Simulating Arctic Ice Clouds during Spring Using an Advanced Ice Cloud Microphysics in the WRF Model
, Jacques Pelon 2, Jean-Pierre Blanchet 1, Olivier Lemoine 1 and Tatsuo Onishi 2Round 1
Reviewer 1 Report
I support publication.
While I applaud efforts to represent ice nucleation in models like WRF with a more physically based parameterization, some caution is warranted.
For example, the following statement appears (line 54): “They [heterogeneous nucleation processes] all require different water vapor mixing ratios”
While the Meyers et al parameterization for ice nucleation does have a dependence on supersaturation, more recent parameterizations do not. See for example, DeMott et al., 2010. I agree that deposition nucleationshould have a supersaturation dependence, but the sense of the field is that temperature is the more relevant variable for immersion freezing.
The parameterization used here is based on laboratory data from Eastwood et al., 2008 and 2009. It is worth noting that the parameterization is based on a contact angle between ice and a substrate, but that contact angle is not measured. It is derived from the heterogeneous nucleation rate, which is in turn derived from ice onset measurements.
These parameterizations are for deposition nucleation. Is that really the dominant nucleation mechanism for these situations. I would have thought that immersion or condensation freezing would be more appropriate. The text suggests that ice-only clouds are the focus here, but if that is the case, please state it explicitly. (Lines 650 to 652 make be think that all these cases were ice-only, but state that earlier in the paper, if it is the case.)
As I suggest above, I am not asking that the authors revise the nucleation schemes in WRF. But a comment on the issues I’ve raised above is warranted.
Other comments:
lines 45 and 46: “… considering particles conversion.”
I was confused by this sentence. What are the particles converting to? Which particles?
Please revise.
Line 783: check spelling of “Meyers”.
References:
DeMott, P.J., Prenni, A.J., Liu, X., Kreidenweis, S.M., Petters, M.D., Twohy, C.H., Richardson, M.S., Eidhammer, T. and Rogers, D.C., 2010. Predicting global atmospheric ice nuclei distributions and their impacts on climate. Proceedings of the National Academy of Sciences, 107(25), pp.11217-11222.
Author Response
Montreal, July 12 2019
1
The authors thank Reviewer#1 for this comprehensive review of the paper. We address below each comment individually (in blue color). Line numbers refer to the original manuscript.
Open Review
(x) I would not like to sign my review report
( ) I would like to sign my review report
English language and style
( ) Extensive editing of English language and style required
( ) Moderate English changes required
(x) English language and style are fine/minor spell check required
( ) I don't feel qualified to judge about the English language and style
Yes | Can be improved | Must be improved | Not applicable | |
Does the introduction provide sufficient background and include all relevant references? | (x) | ( ) | ( ) | ( ) |
Is the research design appropriate? | (x) | ( ) | ( ) | ( ) |
Are the methods adequately described? | (x) | ( ) | ( ) | ( ) |
Are the results clearly presented? | (x) | ( ) | ( ) | ( ) |
Are the conclusions supported by the results? | (x) | ( ) | ( ) | ( ) |
Comments and Suggestions for Authors
I support publication.
The authors thank Reviewer#1 for these comments
While I applaud efforts to represent ice nucleation in models like WRF with a more physically based parameterization, some caution is warranted.
For example, the following statement appears (line 54): “They [heterogeneous nucleation processes] all require different water vapor mixing ratios”
While the Meyers et al parameterization for ice nucleation does have a dependence on supersaturation, more recent parameterizations do not. See for example, DeMott et al., 2010. I agree that deposition nucleationshould have a supersaturation dependence, but the sense of the field is that temperature is the more relevant variable for immersion freezing.
We agree that the temperature dependence is more relevant than the supersaturation dependence, even if that latter might not be negligible for some heterogeneous nucleation processes, like deposition nucleation. The sentence has been replaced in the manuscript by :
“They are all strongly dependent on the temperature [DeMott2010], but may also require different mixing ratios of water vapor, in addition, the heterogeneous processes also depend on the type and concentration of the particles.”
The parameterization used here is based on laboratory data from Eastwood et al., 2008 and 2009. It is worth noting that the parameterization is based on a contact angle between ice and a substrate, but that contact angle is not measured. It is derived from the heterogeneous nucleation rate, which is in turn derived from ice onset measurements.
Our sentence was indeed ambiguous. We have rephrased it:
“In the parameterization used in Girard et al. [19], the contact angle is derived from the heterogeneous nucleation rate, which is in turn derived from ice onset measurements of Eastwood et al. [37,38]. Their laboratory studies used uncoated (θ = 12o) and sulfuric acid-coated (θ = 26o) kaolinite particles to derive a parameterization based on a contact angle between ice and a substrate.”
