Review:

Mexican agricultural soil dust as a source of ice nucleating particles

Summary:

In their manuscript, Pereira and colleagues studied the ice-nucleating ability of dust aerosols originated from different agricultural fields at four sites in Mexico. They collected soils from different croplands that were later aerosolized and analysed in the laboratory and compared the lab data set to aerosol samples collected in-situ in the field. The ice nucleation activity (INA) of aerosols was measured in immersion freezing mode which is relevant to ice formation in mixed-phase clouds. The authors draw conclusions on the extent to which organic carbon and certain minerals contribute to the freezing activity of dust particles and discuss the importance of particle chemical composition and size for efficient ice nucleation.

The Introduction is very well written and represents the current scientific knowledge in this field. The methods and sampling design used in this study both in the field and in the laboratory are adequate and the presented results are of general interest to the INP community. Overall, the work presented merits publication in Atmospheric Chemistry and Physics after minor revisions.

General comments:

One conclusion reached by the authors is that agricultural soil dust contributes to the formation of mixed-phase clouds (line 368). However, the data presented in this manuscript do not fully support this conclusion. Specifically, transport measurements were not part of this study and consequently, the influence of the sampled aerosol on cloud formation can only be assumed. Indeed, it could be that these soil dusts are transported to altitudes where mixed-phase clouds appear and the INP concentrations from this study match with literature data, however more data is needed to proof this hypothesis in this context. Therefore, I propose to weaken the according argument in the conclusion. The manuscript could be improved if the authors deepened the discussion of possible transport processes of soil from fields to high altitudes.

The authors use freezing curves to depict the INA of their aerosol samples. However, in the discussion the authors focus the interpretation rather on T50 values than the freezing curves themselves. T50 values have been used in the scientific community to represent a sample’s INA. For pure ice nuclei that freeze immediately at a given temperature, T50 is a good parameter for comparison studies. However, if one sample consist of more than one ice nucleus (e.g. OC and K-feldspar), freezing curves may show steps in the spectrum (see e.g. onion (F), wheat (F) in Figure 3). Consequently, the T50 values does not fully represent such samples. Have the authors considered also taking T10 and T90 into consideration or include a more detailed discussion about the spectra?

Specific comments:

Abstract:

Line 32: The authors state that T50 values and aerosol particle size are correlated. There is no mention of what exactly is meant by particle size. I assume aerodynamic diameter? I would recommend mentioning this in the text.

Line 33: Please indicate precisely which efficiency is meant. Ice nucleation efficiency?

Introduction:

Line 61: not all bacteria and fungi are ice nucleation active. I would recommend to add ‘certain’ to the sentence: ‘[…] (e.g. certain bacteria, fungi) […]’

Methods:

Line 111: I was wondering what the weather conditions were like during the campaign? Did the authors record any data on the meteorological conditions?

Line 111: ‘[…] samples collected at the ground level.’ How much distance was between the instruments and the ground? Did they actually stand on the ground or were the instrument mounted onto something?

Line 116: Is there a specific reason why aerosol and soil samples were stored at different temperatures?

Figure 2: I was wondering if the ‘OPC sign’ shouldn’t be indicated before the ‘aerosol collector sign’? Or was the OPC attached after the impactor/MiniVol?

Line 115 and line 123: For how long did you collect your aerosols? Was the time period in the field the same as in the laboratory?

Line 130: Did the authors collect the aerosol samples for the ice nucleation and mineral analysis in parallel or were the measurements performed sequentially?

Line 146: Are the authors referring to the diameter of the droplets (d=170 um) as the size?

Line 147: I would recommend writing ‘dry nitrogen’, rather than dry air.

Line 147: I assume that the droplets were not evaporated completely but rather reduced to the desired size? Please clarify.

Line 152: Could the authors give a little bit of insight into the statistics of their setup? How many droplets/particles per glass slide were analyzed for one sample?

Line 181: This sentence is incomplete. What is meant by ‘at 0.85%’? Does that refer to a concentration of the buffer? Further, could the authors specify which sterile solution they used? In addition, was the cultivation performed at room temperature? Maybe the authors could mention that in the text.

Results and Discussion:

Figure 3: For clarification to the reader, I would suggest to mention the aerodynamic diameter as the size in the caption.

Figure S1: It is not clear to me how the particle size distribution was recorded for the field measurements. Could the authors please state in the ‘Method’ section how the measurement was performed?

Line 212: I was impressed that the type of crop which previously grew on the field influenced the ice nucleation ability of the soil. Do you think field treatments (e.g., pesticides, fertilizer or no artificial treatments that could promote biodiversity and possibly ice-active microbes) could affect the INA of the soil? In addition, do you think that freeze tolerant plants (see e.g. Marcellos and Single, 1979) may leach ice nuclei into the soil?

