A Parallax Shift Effect Correction Based on Cloud Top Height for FY-4A Lightning Mapping Imager (LMI)
Round 1
Reviewer 1 Report
This paper proposes that the lightning location accuracy of LMI is affected by the cloud top height (CTH), and establishes an ECPC model to correct for this influence. This model uses CTH data from the AGRI on FY-4A to correct the lightning positioning data of LMI. The work compares the corrected data of LMI with data from the BLNET and radar data to evaluate the effectiveness of the correction. The work presented in this paper is interesting, and also practical for the calibration of space-based lightning detection data. Although this paper is valuable in terms of calibration methods, there are still some problems that need to be considered.
1. The figure name is incorrect. (line 346, line 385)
2. The figure 11 and 12 could use a different color scheme.
3. The temporal and spatial thresholds for LMI and BLNET proposed in the paper are 330ms and 16.5km. Due to the significant impact of different spatiotemporal threshold choices on matching results, I wonder how to determine these values? Some description or references are required to explain how to estimate the temporal and spatial thresholds. (line 510)
4. The three levels of "Event", "Group", and "Flash" in the LMI seem to have the same temporal and spatial thresholds matched with BLNET. This seems unreasonable and the spatial and temporal thresholds for matching the "Flash" scale should be larger.
5. What does the "BLNET flash" mentioned in the paper refer to? Flash? Or return stroke? If the "BLNET flash" mentioned in the paper is a flash, a clustering method should be provided. (line410, line421……)
6. I hope the author can further enrich the evaluation of other LMI performance, especially the detection efficiency.
Minor editing of English language required.
Author Response
Please see the attachment
Author Response File: 
Author Response.docx
Reviewer 2 Report
The study concerns the solution of an important problem of the correction of lightning location obtained by the satellite optical sensor. The model of ellipsoid cloud top height parallax correction is proposed. The proposed method is fully described. The comparison with the groun-based observations resulted disparities of order slightly greater than the instrument's spatial resolution. The text is clear, the visual material corresponds well to the context.
The research will be of interest to readers. I recommend the manuscript for publication after correcting some typos.
Some questions:
1) introduction: the authors didn't compare different methods for solving this problem used for data of other geostationary satellites (maybe GLM/GOES-R: https://doi.org/10.1175/JTECHD-19-0100.s1 or Himawari-8: https://doi.org/10.3390/rs12152472). However, this is not necessary for this study: the papers describing methods close to the model used in the present study are cited.
2) why the comparison of data with BLNET data was done only for 1 day? That storm was the most productive for LMI, but in propose to increase the statistics it would be useful to consider some other days with severe storms too (July 13, August 1 and 6, for example).
3) The day of June 28 is interesting - what could be the reason that the LMI Event number exceeded the BLNET flash number?
The typos:
1) line 171: 'T' - is it should be T' as it is on the fig. 1?
2) 199: S_0 - '0' should be subscript according to fig. 3
3) 204: the Fig. 1 doesn't contain marks of X, Y, Z axes. Maybe it should fig. 3?
4) 220: what is '-f'? Not found in text
5) 265: dot should be replaced by comma or space...
6) 345: Figure 1 - replace by Figure 8
7) 385: Figure 2 - replace by Figure 9
8) Page 12: in the axis label the 'g' is missed: 'Heiht' -> 'Height'
Author Response
Please see the attachment
Author Response File: Author Response.docx
Reviewer 3 Report
[Title]
A Parallax Shift Effect Correction Based on Cloud Top Height for FY-4A Lightning Mapping Imager (LMI)
[Summary]
The authors reconstructed ECPC model in order to evaluate the parallaxshift of F-4A LMI. Then, they applied to the model to LMI data and after the correction, the LMI data agree well with ground-based BLNET data.
[Broad Comment]
This kind study becomes more and more important to evaluate lightning detection from geostationary satellites.
I almost agree with the author's discussion.
If I am allowed to wish, the case used for comparison between LMI and BLNET may be changed to other case(s) with more plentiful lightning data, if you have data of extra case(s).
Additionally, there are some minor mistakes in the manuscript, so please check and revise carefully.
[Specific Coments]
Line 313: (Lightning Mapping Imager)
The words appears Line 52 already and are not necessary here.
Line 346: Figure. 1
-> Figure. 8
Line 385: Figure. 2
-> Figure. 9
There are no link (or how to access) to the data used for the analysis of this study. Please recheck the data policy of the journal.
Sorry for delaying my review. I hope my review report helps the revision process.
Author Response
Please see the attachment
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
There are still some problems with the article, hopefully some revisions can be made and then published.
1. The authors explain that the data provided by BLNET is flash, but this does not seem to be the smallest level of discharge events that BLNET can detect. Why not use the same level of discharge event for comparison? For example, Figure 11(a) and (b) could be more intuitive if compared with BLNET strokes (discharge events).
2. Figure 5 shows a comparison between LMI 'event' and BLNET flash, why are different levels of discharge events used for the comparison? The 'event' is pixel units of the LMI CCD array illuminated by light pulses of flash, while the BLNET flash is a single flash process. It should be a comparison between LMI 'flash' and BLNET lightning.
3. As mentioned in the paper, Figure 5 gives a rough indication of the relative detection efficiency between LMI and BLNET, and it is hoped that a value for this relative detection efficiency can be given. Similar to the previous comment, 'event' is compared with BLNET flash for relative detection efficiency, which can overestimate the detection efficiency of LMI. (line 295-297 in Revised version (marked) or line 293-295 in Revised version (clean))
4. A common problem in the paper is that the comparison of the two systems, LMI and BLNET, was not performed at the same level of lightning discharge events.
Minor editing of English language required
Author Response
Please see the attachment
Author Response File: Author Response.docx