These parameterizations are for deposition nucleation. Is that really the dominant nucleation mechanism for these situations. I would have thought that immersion or condensation freezing would be more appropriate. The text suggests that ice-only clouds are the focus here, but if that is the case, please state it explicitly. (Lines 650 to 652 make be think that all these cases were ice-only, but state that earlier in the paper, if it is the case.)
We hypothesis that the dominant nucleation mechanism on those ice clouds are deposition nucleation on uncoated IN (θ = 12o) and immersion or condensation freezing on sulfuric acid-coated IN (θ = 26o) kaolinite particles. This hypothesis is based on observations: cloud conditions below saturation with respect to water, low temperature at the top of the cloud, and presence of a large occurrence of haze, smoke and dust particles during the ISDAC period (Burton et al. 2012 Atkinson et al. 2011 Jouan et al., 2012, Warneke et al., 2009). It is explicitly written in the description of the simulations and in the conclusions (Points 3 and 4).
The authors confirm that this study focused on ice-only clouds. This has been explicitly mentioned in the abstract and in the conclusions. We have added “ice clouds” in the title of the manuscript.
As I suggest above, I am not asking that the authors revise the nucleation schemes in WRF. But a comment on the issues I’ve raised above is warranted.
Other comments:
lines 45 and 46: “… considering particles conversion.”
I was confused by this sentence. What are the particles converting to? Which particles?
Please revise.
The last part of this sentence was confusing and has been removed.
Line 783: check spelling of “Meyers”. Meyers,
Done
References:
DeMott, P.J., Prenni, A.J., Liu, X., Kreidenweis, S.M., Petters, M.D., Twohy, C.H., Richardson, M.S., Eidhammer, T. and Rogers, D.C., 2010. Predicting global atmospheric ice nuclei distributions and their impacts on climate. Proceedings of the National Academy of Sciences, 107(25), pp.11217-11222.
Submission Date
27 June 2019
Date of this review
05 Jul 2019 19:25:03
Author Response File: 
Author Response.docx
Reviewer 2 Report
I would like to see some detail on the shape of the individual ice crystals. To repeatedly discuss theta and CNT connections and to talk of small and large crystals - but not talk about shape seems odd.
Perhaps a diagram of the shape you think they adopt - unless they are all just irregular?
Also, a couple of further sentences about the IN being heterogeneous or homogeneous please.
Author Response
Montreal, July 12 2019
2
The authors thank Reviewer#2 for this comprehensive review of the paper. We address below each comment individually (in blue color). Line numbers refer to the original manuscript.
Open Review
(x) I would not like to sign my review report
( ) I would like to sign my review report
English language and style
( ) Extensive editing of English language and style required
( ) Moderate English changes required
(x) English language and style are fine/minor spell check required
( ) I don't feel qualified to judge about the English language and style
Yes | Can be improved | Must be improved | Not applicable | |
Does the introduction provide sufficient background and include all relevant references? | (x) | ( ) | ( ) | ( ) |
Is the research design appropriate? | (x) | ( ) | ( ) | ( ) |
Are the methods adequately described? | (x) | ( ) | ( ) | ( ) |
Are the results clearly presented? | (x) | ( ) | ( ) | ( ) |
Are the conclusions supported by the results? | (x) | ( ) | ( ) | ( ) |
Comments and Suggestions for Authors
I would like to see some detail on the shape of the individual ice crystals. To repeatedly discuss theta and CNT connections and to talk of small and large crystals - but not talk about shape seems odd.
Perhaps a diagram of the shape you think they adopt - unless they are all just irregular?
Also, a couple of further sentences about the IN being heterogeneous or homogeneous please.
In this paper, we use a double-moment microphysical scheme to perform simulations at the regional scale. Such models do not actually use a variety of shapes for ice crystals. Even in the ISDAC dataset, the mean ice crystal radius is calculated assuming spherical shape of ice particles and monodisperse size distributions (L224-225). Here, in the model, all ice particles are assumed to be just irregular particles under the form of bullet rosettes. This has been added in the revised manuscript.
