Review Reports - Case Study of a Retrieval Method of 3D Proxy Reflectivity from FY-4A Lightning Data and Its Impact on the Assimilation and Forecasting for Severe Rainfall Storms

Round 1

Reviewer 1 Report

Review of “A Retrieval Method of 3D Proxy Reflectivity from FY-4A Lightning Data and Its Impact on the Assimilation and Forecasting for a Severe Rainfall Case”

The manuscript describes a method for assimilating geostationary lightning data from the Lightning Mapping Imager into a numerical weather prediction model by first converting the lightning data into proxy radar reflectivity. The dataset and methods are well described. My main comment is that only one case study storm is analyzed, which had the largest precipitation intensity in the past seven year for the Beijing region. One case study, especially an outlier storm in terms of precipitation, is insufficient to validate a new data assimilation method.

Main Comment:

Only one case study storm presented. I see two possible paths for the authors:

Publish the case study result, but make clear in the title, abstract, and main text that these results are only for one rather extraordinary storm and that the method is not validated. They could frame the method as an example of how to use lightning for data assimilation, rather than claiming that a new, validated method has been developed. Title could be something like “Case Study of a Retrieval Method of 3D Proxy Reflectivity from FY-4A Lightning Data and Its Impact on the Assimilation and Forecasting for a Severe Rainfall Storm”.
Conduct similar analyses for more storm cases with varying precipitation intensity. Publish these results as a new method with preliminary validation.

Other comments:

Figure 5: Does log10 give the best fit to the lightning/radar data? To me, a single linear fit would likely work quite well. Something similar to the blue non-linear fit but a straight line. The authors should provide fit parameters to argue that log10 works best.

Section 2, line 98, ADTD stands for Advanced Time of Arrival and Direction, not Active Divectory Topology Diagrammer.

Grammar and style:

First sentence of Intro: “systems” seems like the wrong here. I think the authors simply want to state that there exists in cloud and cloud to ground lightning and LMI can detect both of these.

Second sentence of Intro: replace “reflect” with “identify”

Intro, line 53: Replace "forecast" with "systems" and change "location" to "locations". “…the locations of the convective systems were improved…”

Figure 11 needs axes labels

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

To the Authors:

The paper is well written, but some parts need further clarification. I have indicated them below.

Main Points:

1) The methods used and explanations concerning converting lightning to radar proxy data need clarification, more details (improvement).

2) The results are not tested for significance.

3) The neighborhood radius is really quite small, and the authors should check that the results "hold-up' when the radius is increased incrementally, perhaps to 48 km, by 12 km increments, for example.

Lines 140-141: "Last, the maximum FY-4A proxy reflectivity is expanded to 3D FY-4A proxy reflectivity using the empirical profiles of reflectivity (Figure 3)."

--> It is not really a 3-D expansion, but a 1-D expansion within the vertical grid-column.

Lines 153-155:"The maximum FY-4A lightning density and the maximum radar reflectivity can reach more than 25 strikes and 55dBZ at all the moments."

--> I am not sure what the authors mean by "at all the moments." Also, do they mean observed (for this case study)?

Lines 177-179: "The mean and median of maximum radar reflectivity corresponding to the lightning density are taken to determine the logarithmic curve-fitting between FY-4A lightning density and the maximum radar reflectivity,"

--> I don't see how the mean and median are used in this equation or how/why they are displayed in Fig. 5

--> Also, I would first discuss Fig. 5 before giving the equation on line 180.

Lines 226:228: "The real-time radar reflectivity at each analysis time are used in this section. The height from ground to 14 km is divided into 200 m vertical layers. The radar reflectivity of each layer are averaged for each threshold."

--> Taken together, these sentences are confusing to read and non-sequiturs.

--> A much better explanation must be given for the interpolation scheme.

Lines 237-248: The authors claim that the profiles are "similar." Yet, they write about under- and over-estimations. How do they define similar here, and are they really similar if there are under- and -over-estimations? The authors need to be much clearer in how they describe these curves. One could ask: what is the sensitivity of the results to errors in the estimation of these curves?

Line 260-264: Are confusing. How are initial boundary conditions from the ECMWF model and the recycled boundary conditions blended. This part needs to be a lot clearer.

Line 296: The influence radius used in this study is 5 km..."

