Relating Convection to GCM Grid-Scale Fields Using Cloud-Resolving Model Simulation of a Squall Line Observed during MC3E Field Experiment

Round 1

Reviewer 1 Report

The present-day GCMs still have a lot of discrepancies and bias when compared with the observation, especially over convective regions. This study uses a WRF simulation to examine the relationships between convective precipitation and coarse-grained variables averaged over a range of subdomain sizes equivalent to one single grid in GCMs. The goal is to determine the relationship between convection and three large-scale variables, including 500hPa vertical velocity, moisture convergence, and CAPE.  Many interesting results are obtained including their correction and time lead/lag, which should provide useful implications for improving GCM convective parameterizations. Considering its scientific merits, I believe this study should match the broad scope of the journal Atmosphere.  

I have only a few comments concerning the presentation of this work. Meanwhile, I have several minor questions that got me confused, which I think the authors can readily resolve.  Overall, I would suggest the authors undertake a minor revision before it is considered to be published. 

 

My comments:

1) Line 17-20: you mentioned "The results show that... but the correlation decreases with decreasing subdomain size". Which figure do you use to conclude this? By comparing Figure 6 and 8, I think the correlation coefficient with W500hPa does decrease, but those with dCAPE and moist convergence increase a lot! Please clarify this in the context. Since you have all the data, I would suggest the authors add a new table with 7 rows and 3 columns and list all the correlation coefficients. The row of this table corresponds to different subdomain size from 256 km to 2 km, and the column corresponds to these three large-scale variables. 

2) Line 25-27: "...and the lead time corresponding...with convection". This sentence sounds confusing. Please rewrite it. 

3) Figure 1 caption: I got confused about the subdomain indicated by the dotted rectangle when I first read here. I am wondering its meaning until I read the beginning of Section 3. Please rewrite this caption to clarify that the WRF solution in this subdomain will be used for analysis in the following sections.

4) Line 169-171: as mentioned by the authors, the simulation solution has a big bias in simulating vertical velocity. This will reduce the credibility of your following analysis about 500hPa W and mass flux. Please clarify or admit this potential drawback in the context. 

5) Line 201: please add the sentence that you use mass continuity equation to simplify Eq. 2. 

6) Line 211: please explain the variable \bar{T}_{v}(t) in the context. It should be environmental virtual temperature. 

7) Line 241-245: another interesting scenario you may consider is excluding data points with small or no precipitation. In that way,  you only use the data points at the right-hand side above zero precipitation to calculate the correlation coefficients. 

8) Figure 6 & 8: please re-order the panels to be the same as Figure 4 & 5. Also, why is there a clear separation between green dots and red dots in Figure 6? Normally I would expect a cluster of data dots.  Is it because the period of data used here is too short (only 12 hrs) here? 

9) Figure 7: please add a legend to explain curves in different colors. 

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

In this study an attempt is made to correlate convection with GCM large scale parameters in a case of squall line using the WRF model. I found the paper very interesting providing a fruitful input in the mesoscale meteorology and nowcasting of squall lines. However, I have some comments and queries.

Lines 89-93: The objectives are not very clear compared with the previous part of the introduction and the abstract. These should be reformulated Lines 138-149: The synoptic scale analysis is described briefly referred to the paper by Tao et al (2013). However, I think that the authors should present one or more related figures to facilitate the reader. The figures could be reproduced by the paper Tao et al (2013) or other sources with relevant reference Line 152: I do not agree with the statement “the model captures main observed features of the squall line, such as its intensity and distribution at 8 km and 2.5 km” based on the two patterns of Figure 2. The resemblance of the two patterns is very poor. The authors overestimate the model performance and should use a more realistic statement. Lines 167-168: Similarly with previous comment. In Figure 3, the comparison is not correct. The authors overestimate the performance of WRF. The WRF represents ascending motion from the lower levels after 3 h while there are observed descending motion while fails in representing the ascending motion at the lower levels in the following hours. It simulated correctly the ascent at the upper levels but it fails in capturing the hour of the peak values. Lines 190-193: How this definition is verified? A reference is required. Line 219: The abbreviation dCAPE has not defined before in the text and should be clarified. Lines 226-227: I do not agree with the statement “The precipitation is well correlated with dCAPE and 500 hPa vertical velocity” when the correlation coefficient is 067 and 0.60, respectively. The correlation is fairly good. Line 292: why the maximum correlation appears at these specific domains? Physical explanation is needed. Furthermore, the authors do not comment on the reason that the correlation is very low at domains with smaller subdomain sizes. Line 310: Similarly to previous comment. The correlation is very low for smaller size subdomains. Why??? Line 326: “… peaks of large-scale variables increase with decreasing subdomain size”. An explanation is needed. Conclusions: The authors should comment if and how their results apply only in this specific case in this specific midlatitude area and this specific time of year.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

The authors made a large effort to incorporate the comments of both reviewers and the paper has been greatly improved. I suggest that it can be published in its present form.