Submission Date
27 June 2019
Date of this review
02 Jul 2019 08:24:44
Author Response File: Author Response.docx
Reviewer 3 Report
Keita et al. introduce modifications to the Weather Research and Forecasting model to simulate ice formation in arctic clouds using parameterizations of ice nucleation for polluted and non-polluted aerosol types. Ice formation remains a large uncertainty in climate system, so the manuscript is certainly relevant and worthy of publication in Atmosphere. Broadly, the authors could improve the manuscript by discussing more the broader aspects of their findings. Specifically, they demonstrate that the stochastic approach to nucleation improves the model, even with rather simple parameterizations of heterogeneous nucleation. Discussion of future improvements to parameterizations would be helpful. Below, I provide specific comments that the authors can have the opportunity to consider prior to publication.
- I did not gather the specific motivation for using the dates “1 to 30 April 2008” (line 108). Was it a matter of availability of observational data for this time? Or other reason? Brief statement of why this timeframe was used would be helpful. Unless I missed this, there was no such statement.
- The authors use the parameterizations of Girard et al. (reference 19). As acknowledged by the authors, “In the parameterization used in Girard et al. [19], the contact angle is derived based on laboratory studies of Eastwood et al. [37,38] on uncoated and sulfuric acid-coated kaolinite particles.” (Lines 160-161). It is expected that real atmospheric INP are more varied than coated and uncoated kaolinite alone. Some justification of the validity of the parameterization of Girard et al. should be given. That is, is kaolinite found in abundance in the region simulated?
- In Figures 8b to 15b, why is the mean shown only for either the WRF_np or WRF_p simulation and for DARDAR, and not for all simulations?
- In Summary and Conclusions, a brief discussion of ways in which models could be improved would be helpful. The authors mention variable aerosol concentration and acidity (lines 680-681) but what about parameterizations of heterogeneous nucleation that include a wider range of particle types? In other words, considering the results from the present manuscript and the authors past work (e.g., Keita et al. 2016, ref 15), will improving parameterizations for mixed-composition aerosol be necessary to improve ice nucleation simulations?
Author Response
Montreal, July 12 2019
3
The authors thank Reviewer#3 for this comprehensive review of the paper. We address below each comment individually (in blue color). Line numbers refer to the original manuscript.
Open Review
(x) I would not like to sign my review report
( ) I would like to sign my review report
English language and style
( ) Extensive editing of English language and style required
( ) Moderate English changes required
(x) English language and style are fine/minor spell check required
( ) I don't feel qualified to judge about the English language and style
Yes | Can be improved | Must be improved | Not applicable | |
Does the introduction provide sufficient background and include all relevant references? | (x) | ( ) | ( ) | ( ) |
Is the research design appropriate? | (x) | ( ) | ( ) | ( ) |
Are the methods adequately described? | (x) | ( ) | ( ) | ( ) |
Are the results clearly presented? | ( ) | (x) | ( ) | ( ) |
Are the conclusions supported by the results? | (x) | ( ) | ( ) | ( ) |
Comments and Suggestions for Authors
Keita et al. introduce modifications to the Weather Research and Forecasting model to simulate ice formation in arctic clouds using parameterizations of ice nucleation for polluted and non-polluted aerosol types. Ice formation remains a large uncertainty in climate system, so the manuscript is certainly relevant and worthy of publication in Atmosphere. Broadly, the authors could improve the manuscript by discussing more the broader aspects of their findings. Specifically, they demonstrate that the stochastic approach to nucleation improves the model, even with rather simple parameterizations of heterogeneous nucleation. Discussion of future improvements to parameterizations would be helpful. Below, I provide specific comments that the authors can have the opportunity to consider prior to publication.
- I did not gather the specific motivation for using the dates “1 to 30 April 2008” (line 108). Was it a matter of availability of observational data for this time? Or other reason? Brief statement of why this timeframe was used would be helpful. Unless I missed this, there was no such statement.
To evaluate the improvements of the parameterization developed in the paper, we have used both in situ measurements from the ISDAC airborne campaign and co-localized satellite measurements (CALIPSO/CloudSat). The dates “1 to 30 April 2008” correspond to the period of the flights during the ISDAC campaign. This was explicitly stated in the text lines 115,118, 212 and 216.
- The authors use the parameterizations of Girard et al. (reference 19). As acknowledged by the authors, “In the parameterization used in Girard et al. [19], the contact angle is derived based on laboratory studies of Eastwood et al. [37,38] on uncoated and sulfuric acid-coated kaolinite particles.” (Lines 160-161). It is expected that real atmospheric INP are more varied than coated and uncoated kaolinite alone. Some justification of the validity of the parameterization of Girard et al. should be given. That is, is kaolinite found in abundance in the region simulated?
