|The authors spent a great deal of effort addressing the reviewer suggestions, which is much appreciated, and the paper is substantially improved. The logic of the paper and the figures are now much clearer, and I am happy to recommend it for publication, after a few minor comments detailed below are addressed.|
p.16, l.20: “Because enhanced rBC and CN concentrations are expected to be good indicators of anthropogenic activity, they are used to isolate clouds impacted by anthropogenic emissions.”
I agree that rBC is a good indicator, but I am not so sure about CN. In this paper, CN is defined as the number count of aerosols between 3 or 10 nm up to 3 microns in diameter (p.5, l.35). Many local non-combustion derived aerosols pop up in the lower diameter range between 3-60 nm (e.g., Tunved et al. 2013). These very small background particles can be so numerous that they can dominate the CN count (e.g., Zamora et al., 2016). That could be one explanation of the large discrepancy between Figures 8 and 10 in the aerosol distributions relative to reff, why Figure 12 and Figure 8 are so similar, and why there are more brown points in Figure 13b than in Figures 13a and 13c.
p.17 l.4: “We conclude that CN and rBC particles, which were used to classify local clouds, have the potential to grow to accumulation mode particles measured by the PCASP.”
Again, I feel that CN is inappropriate to use for this purpose – PCASP concentrations would likely be better. Also, from the author’s definition of CN, the CN data should already include the data in the PCASP ranges. This is confusing to me.
P.19, l.7: “When applying the linear regression to the data sets corresponding to the two sites separately, the obtained ACI values differ (Table 1), with OLI having a lower ACI value (0.12±0.05) than NSA (0.20±0.07). Given the small sample size (24 and 16 cases for OLI and NSA) which was caused by the PCASP data being quality-flagged for some cases, it is not possible to determine whether this is caused by a difference in nucleation efficiency between aerosols at the two sites or a random effect.”
How were the errors in ACI value calculated? Is the difference in ACI values between the two sites significant? Given the error ranges listed here, and the high spread of data shown in Figure 14, it looks likely to me that they are likely not significantly different, in which case discussion of the differences is not appropriate.
P20, l.5: For the following reasons, I feel that the following new text is too speculative, and suggest removing it:
“The lower R2 value for OLI (0.24) in comparison to NSA (0.40) could indicate that the assumption that PCASP particle concentrations represent a good approximation for CCN concentrations is partly violated at OLI.”
As mentioned, I’d like to know if the ACI values actually were significantly different.
“This could result from those particles being less aged and consequently less coated by sulphates and organics in comparison to those observed around NSA.”
These data are not provided.
"In addition, some data points lie above the 1:1 line which might indicate that particles smaller than the PCASP size range (i.e. < 100 nm) are acting as CCN (Leaitch et al., 2016)."
What figure is being referred to here - Figure 14?
"Further, the assumption that the below-cloud aerosol properties govern the cloud microphysical properties might not be true for all clouds depending on sub-cloud vertical mixing. Therefore, we identified all clouds where the above-cloud PCASP concentration is larger than below-cloud (red dots in Fig 14), and indeed half of these clouds are above the 1:1 line. When using the above-cloud concentration for these clouds, only two of these clouds are above the 1:1 line. However, there are still 11 more clouds above the 1:1 line. Since these clouds generally feature PCASP concentrations < 50 cm-3, the fact that they are above the 1:1 line could be related to increasing sampling errors for small concentrations. They may also confirm the finding of Leaitch et al. (2016) that aerosols below 100 nm can act as CCN for thin clouds.”
This last argument is based on a very low sample number. Especially due to this low sample size, detailed description of the instrument/sampling error is key to giving any credence to the hypothesis. To “confirm” the Leaitch et al findings, the authors would need to show a lot more information on this error than what is provided here. Without that information, I believe the above discussion is much too speculative.
p.2, l.3: Because of these observations…
p.5, l.5: Suggest: In the following work
p.6, l.15: I still disagree with the phrasing here, because while these tracers are emitted by fires, they are also affected by other processes. BC and CO are only good tracers of smoke for in situ data if the other processes affecting the BC/CO concentrations are unimportant at a specific place and time. I suggest rephrasing to something like, “Therefore, we manually inspected the vertical profiles of rBC and CO, which together can be used to trace biomass burning in otherwise clean environments (Warneke et al., 2009, 2010).” (Note: if you make this change, the reference is probably not appropriate anymore, since I believe Warneke et al. used a model to identify smoke-related BC in non-clean environments)
For the same reason, on p.6, l.17, I suggest you change to: “For each spiral obtained at the two sites, elevated layers with CO >= 0.1 ppmv or rBC >= 20 ng kg-1 were flagged as potentially corresponding to forest fires.”
p. 6, l.18-19: “Local emissions, on the other hand, are expected to be found in a layer connected to the surface” – how sure are you about that? Do you mean “Local emissions, on the other hand, are expected to be concentrated in the layer…”?
p.8, l. 29: e.g.?
p.10, l.24: what does the question mark mean?
Figure 1: it would be helpful to state in the caption how fire source was identified
Tunved, P., Ström, J., and Krejci, R.: Arctic aerosol life cycle: link- ing aerosol size distributions observed between 2000 and 2010 with air mass transport and precipitation at Zeppelin station, Ny-Ålesund, Svalbard, Atmos. Chem. Phys., 13, 3643–3660, doi:10.5194/acp-13-3643-2013, 2013.
Zamora, L. M., Kahn, R. A., Cubison, M. J., Diskin, G. S., Jimenez, J. L., Kondo, Y., McFarquhar, G. M., Nenes, A., Thornhill, K. L., Wisthaler, A., Zelenyuk, A., and Ziemba, L. D.: Aircraft- measured indirect cloud effects from biomass burning smoke in the Arctic and subarctic, Atmos. Chem. Phys., 16, 715–738, https://doi.org/10.5194/acp-16-715-2016, 2016.