Figure 5: I would recommend to add the temperature and duration of the heat treatment to the caption.

Figure 3 and 5: When printing out the manuscript the lines of Corn 2 (L), Bean (L) and Onion (L) are hard to distinguish. The authors may consider changing the colours.

Line 281: I recommend to also cite Zolles et al. (2015) here, as they showed that the INA of e.g. microcline decreased by 2 degrees after the sample was heated to 250°C.

Technical remarks:

Line 45: INP should be in plural (INPs)

Line 145: Missing the letter s: ‘Afterwards, […]’

Line 640: The micrometre symbol is in a different style

References:

Marcellos, H., & Single, W. V. (1979). Supercooling and heterogeneous nucleation of freezing in tissues of tender plants. Cryobiology, 16(1), 74-77.

Zolles, T., Burkart, J., Häusler, T., Pummer, B., Hitzenberger, R., & Grothe, H. (2015). Identification of ice nucleation active sites on feldspar dust particles. The Journal of Physical Chemistry A, 119(11), 2692-2700.

General Comments

This work evaluates the ice nucleating ability of Mexican agricultural soil dust and discuss the possible factors influencing its ice nucleating ability. Since “the ice nucleating abilities of agricultural dust particles from the Mexican territory have not been reported up to date (Lines 84-85)”, the dataset in this area will be potentially helpful to improve our knowledge of the possible impact of agricultural soil dust on ice nucleation in mixed-phase clouds. Unfortunately, however, there are some problems regarding the experimental approaches. For example, it is not clear when and how the soil and aerosol samples were obtained. The use of sized-resolved INP data are unique, but it is disappointing that only limited INP data in a narrow size range are reported in this manuscript. Although this work tries to evaluate the possible importance of organic and mineral components in the agricultural soil dusts using several approaches, the conclusions are obscure and not well summarized. In addition, it is difficult to compare the results presented in this manuscript with those from other previous studies which evaluated the ice nucleating ability of agricultural soil dusts in other locations, because this study uses T₅₀ values while most studies have used other parameters like the ice-nucleation active site density (INAS). For these reasons, it is almost impossible to evaluate the quality of the data and whether the key conclusions presented here are scientifically appropriate. I think that the quality of this manuscript should be significantly improved before considering publication.

Specific Comments

1) It is not clear when the soil and aerosol samples were collected. As for the soil samples, please clarify when the samples were collected in Table 1. As for the aerosol samples, please clearly explain when and where each sampling was started and finished (this information should be summarized in a table).

2) When did you collect the soil samples used for microbiological analysis and then when these were analyzed? If the authors would like to measure the colony forming units (CFUs) per gram, the analysis should be performed just after sampling, because the microorganisms in soils would be significantly changed once these were collected and stored at room temperatures. If the samples were not analyzed immediately, the data would not be scientifically valuable, and all the data and description about the CFU data (Lines 179-185; Lines 264-275; Figure S4) should be removed.

3) I could not understand why only the INP data obtained from the MOUDI stages 3 to 6 (Figures 3 and 5) are reported, despite the facts that the samples were likely to be collected on the eight stages of the MOUDI (Lines 130-133). Although the authors explain that “the present results focus on particles >0.56 μm as it has been shown that particles >0.5 μm have a higher potential to act as INPs (e.g., DeMott et al., 2010) (Lines 190-191)”, I think that they could evaluate whether the particles larger than 0.5 μm indeed have a higher potential to act as INPs than those smaller than 0.5 μm based on their data collected on the MOUDI stages 7 and 8. It is also unclear why the INP data larger than 5.6 μm (MOUDI stages 1 and 2) are excluded.

4) In addition, I would like to suggest showing the total INP data (i.e., the sum of INP data obtained from all the MOUDI stages).

5) A fatal flaw of this work is a lack of the reliability of the field INP data. As I already point out in Comment 1, it is not clear when and how the field samples were collected. However, if only 4 field INP data (only 1 field sample at each location) are available as shown in Figure 3, the numbers of the data are too small. Furthermore, it is also unclear whether the INP population in the field data were indeed well characterized by local agricultural soils, because their source and composition are not evaluated. Given the location of the sampling sites (Figure 1), there is the possibly that the aerosol population in the field INP data were characterized not only by local agricultural soils, but also by various types of aerosols from oceanic and urban/rural sources. If the authors would like to discuss the difference between the field and laboratory INP data based on only 4 field data, they should provide strong evidence that the available field data were mostly characterized by local agricultural soils. Otherwise, the discussion and conclusion about the comparison of the field and laboratory data presented in this manuscript (e.g., Lines 28-30; Lines 189-210; Lines 369-372) would not be scientifically valuable and hence should be removed.