--> This is tiny and the results should be checked using a larger neighborhood radius, perhaps up to 48 km. Otherwise, the differences might just be luck. The authors might consider checking their results for significance.

Lines 346-348: "Cycling assimilation of reflectivity adjusts the rain water mixing ratio below 4km and the graupel mixing ratio above 4km (Figure 13a, d, c, f), and the analysis increments of water vapor mixing ratio in FLASH_RF_RV and RADAR_RF_RV are similar."

--> The last phrase does not follow from the previous two.

Lines 369-372: "After cycling assimilation, lightning and radar data are better assimilated into the NWP models, the effect of lightning and radar data in hydrometeors delivers to extended to temperature and wind, which makes the thermal and dynamic analysis more reasonable, and results in the increment of CAPEs."

--> How does cycling assimilation extend to affect temperature and wind? More details are needed. What is meant by: "extended to."

Lines 399-401: Please see: https://journals.ametsoc.org/doi/full/10.1175/WAF-D-13-00028.1

Lines 403-404: However, as FY-4A is a geostationary satellite for test, it is necessary to conduct a more detailed assessment of the FY-4A LMI lightning data [30]."

--> I don't see the connection between the two parts of the sentence.

Lines 428-429:
that assimilating 3D FY-4A proxy reflectivity achieves a resembling effect with assimilating observed reflectivity"

--> The wording: "resembling effect" is not precise.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

Review of the paper:

A Retrieval Method of 3D Proxy Reflectivity from FY-4A Lightning Data and Its Impact on the Assimilation and Forecasting for a Severe Rainfall Case

By:

Yaodeng Chen, Zheng Yu, Wei Han , Jing He, and Min Chen

General comment

This paper presents a method to assimilate the lightning density observed by the LMI (Lightning Mapping Imager) onboard the FY-4 satellite in the RMAPS-ST (rapid update multi-scale observation data assimilation analysis and short-term prediction system, which is based on WRF and WRFDA). More specifically, lightning density is converted into reflectivity and the latter quantity is assimilated.

The method is tuned using seven cases of severe weather over China and then applied to one of them.

In general, the paper presents an interesting subject and a methodology to assimilate local data (lightning) in remote areas where no radar data are available. There are, however, major flaws in the current form of the paper, and additional verification of the methodology are necessary. They are reported in the major point below.

Major points

The introduction doesn’t present well the past studies on the subject. In addition to missing some important studies on the subject with different models than WRF, there is a lack of physical interpretation of the results obtained in previous studies (using or not using the convective precipitation parameterization schemes for example, using water vapour pseudo-profiles etc).

The main problem with the paper is that the method is used to assimilate lightning with the purpose to use it when radar data are not available (for example in remote regions or over the sea). Nevertheless, the performance of the method is evaluated in the Beijing and Tianjin areas, where radar coverage is good. The comparison proposed in the study, i.e. comparing the results of radar data assimilation with those of lightning data assimilation (after converting lightning density to reflectivity) is certainly useful to understand that the methodology has potential, nevertheless a case when radar data are not available is necessary to complete the study.

The case study analyzed in the paper is also used to set the relationship between the lightning density and radar maximum reflectivity. In this way, the application of the method to the analyzed case is not completely independent, from a statistical point of view, because the same case study was used to set-up the relationship used for lightning data assimilation. This must be highlighted into the paper.

The English of the paper need a revision. Some sentences are difficult to understand. I put sticky notes in the pdf of the paper attached to this review to show some of them, nevertheless a general revision of the language is necessary.

Minor points

There are few minor points that the author must consider. There are sticky notes in the paper with these points. In particular, consider the following:

The Discussion (Section 6) is not a critical revision of the results obtained in the paper. If maintained in this form, it can be included in the Conclusions.

Section 3.2.2 is unsatisfactory for the following two points:

a) to what I can understand the profile used in GSI varies from 20-25 to 45-50 dBZ ranges, while in this paper different ranges are used without justification.
b) It is not clear how the profiles of Figure 7 are obtained (from the observations of the seven convective cases over China?). In Section 5 you say that those cases are used to establish the relationship between maximum radar reflectivity and lightning and not to extend in the vertical direction the reflectivity profile. Explain better.

Figures

Figure 1: I can’t understand why the hourly precipitation are shown while lightning are presented accumulated over 15 minutes.

Figure 5: I cannot understand why you use at the same time the mean and the median of maximum radar reflectivity. Explain.