Assimilation and Inversion of Ionospheric Electron Density Data Using Lightning Whistlers
Round 1
Reviewer 1 Report
Comments on "Assimilation and inversion of ionospheric electron density data using lightning whistlers" by Tian Xiang, Moran Liu, Shimin He and Chen Zhou to Remote sensing Journal (2023)
Comments:
In this paper, authors have studied the time-frequency information from the dispersion of lightning whistlers to investigate the electron density profiles along with actual electron density profiles from ionosonde and IRI model and correct the electron density profile of the ionosphere by assimilation of these electron density obtained from lightning whistlers. The Kalman filtering is adopted to assimilate the electron density distribution along the propagation path of lightning whistlers. The results suggest that estimation of electron density profiles has been significantly improved when the data has been assimilated using Kalman filtering. Authors suggest that the electron density after assimilation is in good agreement with the true value. Based on this, authors have studied the influence of the frequency difference on the assimilation inversion. Further it is found that when the frequency difference between frequency points is less than 1 kHz, its effect is worse. Based on my reading of the manuscript, I can say that manuscript looks good and reading is smooth and the manuscript may be suitable for publication. However, I have following comments/suggestions and please respond to following major comments.
Major comments:
1. Figure-5 what are the bottom and top heights. Are you extracting the D region density or total ionosphere? Whether data assimilation is done only for lower ionosphere or from 80-180 km ionosphere? Please clarify.
2. Authors are mentioning about Gaussian distribution in vertical direction. But what about horizontal direction? Similarly, why 100 m separation is assumed?
3. Though the electron density profiles are comparable to the observed values, it has to be kept in mind that these electron density profiles are very smooth. Rocket observations of electron density profiles look very jagged? Why your observations are looking so smooth. Please clarify.
4. I believe that authors are mentioning about only quiet time electron density profiles and their assimilation. But how electron density profiles behave when there is a geomagnetic storm or solar flares? Please clarify.
5. Please provide diurnal variation of electron density profiles before data correction, actual profiles and corrected density profiles. Similarly, show such variations for different seasons. This will provide complete picture.
6. It is known that ground VLF receivers are also used to estimate electron density profiles. Whether ground and space based observations can be correlated?
7. Why height information can’t be retried from time-frequency information. It is possible. Please clarify.
8. Why only Gauss-Markov algorithm has been tested? Why not other methods?
9. Why assimilation becomes worse when the frequency difference between frequency points is less than 1 kHz? Please clarify both what is this frequency difference and its difference in assimilation.
10. Height correlation distance is mentioned as 30 km? But what are their boundaries.
11. Other comments are highlighted in the manuscript. Please go through the manuscript for few minor corrections.
Author Response
Please see the attachment
Author Response File: 
Author Response.docx
Reviewer 2 Report
The manuscript entitled “Assimilation and inversion of ionospheric electron density data using lightning whistlers” by Xiang et al. [Manuscript ID: remotesensing-2401505] presents a new assimilation algorithm to obtain the ionospheric electron density profile using the time-frequency dispersion spectrum of lightning whistlers observed by satellite.
The authors proposed own inversion method to calculate the ionospheric electron density on the basis of the time-frequency information in the dispersion spectrum of lightning whistlers received by the ZH-1 satellite (or The China Seismo-Electromagnetic Satellite, CSES) as the observed value, whereas the International Reference Ionosphere (IRI) model was used as the background model. Then they adopted the Kalman filtering algorithm to assimilate the observed and background values of electron density along the propagation path of lightning whistlers and compared obtained results with the electron density profile calculated by a data from digisonde stations (for example, the Wuhan Zuoling digisonde station) as the true values for the effectiveness evaluation. The authors got good agreement between the ionospheric electron density after assimilation with the true value, and showed effective implementation of assimilation algorithm to correct the background value using observed values.
The topic of the study is within the scope of Remote Sensing, and the used data and methods are suitable to investigate the research question. However, I have to note that the section “Reference” of the manuscript was not prepare in an appropriate way.
1) There a lot of misprints and additional information. It seems that the section “Reference” of the manuscript was not proofread by the authors before submission.
2) There are a lot of papers published not in English that cannot be checked and read by readers, but then the authors did not mention recent article on data assimilation in space weather study, detection algorithms for lightning whistlers recorded by satellite and development of ground-based ELF/VLF receiver system for similar task.
3) For the convenience of readers, please indicate a DOI identifier for papers published in journals and proceedings.
4) This manuscript would benefit from a close reading, there are many mistakes in paper citing that take away from the clarity of the argument.
- Please correct a citation in the text of the manuscript.
