one filter and V wash is the volume of water the particles were rinsed off with and suspended in.

Collocated Measurements and Supporting Observations
In addition to the sampling of INPs, the physico-chemical properties of the prevailing atmospheric aerosol particles were 215 measured inside a temperature controlled measurement container located on the monkey island of the RV Polarstern. The temperature inside the container was held at ca. 24°C, while the aerosol inlet was heated to 30°C to prevent icing. The aerosol inlet consists of a 6 m long stainless steel tubing (inner diameter of 40 mm), which faces upwards at a 45°angle to the bow of the ship. The flow through the inlet was set to 40 L min -1 (Reynolds number < 2000). With an isokinetic splitter the aerosol was distributed between the different instruments. The aerosol instrumentation relevant to this study included: a mobility particle 220 size spectrometer (MPSS) to measure particle number size distributions (PNSD), a condensation particle counter (CPC) to measure total particle concentration (N tot ), and a cloud condensation nuclei counter (CCNC) to measure the concentrations of cloud condensation nuclei (N CCN ).
PNSDs in the size range between 10 and 800 nm were measured with a TROPOS-type MPSS (Wiedensohler et al., 2012).

225
The time resolution of an up-and down-scan was 5 min. PNSDs were derived with the inversion algorithm by Pfeifer et al. (2014) and corrected for transmission losses as well as counting efficiencies according to Wiedensohler et al. (1997). The sizing of the MPSS was calibrated according to Wiedensohler et al. (2018) at regular time intervals during the campaign (for further details on the MPSS and the measurement container refer to Kecorius et al., 2019). N tot was measured with a CPC (Model 3010, TSI Inc., Shoreview, USA; lower cutoff: 10 nm). A CCNC (CCN-100, DMT, Boulder USA Roberts and Nenes, 2005) 230 was used to measure N CCN at six different supersaturations (SS; 0.1%, 0.15%, 0.2%, 0.3%, 0.5%, 1%,). Each SS was sampled for 10 min and averaged over that period, hence a certain SS has an time resolution of 1 hour. The instrument was calibrated with ammonium sulfate particles before and after the campaign according to the ACTRIS protocol (Gysel and Stratmann, 2013).
the exception of the two lowest temperature steps.
The SML has been found to be enriched in particulate organic matter and surface-active substances compared to the underlying 345 bulk seawater, with enrichment factors (EF) of up to 10 and 50 respectively being reported (Engel et al., 2017;Kuznetsova and Lee, 2002). And, as described in the introduction, the SML is known to be highly ice active. It is therefore an interesting The larger markers in Fig. 6 indicate samples where the SML showed significantly higher ice activity compared to the others, i.e., higher INP concentrations (see above). Interestingly, almost exclusively the highly ice active SML samples are the samples which feature the highest EFs, suggesting that enrichment could be an important factor in controlling SML ice activity.

INPs in fog water
Analogous to the SML and BSW samples, N INP was also determined in collected fog water samples. At -10°C we find N INP between the lower limit of our detectable range of 2*10 2 and 2*10 4 L -1 . At -15°C N INP between 6*10 2 L -1 and the upper limit 375 of our detectable range 9*10 4 L -1 , were observed. At -20°C values between 1*10 4 L -1 and the upper limit of our detectable range, 9*10 4 L -1 were found. 14 fog samples (63.6% of all fog samples) have a freezing onset above -10°C suggesting the presence of biogenic INPs as mineral dust only starts to contribute to the INPs population at temperatures below -15°C (e.g. Murray et al., 2012;O'Sullivan et al., 2018). The highest freezing onset we observed in a sample was at -3.47°C. The samples are divided into two groups by a clearly recognizable gap. The occurrence of these two groups could not directly be related to 380 meteorological parameters. However, as will be discussed in section 3.3.1, the group of more ice active fog samples may be associated with the more ice active atmospheric filter samples.
In general the fog samples tend to be more ice active and show higher N INP at a given temperature than the seawater samples presented in section 3.2. A qualitatively similar observation was already made by Schnell (1977). For seawater samples they collected near Nova Scotia (Canada) they found that some of the samples were very ice active, although the majority of their 385 seawater samples contained no INP active at temperatures warmer than -14°C. On the other hand half of their fog water samples were ice active at temperatures above -10°C with the most ice active sample initiating freezing at -2°C. Schnell (1977) also described that they found N INP in seawater, fog and air to vary independently from each other. An observation that also largely applies to this study, but a more detailed investigation of the relation between N INP in the different compartments is presented in the following sections (3. 3.1 and 3.4).