Kaolinite (silicate mineral) is a common component of desert aerosols (Zimmermann et al. 2008) and dust particles are in general the most abundant INPs in the atmosphere. Our hypothesis is supported by the favorable meteorological situation during the ISDAC period. Field observations have highlighted the presence of a large occurrence of haze, smoke and dust particles during ISDAC period (Burton et al. 2012 Atkinson et al. 2011 Jouan et al., 2012, Warneke et al., 2009).
In the Arctic region, a large fraction of the aerosol particles (including insoluble aerosols such as mineral dust) can be coated with acidic sulfate [Bigg , 1980]. Eastwood et al.2009 results support the idea that anthropogenic emissions of SO2 and NH3 may influence the ice nucleating properties of mineral dust particles by increasing the relative humidity required for ice nucleation. This shift in ice nucleation conditions may influence the frequency and properties of ice clouds.
In the revised manuscript we have added the following text :
“Using a parameterization of ice nucleation on kaolinite particles is supported by the favorable meteorological situation encountered during the ISDAC period. Field observations have highlighted the presence of a large occurrence of haze, smoke and dust particles during ISDAC period [24,40-42]. Besides, in the Arctic region, a large fraction of the aerosol particles (including insoluble aerosols such as mineral dust) can be coated with acidic sulfate [43]. Eastwood et al. [39] results support the idea that anthropogenic emissions of SO2 and NH3 may influence the ice nucleating properties of mineral dust particles by increasing the relative humidity required for ice nucleation. This shift in ice nucleation conditions may influence the frequency and properties of ice clouds.”
- In Figures 8b to 15b, why is the mean shown only for either the WRF_np or WRF_p simulation and for DARDAR, and not for all simulations?
In Figures 8b to 15b, we have followed the convention of Fig. 7. For the sake of clarity, we only represent the modeled average profile showing the best statistical score according to results discussed in Section 4.2 at larger scale. Along F12 and F13 flights, the WRF_np_mean is therefore represented, whereas the WRF_p_mean profile is shown along F21 and F29 flights. It has been mentioned in the text, following the description of Figure 7 (L393-394).
- In Summary and Conclusions, a brief discussion of ways in which models could be improved would be helpful. The authors mention variable aerosol concentration and acidity (lines 680-681) but what about parameterizations of heterogeneous nucleation that include a wider range of particle types? In other words, considering the results from the present manuscript and the authors past work (e.g., Keita et al. 2016, ref 15), will improving parameterizations for mixed-composition aerosol be necessary to improve ice nucleation simulations?
In the conclusions, we have indeed already mentioned that the improved parameterization planned in our following study will take into account the temporal and spatial variation of the aerosol concentrations and their degree of acidity using the proper contact angle. We agree with Reviewer#3 that our parameterization is also limited to the ice nucleation on dust particles only. Further work could consider heterogeneous ice nucleation on other types of particles, e.g. primary biological particles (bacterias, fungal spores and pollens). Taking into account the mixed composition of aerosols to develop specific parameterizations of ice nucleation could improve simulations of ice clouds where aerosol plumes of various types are observed. This has been added in the conclusions.
Submission Date
27 June 2019
Date of this review
07 Jul 2019 20:37:26
Author Response File: Author Response.docx
Reviewer 4 Report
The paper is dedicated to the improvement of the Weather Research and Forecasting (WRF) model by implementing a new parameterization of ice nucleation. The paper represents a valuable piece of work and is certainly interesting for the readership of Atmosphere and other relevant journals. The manuscript, which is clearly written and easy to follow, includes a number of properly designed case studies and comparisons with observations. The key conclusion of the paper-new parameterization works better than earlier ones- is fully supported by calculations.
The paper is of good quality, reports new important results and , thus, deserves publication.
However, a few issues should be addressed before publication can be recommended.
1. The authors claim that they use " advanced ice cloud microphysics" in their simulations. However, all they did in regards to "ice cloud microphysics" is the application a minor version of Classical Nucleation Theory (CNT), which is limited to adjustment of the contact angle. Moreover, they approximate, with no justification, nucleation of binary (sulfuric acid-coated ) ice crystals by unary nucleation with different contact angle. Please, comment.
2. The uncertanties in measurements/observations are mentioned only briefly ( lines 207-209). An extented analysis of experimental uncertanties and their impacts on conclusions made in the paper would be helpfull.
Technical issue
Line 16, please, delete "model"
Author Response
Montreal, July 12 2019
4
The authors thank Reviewer#4 for this comprehensive review of the paper. We address below each comment individually (in blue color). Line numbers refer to the original manuscript.
