The following points are suggested for revision of the manuscript:
Main points:
- Improving and completing the description of the CCN data analysis:
• The correction of the CN and CCN data for multiple charges and the transfer function is not trivial. The authors should either describe exactly how their data were corrected (e.g. in appendix, if not wished in the text) or at least reference the papers that describe the methods they used.
• It is not always clearly distinguished between bulk and size-resolved CCN measurements. This could be easily done, e.g., by using different symbols and/or using distinct terms for the derived parameters (e.g. total CCN number concentration vs. CCN number concentration (measured behind the DMA, in size-resolved mode) and activation ratio vs. size-resolved activation ratio).
- Improving the Results section:
• The whole section on the CCN closure is totally inscrutable. It needs to be clarified which data sets exactly were used for which type of closure. I have read this section at least ten times and I still cannot figure out from the text which parameters are plotted in Figs. 5 and 6. I have done this kind of closure tests myself several times. Thus, I have an idea of what might be shown in the two figures but it is certainly not clearly described in the text. Also, because the method is not described properly, it is hard to understand what the results tell us. Therefore the description of the method as well as the discussion of the closure results needs to be revised.
The parameters I would calculate and compare with each other for your study would be the following:
a. The actually measured total CCN concentration directly measured by the CCNC.
b. The observed total CCN concentration obtained from size-resolved CCN measurements: the actually measured CCN efficiency spectrum is multiplied by the actually measured CN size distribution, which yields the CCN size distribution; then this is integrated over the whole size range to obtain the total CCN concentration.
c. The predicted total CCN concentration obtained from the average CCN spectrum: the average measured CCN efficiency spectrum (would correspond to the spectra plotted in your Fig. 1) is multiplied by the actually measured CN size distribution, which yields the CCN size distribution; then this is integrated over the whole size range to obtain the total CCN concentration.
d. The predicted total CCN concentration obtained from the average CN size distribution: the actually measured CCN efficiency spectrum is multiplied by the average measured CN size distribution, which yields the CCN size distribution; then this is integrated over the whole size range to obtain the total CCN concentration.
The following comparisons of parameters would be interesting:
o a and b: This would be the easiest way to check if in general the size-resolved CCN measurements are consistent with bulk measurements. This comparison cannot be done in this study since no parallel bulk and size-resolved measurements are available
o b and c: For the time period of the size-resolved CCN measurements the total CCN concentration is 1) derived from the actually measured CCN spectrum and 2) calculated from the averaged CCN spectrum. With this comparison the influence of the variation of the chemical composition on the CCN concentration can be investigated, since the CN size distribution is the same for both the parameters. I guess this is what is shown in Fig. 5
o a and c: For the time period when the bulk CCN measurements were performed the measured total CCN concentration is compared with the calculated concentration from the averaged CCN spectrum. With this comparison again the variation of the chemical composition can be investigated but also how the bulk compares with the size-resolved CCN method. I guess this is what is shown in Fig. 6 but if so which average CCN spectra were then used for the calculation? The size-resolved CCN measurements, from which the CCN spectra can be derived, were made at different supersaturations than shown in Fig. 6.
o b and d: This comparison would be interesting for investigating the influence of the variation of the size distribution on the CCN concentration because here the CCN spectrum is the same for both the parameter. I guess that this reveals an even weaker correlation than for Fig. 6.
• If MAF<1 is only due to an error of the CDF fit (p. 16, l.9-10), why the whole discussion on Da vs. Dcut is needed? It should be enough to report Da values only since Dcut is anyhow practically the same. Btw, whether MAF<1 is an error of the CDF fit could be shown by plotting the data points in addition to the fit lines in Fig. 1!
- Improving the presentation quality:
• The manuscript is written in bad English. In some cases, grammar mistakes even change the meaning of the sentence. Apart from that, the methods and results are not presented in a clear and well-structured way. While revising this manuscript you should make sure that “the description of experiments and calculations is sufficiently complete and precise to allow their reproduction by fellow scientists“ (cf. ACP review aspects).