390
The N INP (T) we observed in Arctic fog water is similar to what Gong et al. (2020) found in cloud water samples on the Cape Verde Islands, but tends to be lower than what was observed by Joly et al. (2014), who measured at Puy de Dôme (France) and reported a correlation between high concentrations of biological particles and INP concentrations. However the freezing onset temperature of around -6°C is almost identical in the three studies. In this section we relate and compare N INP in fog water samples with those measured in fog-free air (see section 3.1), following the procedure introduced in Gong et al. (2020), which is briefly described in the following. The number concentration of CCN (N CCN ) at a particular supersaturation (SS) is used as proxy for the fog droplet number concentration. Furthermore, Gong et al. (2020) made the legitimate assumption, that all INPs act as CCN. Together with an estimated fog droplet diameter (d drop ), the volume of fog water per volume dry air, LWC fog can be calculated as follows: For determining N CCN a SS needs to be defined. Since fog, unlike clouds, is characterized by low updrafts, SS is also typically low (0.02% -0. 2% Pruppacher and Klett, 2010). Thus we choose N CCN measured at SS = 0.15% as proxy for the droplet number concentration. Please note that we do not use N CCN measured at SS = 0.1%, because after the removal of data points due to quality assurance, the data coverage for SS = 0.15% is significantly better than for SS = 0.1%.

Connecting atmospheric INPs to sea spray 430
In order to assess the ocean as possible source of atmospheric INPs, we derive potential atmospheric N INP by virtually dispersing the characterized seawater samples as sea spray (Irish et al., 2019b;Gong et al., 2020 where NaCl mass, air , and and NaCl seawater are the mass concentrations of sodium chloride in corresponding air and seawater 440 samples, respectively. NaCl mass, air varied between 0.04 and 1.9 µg m -3 during the campaign with an average of 0.48 µg m -3 . The average NaCl seawater of all SML and BSW samples is 32.5 g L -1 with actual concentrations varying between 25.7 and 34.5 g L -1 . NaCl seawater was derived from the salinity of the samples with the simplifying assumption that NaCl is the only salt in the sea water. This assumption is justified as non-NaCl salts represent only minor constituents of the sea water. Samples from meltponds are excluded here and also in the following as they are mostly fresh water and therefore not suited for this 445 approach that is based on NaCl concentration. Fig. 10 shows atmospheric, filter derived N INP in gray and the sea spray derived N INP sea spray, air (red symbols correspond to SML samples and blue ones to BSW samples). As can be seen, N INP sea spray, air falls mostly in the range between 10 -6 and 10 -2 m -3 , which is approximately 4 to 5 orders of magnitude lower than the atmospheric N INP derived from our atmospheric filter samples. Our concentration range of 10 -6 to 10 -2 m -3 is also roughly two orders of magnitude lower than the ranges 450 reported by Gong et al. (2020), who sampled near the subtropical islands of Cape Verde during late summer, and Irish et al.  (Keene et al., 2007;Van Pinxteren et al., 2017) and 10 2 for supermicron SSA (Quinn et al., 2015;Keene et al., 2007). As we have no information about the size of the INPs, except that they are larger than 0.2 µm, we cannot say what enrichment factor would be an appropriate assumption in regard to INPs, but the above mentioned literature indicates that processes exist that can produce sufficiently high enrichment factors at least for some substance classes. But it should be also noted, that the laboratory study by (Ickes et al., 2020) ::::::::::::::: Ickes et al. (2020) did not find a correlation between total organic carbon content of algal culture 465 samples and the freezing of the sample. The same study confirmed that the transfer of ice nucleating material from the seawater to the aerosol phase can indeed happen. Therefore a marine source for the INPs in the Arctic atmosphere cannot be ruled out, but considerable enrichment of INPs during the transfer from the ocean surface to the atmosphere would have to take place. In section 3.1, we described that INP concentrations are different in the ice free ocean, within the ice pack, and close to land.
In the following we will show, that merely the proximity to land does not make marine INP sources inferior to terrestrial ones. To elucidate this, we consider a time period of several filter sampling intervals which occurred around a time when both, atmospheric and INP concentrations in the SML, were highest. This happened close to Svalbard and in the vicinity of the ice edge, which makes the situation even more interesting.