Open Review
(x) I would not like to sign my review report
( ) I would like to sign my review report
English language and style
( ) Extensive editing of English language and style required
( ) Moderate English changes required
(x) English language and style are fine/minor spell check required
( ) I don't feel qualified to judge about the English language and style
Yes | Can be improved | Must be improved | Not applicable | |
Does the introduction provide sufficient background and include all relevant references? | (x) | ( ) | ( ) | ( ) |
Is the research design appropriate? | (x) | ( ) | ( ) | ( ) |
Are the methods adequately described? | (x) | ( ) | ( ) | ( ) |
Are the results clearly presented? | (x) | ( ) | ( ) | ( ) |
Are the conclusions supported by the results? | (x) | ( ) | ( ) | ( ) |
Comments and Suggestions for Authors
The paper is dedicated to the improvement of the Weather Research and Forecasting (WRF) model by implementing a new parameterization of ice nucleation. The paper represents a valuable piece of work and is certainly interesting for the readership of Atmosphere and other relevant journals. The manuscript, which is clearly written and easy to follow, includes a number of properly designed case studies and comparisons with observations. The key conclusion of the paper-new parameterization works better than earlier ones- is fully supported by calculations.
The paper is of good quality, reports new important results and , thus, deserves publication.
We thank Reviewer#4 for his/her positive comments.
However, a few issues should be addressed before publication can be recommended.
1. The authors claim that they use " advanced ice cloud microphysics" in their simulations. However, all they did in regards to "ice cloud microphysics" is the application a minor version of Classical Nucleation Theory (CNT), which is limited to adjustment of the contact angle. Moreover, they approximate, with no justification, nucleation of binary (sulfuric acid-coated ) ice crystals by unary nucleation with different contact angle. Please, comment.
In this study, we have used the two-moment microphysical scheme of Milbrandt and Yau. In contrast to the original version of the official release of the code, based on the Meyers et al. parameterization ( relying only on supersaturation with respect to ice), we have improved the ice nucleation using a much more physically parameterization. That latter has been developed by Girard et al. and is based on the CNT. This parameterization is built to simulate Arctic ice clouds influenced polluted and non polluted air masses. In the first case, the deposition nucleation occurs on an uncoated IN, and in the second case, the immersion or condensation freezing nucleation occurs on an IN coated with sulfuric acid. This hypothesis is supported by the meteorological situation and observations indicating the presence a large occurrence of haze, smoke and dust particles during the ISDAC period. (Burton et al. 2012 Atkinson et al. 2011 Jouan et al., 2012, Warneke et al., 2009).
In the parameterization used in Girard et al. [19], the contact angle is derived from the heterogeneous nucleation rate, which is in turn derived from ice onset measurements of Eastwood et al. [37,38]. Their laboratory studies used uncoated (θ = 12o) and sulfuric acid-coated (θ = 26o) kaolinite particles to derive a parameterization based on a contact angle between ice and a substrate.
2. The uncertanties in measurements/observations are mentioned only briefly ( lines 207-209). An extented analysis of experimental uncertanties and their impacts on conclusions made in the paper would be helpfull.
We agree with the reviewer that knowing the uncertainties on the measurements is certainly an important aspect of this study. Jouan et al. (2012) have performed a comprehensive study of the ISDAC measurements derived by a variety of instruments and discussed the uncertainties in measurements/observations. They found that the uncertainties on the measurements are the following: ±11% for RHi, ±50% for Ni, ±75% for IWC and ±0.5oC for T. Using the IWC and Ni uncertainties, it can be shown that the uncertainty on Ri is ±97%. More details on uncertainties and biases related to these measurements are described in details in McFarquhar et al. (2011) and Jouan et al. (2012). Concerning DARDAR data, the RMSE (root mean square error) in IWC and Re derived from DARDAR are retrieved according to the methodology of Delanoë and Hogan [44]. Jouan et al. [24] found that within TICs, the normalized RMSE of IWC and Re from DARDAR product ranged from 10 to 30% for IWC and less than 20% for Re.
In the revised manuscript we have added the following text :
“They found that the uncertainties on the measurements are the following: ±11% for RHi, ±50% for Ni, ±75% for IWC and ±0.5oC for T. Using the IWC and Ni uncertainties, it can be shown that the uncertainty on Ri is ±97%.
“The RMSE (root mean square error) in IWC and Re derived from DARDAR are retrieved according to the methodology of Delanoë and Hogan [44]. Jouan et al. [24] found that within TICs, the normalized RMSE of IWC and Re from DARDAR product ranged from 10 to 30% for IWC and less than 20% for Re.”
Technical issue
Line 16, please, delete "model"
Done
Submission Date
27 June 2019
Date of this review
09 Jul 2019 01:08:20
Author Response File: Author Response.docx