Specific remarks:
p. 2, l. 5-7: “A gradual increase…” This sentence is weird. I think what you mean is “A gradual increase of the activation ratio (AR) with particle diameter suggests that aerosol particles consisted of a variety of different hygroscopicities.”
p.2, l.7-10: Better formulation: “During pollution events the activation diameter (Da) measured at low supersaturation (SS) was significantly increased compared to background conditions; the increase was not observed when SS>0.4%”
p. 4, l. 5: ”aerosol particle formation” instead of “aerosol formation”
p. 4, l. 17: add “particle number” before “size distribution”
p. 5, l. 23-24: it must be “Rose et al., 2010; 2011; Gunthe et al., 2011” instead of “Rose et al., 2008; Gunthe et al., 2009”
p. 7, l. 12-13: “Aerosol chemical composition…” I would shift this sentence to the end of this paragraph.
p. 8, l. 12: Which temperature sensor do you mean, inlet or sample temperature?
p. 15, l. 15: Please give an error estimation for the supersaturation levels.
p. 9, l. 22-23: What means “CCN data … were filtered according to the instrument recorded parameters”? Do you mean that you sorted the data? Did you throw out any data points? If so, explain why?
p. 10, l. 24: please add “of a particle” after “describe the ability”
p. 14, l. 16: it says that Fig.1 shows the “measured CCN efficiency”, but this is not true. The figure shows the “CDF fit lines of the average CCN efficiency spectra”. In fact, I would really appreciate if you could show the data points of the average spectra, too.
p. 14, l. 21-23: “A gradual increase…”. This is basically the same weird sentence as in the abstract. Please revise it according to my above suggestions.
p. 15, l. 1-2: Please reword this sentence: “The slope of AR around Da is less steep during polluted events than under background conditions, especially for low SS.”
p. 16, l. 5-7: “It should be noticed that… is equal to 1” I have no idea what these two sentences mean. Please revise.
p. 17, l. 18: change to “… may be due to the high amounts of organics freshly emitted…”
p. 18, l. 18-19: I do not see that “Da_BG and Da_POL show larger variations at lower SS”. Relatively seen, I would say that the largest variation is seen at SS = 0.11% and 0.23%. This would be also consistent with the variations for Da given in Tab. 1.
p. 21, l. 14-18: Here you discuss the correlation of AR at SS=0.2% with kappa_chem, which is also shown in Fig. 8a. However, Fig. 9, which you also refer to at this point, shows AR at SS=0.5%. Can you explain?
p. 23, l. 16-20: “Based on the CDF fit method,… was not observed when SS>0.4%.” These are the same strange sentence as in the abstract. Please see my comments above.
p.23, l. 25: Rephrase as “However, the case is more complex for particles originating from heavy pollution…”
Tab. 1:
- Please indicate the meaning of the values in this table. Are they the arithmetic mean values +/- standard deviation over the entire measurement period?
- I think it would be sufficient to write only one digit after the comma for Da, Dcut, AR.
- Your MAF values decrease with increasing supersaturation. I think it makes physically no sense that all particles with ~ 300nm activate at SS=0.1% but not at 0.8%. If a particle is activated at SS=0.1% it is activated even more easily at 0.8%. Since the MAF values smaller than one seem to be an artifact, which you also mention on p. 16, l.9-10, the whole distinction between Da and Dcut is meaningless. In my point of view, in this manuscript, you should present the Da and respective kappa_a values only.
Fig. 1:
- Please plot the data points of the average measured CCN efficiency spectra, too.
Fig. 2:
- Please indicate in the caption the meaning of the error bars and that this graph is depicting the size resolved CCN data.
- Btw., it would be also interesting to calculate the kappa from the bulk measurements.
Fig. 4:
- What is plotted here? kappa_cut as mentioned in the caption or kappa_a as indicated in the x-axis label?
- I guess the particle size ranges as indicated in the individual panels should be consistent with the values of Tab. 1, but they are not.
Fig. 7:
- Why do you write “AR (*100%)” in the y-axis labels? Is it not just the unitless AR that is plotted here?
- Please define the exact size ranges for the individual modes you integrated the particle number concentration (panel g, h).
- Please revise the figure caption. It is enough to write only once that the background conditions are plotted on the left and polluted conditions on the right.
Fig. 8:
- Please label the individual panels and indicate in the text accordingly.
- In fact, it would be also interesting to see ALL the combinations of AR (for all SS) versus kappa_chem, f44, and N_CN,Acc, N_CN,Ait, N_CN, Nuc; also the cases for which no good correlation is found. Therefore I suggest either to add panels for the other combinations or, what is probably more convenient, to add a table showing the correlation coefficients of all the combinations.
Fig. 9: y-axis label: please remove “(*100%)”. This is not valid, not for AR and especially not for kappa_chem. |
- Printer-friendly version
- Supplement