475
The overall most ice active SML sample, SML37, was taken on July 15th, 10:50, and is highlighted in Fig. 11 (lower panel, light blue symbols). It occurs at the beginning of the sampling period of LV194, the second most ice active atmospheric filter sample (Fig. 11, upper panel, blue symbols). A number of atmospheric samples collected before and after sample LV194 are also shown. Most of the N INP (T) spectra from these samples have a very similar overall shape, featuring a fairly steep increase at temperatures above -10°C, followed by a plateau region between ca. -10°C and -21°C and another, but less steep increase 480 below -21°C. Such a behaviour is indicative for the presence of distinct INP populations and therefore not many mixing events happened during transport (Hartmann et al., 2020;Welti et al., 2018). Additionally, the INPs active at these warmer temperatures are likely biogenic and proteinaceous as indicated by heat tests described in section 3.1.
INPs from the atmosphere are deposited into the SML. However this is highly speculative and needs further research.
To further elucidate the possible connection between atmospheric INPs and INPs in the SML, in the following we consider 495 additionally available aerosol related and meteorological information.
To broaden the perspective beyond the aforementioned measurements at the position of the ship itself, HYSPLIT back trajectories were also assessed. In Fig. 13   shows to which sample the trajectory belongs (consistent with Fig. 11 and 12). The SSMIS sea ice concentration with emphasized ice edge on 15 July 2017 is also shown (Sea ice concentration product of the EUMETSAT Ocean and Sea Ice Satellite Application Facility).
in Fig. 12 are shown. The colour code indicates into which collection time, i.e., sample the respective trajectories fall (corre-515 sponding to background colors used in Fig. 12 material, and therefore these dust related-processes may also explain spatially confined areas of high ice activity without being contradictory to the assumption of biogenic INP.
-Heat tests indicate that INPs active above -15°C are biogenic and proteinaceous.
-The freezing spectra of atmospheric INPs and INPs from the SML feature similar slopes at temperatures above -10°C, suggesting a connection between both compartments, which, however, as discussed above, would need a substantial enrichment of INP during the sea spray production.

545
-Aerosol particle parameters show that clearly different air masses arrive at Polarstern over the course of the case study.
-Backtrajectories indicate that sampled air masses have different regions of origin and travel over different pathways towards Polarstern.
-Elevated Chlorophyll-a concentrations were observed for a short phase directly at at the position of Polarstern (Ferrybox) and also in the wider geographical region in the week-long satellite composite. This indicates a high biological activity 550 in the investigation region.
We interpret these findings as strong indication for a local marine source being present during our case study. Seemingly this is in contradiction to the results gained from the analysis of fog-water as presented above, unless a significant enrichment of INPs takes place during the aerosolization of seawater and/or SML material. In other words, there is a strong need for gaining knowledge concerning the mechanisms of aerosolization and resulting fluxes of INP and related species at the ocean-555 atmosphere-interface.

