This manuscript is well-written in general. While this reviewer is not 100% sure about the atmospheric implication of the study outcomes, this paper discusses an important negative result regarding potential ambient ice nucleation – surface active macromolecules do not dominate INPs in bulk soil samples. The study topic is relevant to the journal scope of ACP. This reviewer supports the publication of this paper in AMT after addressing several questions below.

===

Questions

(1) Can surface tension and micelle formation of heated soil extracts be measured? If yes, can they be reported in this manuscript? Doing so might clarify the next question.

(2) Can heat-treated soil extracts show what the authors find in lignin (an increase in conc is proportional to an increase in IN activity and a reciprocal decrease in surface tension)? Adding discussions regarding surface properties of heat-treated soil suspensions at this stage might be worth to shed light on the relative importance of surface active MMs to IN active protein in soil samples. Recently, as the authors discuss in L35-37, some heat-sensitive proteins/enzymes commonly found in soil, such as Rubisco, have been found to be IN active. Heat treatment presumably removes these proteins, and the remaining soil organics can show lignin-like surface properties?

Comments

P1L4: Why the impact needs to be on only ‘local’ cloud formation?

P4L119-121: How long the tube was opened to sample ambient air? Quantitatively clarifying the blank handling protocol in the manuscript will be helpful for the reader and whoever may follow up on your study.

P5L132: Not SA water?

The study presented in the manuscript was motivated by the hypothesis that hydrophobic interfaces might have an influence on how organic macromolecules nucleate ice. To test this hypothesis, dilutions of lignin and snomax were tested for surface tension and ice nucleation properties. Increasing concentrations of these compounds in water resulted in decreasing surface tension and an increasing number of ice active sites. In contrast, soil extracts were reduced by dilution in the number of ice active sites with little unsystematic changes in surface tension.

All laboratory work was done with great attention to detail. Overall, the study is clearly written and its results offer new insights, as far as the comparison of lignin and snomax with soil extracts is concerned. Regarding the correlation between surface tension and ice active sites observed in dilutions of lignin and snomax, I wonder whether this is not a foregone conclusion. The substances tested were already known before the experiment to have ice nucleating and surfactant properties at the same time. That both properties are less expressed at higher dilution levels is revealing the obvious. More interesting would be to have a closer look and discuss trends in T50 within dilution ranges where surface tension did not change (lignin <10^2 mgC/L; snomax <10 mgC/L), but T50 tended to increase with increasing concentration of C.

Fig 2: Considering the counting error, is the number of freezing events at AWI significantly different from that in other compartments? If I counted correctly, 11 events occurred at the AWI, 3 in the neighbouring shell, and 6 in each of the three inner compartments. The difference between counts in AWI and in the three inner compartments does not appear to be significant because these low numbers of counts are associated with a relatively large uncertainty. Their standard deviation is roughly equivalent to the square root of counts. In identical replications of this experiment, AWI counts might well turn out to be the same (11 -√11)  as in inner compartments (6 + √6).

Line 270: replace 'step-wide' with 'step-wise'

Line 347 replace 'heat-label' with 'heat-labile'

Figure 6: Consider using the same unit in the y-axis label as in Figure 7 ([(mg C)^-1] instead of [mg C^-1]).