6) Although the authors describe that “the differences between laboratory and field environments are also reflected in different PSD observed during the aerosolization process (Fig. S1) (Lines 205-206)”, I doubt if the particle size distribution (PSD) of the filed (F) samples would be totally characterized by local agricultural soils. Please show evidence that all the sizes (0.3-10 μm) of the field samples were entirely characterized by local agricultural soils.

7) If the authors consider that “the highest particle concentration for the L samples was found for particles between 1.0 μm and 5.0 μm (Fig. S1a), while the F samples are enriched in smaller particles, i.e., 0.3 μm (Fig. S2b) (Lines 208-209)” and “the larger particles present in the L samples likely promoted ice nucleation at warmer temperatures (line 210)”, the authors would need to evaluate the ice nucleating ability of the laboratory (L) samples in the size range smaller than 0.56 μm (see also Comment 3).

8) I cannot understand how Figures S2 and S3 are prepared. In addition, I cannot understand why the results of the agricultural dust particles in the laboratory (before heat treatment) are different between Figures S2 and S3. Please explain the details of these figures.

9) Previous studies (Tobo et al., 2014; Steinke et al., 2016) have reported the INAS of agricultural soil dust, instead of T₅₀ values. If the authors think that “the ice nucleation temperatures observed in the present study are on the same order as those reported for agricultural dust in Wyoming (USA), from -18ºC to -36ºC for dp=0.6 μm (Tobo et al., 2014), and Argentina, China, and Germany from -11ºC to -26ºC for dp <5 μm (Steinke et al., 2016) (Lines 231-233)”, the INAS of agricultural soil dust presented in the manuscript should be calculated.

10) Although this work tries to discuss the possible influence of organic matter (Section 3.2) and mineralogy, especially K-feldspar (Section 3.3) using several approaches, it is hard to understand the key conclusion of this study. Do you think which component would have more impact on the ice nucleating ability of the agricultural soils? Please discuss this point and explain in the Abstract and Conclusion sections.

11) Although the correlation map is shown in Figure S5, it is still hard to imagine the relationship of T₅₀ values with OC and mineral compositions for each sample. I would like to suggest preparing some scatter plots that compare these parameters (e.g., INPs vs. OC, K-feldspar, etc.).

12) In addition, I would like to suggest comparing the total INP data (see also Comment 4) with OC and mineral compositions, because the OC and mineral contents reported here would be based on the analysis of the bulk samples smaller than 10 μm. This result might be helpful to answer Comment 10 (which component would have more impact on the ice nucleating ability of the agricultural soils?).

Technical Corrections

13) Line 22: ice crystals formation => ice crystal formation

14) Line 62: improving => enhancing or influencing (or something like this)

15) Lines 68-69: Garcia et al. (2012) reported the results from aerosol sampling in the air and not in soils.

16) Lines 95 and 104: The terminology like Morelos (MOR) and Morelos (ZAC) are confusing.

17) Line 193: solid lines => solid curves?

18) Line 194: dotted lines => dashed curves?

19) Line 209: Fig. S2b => Fig. S1b?

20) Line 243: dp <0.6 μm => dp = 0.6 μm

21) Line 301: What is “Hunucmá sample”?

22) Table 1: Please clearly explain why the samples are labeled as “Nopal”, “Corn”, “Bean”, “Chili”, “Wheat”, “Onion”, and “Corn 2” (please add some explanations in the Methods section). In addition, what do you mean by “°O” and “N°”?

23) Figure 1: It is hard to see the location of each sampling point, because the same symbols with similar colors are used.

24) Figure 2: It is hard to understand the experimental setup. More detail schematic images (particularly, around aerosol collector) should be presented.

25) Figure 3: It is hard to see the difference of curves, because similar curves and colors are used. Please prepare more eye-friendly figures.

26) Figures 3 and 5 caption: d) 0.56.1.0 μm => 0.56-1.0 μm

27) Figure 4 caption: Organic carbon, elemental carbon, and mineral contribution ~.

28) Figure 5: It is hard to see the difference of curves, because similar curves and colors are used. Please prepare more eye-friendly figures.

29) Figure 7: Please explain how the red-colored range of “Field samples (this study)” is defined. In addition, the authors would need to prepare a supplementary figure showing the comparison of the INP data with this red-colored range. In this figure, the INP data should be the total INP number concentration (see also Comment 4).

30) Figure S1: I would suggest the use of dN/dlogD_p instead of aerosol concentrations (N). In addition, please clearly explain when and how the data in Figure S1b were measured.

31) Figures S2 and S3: Please explain what the box-and-whisker plots indicate.