Summary and Conclusions
We of Africa, which indicates that at these low temperatures dust is an important INP even in the Arctic.

570
SML samples from the biologically active MIZ have a higher fraction of highly ice active samples than the other ocean compartments (ice-free ocean, ice pack, melt pond). Besides that the ice activity of SML and BSW samples is not simply correlated with the environment the sample was taken from. In general, few highly ice active samples stand out against the other samples. Except for one case, we found the SML to be weakly to significantly enriched in INPs compared to the underlying BSW. The enrichment factors (EF) varied between close to 1 and 94.97 at -15°C. The most enriched samples featured the 575 highest ice activity : in ::: the ::::: SML ::::::: samples.
From INP concentration in the fog water and the measured CCN number concentrations we derived potential N INP in the air, which we compared to the directly measured N INP , and found good agreement. This indicates that the same, or at least similar, INP populations were present in corresponding fog water and air samples, suggesting that during fog events INPs are activated to droplets and become available as immersion nuclei inside the fog droplets.

580
Using the ratio of NaCl mass concentration in the air and in the seawater as a scaling factor, we assessed if atmospheric N INP can be explained simply by aerosolization of SML and BSW material. At any given temperature we found SML-and BSW-derived N INP to be 4 to 5 orders of magnitude lower than the N INP directly measured in air. This clearly shows that aerosolization of SML or BSW material, without significant enrichment of INPs during aerosolization, does not suffice to explain N INP in air. In other words, a marine source for the INPs in the Arctic atmosphere is possible, but enrichment of INPs In a case study we looked more deeply into a scenario for which coinciding SML and air samples were highly ice active.
Thereby, we found similarities in the temperature spectra of the highly ice active INPs in the SML and in the air. Air mass changes, indicated by changes in aerosol properties and back trajectories, did not cause changes in the observed INP population.
Isolated patches with chlorophyll-a concentrations of about one order of magnitude higher compared to their surroundings 590 underline high biological activity in the investigation region for the time period we investigated in the case study. We consider this as indications for a local biogenic marine source of INPs being present.
Altogether, we found INP concentrations in air, fog water, SML and BSW to be highly variable, with a small number of cases featuring significantly enhanced ice activity. This emphasizes the episodic, highly variable nature of INPs as it was already described decades ago by Bigg (1961). This puts a question mark to the appropriateness of parameterizations based on aerosol 595 particle number in atmospheric models. We found indications for a marine biogenic INP source, however further investigations are needed to gain quantitative knowledge concerning the aerosolization process and the resulting INP fluxes at the interface between the atmosphere and the ocean surface.
Lastly, to take up the questions from the introduction: -What is the abundance of Arctic INPs and in what temperature range can they nucleate ice? 600 We found INP active between -7°C and -38°C over a concentration range from 4 * 10 -1 m -3 to 1 * 10 8 m -3 . Most of the time N INP was at the lower end of N INP range known from mid-latitudes or even lower. Exceptions were the upper and lower end of the temperature range: At -10°C N INP of up to 6 * 10 1 m -3 were observed, while at -32°C N INP was in the same order of magnitude (10 5 m -3 ) as in the outflow region of the Saharan desert.
-What is the nature of Arctic INPs (biogenic material vs. mineral dust)? 605 We find indications that the warmer temperatures (>-15°C) are dominated by biogenic INP, while at colder temperatures (<-25°C) likely mineral dust dominates.
-What is the origin of Arctic INPs (local vs. long range transport, marine vs. terrestrial)?
For the INP at warmer temperatures we find indications that they are marine and locally emitted, which, however, necessitates an enrichment of INP during sea spray aerosol production of several orders of magnitude.. measurements. We also thank Amelie Assenbaum, Audrey Brown, Mareike Löffler and Jasmin Lubitz for their assistance with the cold stage