Line 101: Bayupati [20] --> Bayupati et al. [20]
Line 107: Oike [21] --> Oike et al. [21]
Line 110: Zahlava [22] --> Zahlava et al. [22]
Line 112: Putri and Kasahara [23] --> Putri et al. [23]
- Please add a reference if it needs to be done.
Line 344: Santolik (2009)
5) I would recommend to clarify the following key word.
Line 19: dispersion --> dispersion spectrum (or your variant)
6) Figure 3 is not very clear, namely a small font of numbers and letters.
7) Text description of following figure is absent.
page 10: Figure 6. Electric field waveform.
I would recommend to the authors to make moderate editing of English language.
But big obstacle of the manuscript is design of the section "Reference",
which does not meet the requirements of Remote Sensing.
Author Response
Please see the attachment
Author Response File: Author Response.docx
Reviewer 3 Report
This is a review of the paper by T. Xiang et al. entitled "Assimilation and inversion of ionospheric electron density data using lightning whistler" submitted to the Remote Sensing journal.
The authors of the paper show the interest of using lightning whistlers to measure from space the ionospheric density profile from assimilation techniques. The advantage of making lightning measurements from space is twofold: lightning flahes are numerous, present everywhere (between +/- 60° of latitude) and all the time (day and night) and one can take into account all the ionosphere up to the altitude of the satellite whereas measurements made from the ground with ionosondes only probe up to a certain altitude (maximum of the F2 region).
In the introduction the authors recall what has been done so far for the assimilation of ionospheric data and show the specific interest of whistlers. In a second part, they detail the measurements used in this article and in a third part they present the assimilation algorithm used. In part 4, they present their results and their uncertainty. Finally in a 5th part, they discuss the impact of the altitude of the maximum of the F2 region which cannot be taken into account. Finally they conclude.
There remain in the article some approximations and, in my opinion, also some errors, which will be necessary to correct. Therefore, I propose a major revision.
Specific comments:
Lightning whistler paragraph 1.2: Note that an ionogram from the ground was also done with lightning. Cite: T Farges, and E Blanc, Lightning and TLE electric fields and their impact on the ionosphere, Comptes Rendus Physique, Volume 12, Issue 2, 171-179, 2011, https://doi.org/10.1016/j.crhy.2011.01.013.
Line 93: put spacecraft instead of vehicles
Line 100 and following: replace scholars by searchers
Line 110 to 112: quote the 2 satellites used in this study (as for the other papers quoted) DEMETER and Van Allen Probes
Line 110: you can add the following article related to DEMETER measurements : Santolik, O., M. Parrot, U. S. Inan, D. Buresova, D. A. Gurnett, and J. Chum (2009), Propagation of unducted whistlersfrom their source lightning: A case study,J. Geophys. Res.,114, A03212, doi:10.1029/2008JA013776
Line 128-130: refer to the paper that used a similar method to assimilate ELF measurements of lightning: Coïsson, P., Hulot, G., Vigneron, P., Deram, P., Léger, J. M., & Jager, T. (2019, January). 0+ whistlers in the ELF band recorded by Swarm satellites used to reconstruct the ionosphere below the satellite height. In Geophysical Research Abstracts (Vol. 21).
Move lines 166-167 to the end of the paragraph to separate the satellite description from the results.
Lines 181 and 182: this has already been said in line 90.
Caption of Figure 2: better describe what is in the figure. What is f1, f2, t1, t2? What do the colors correspond to? What happens around 20 kHz with a continuous horizontal line?
Figure 3: the red curve is not described. Why stop the graph at the maximum of the F2 region when the IRI model goes higher and you use it further? Note on the curve the E, F1, and F2 regions that are discussed on line 198.
Line 210: tell what SAO is and what it means?
Line 208: with an (vertical) ionogram you cannot go beyond the height of the maximum of the F2 region. Say this and take it into account later in the analysis of the results.
Line 222 and following: replace grid by mesh when we are going to talk about a calculation mesh.
Figure 5: is the propagation model in 3D or 2D?
Line 238: replace vital by crucial.
Lines 245 to 247: this is especially true if you assume direct propagation (driving) and not propagation from the opposite hemisphere mode. We can take here to illustrate this the reference of the following papers:
Lefeuvre, F., R. Marshall, J. L. Pinc ̧on, U. S. Inan, D. Lagoutte, M. Parrot, and J. J. Berthelier (2009), On remote sensing oftransient luminous events' parent lightning discharges by ELF/VLF wave measurements on board a satellite, J. Geophys. Res. 114, A09303, doi:10.1029/2009JA014154.
Kang, N., & Bortnik, J. (2022). Structure of energy precipitation induced by superbolt-lightning generated whistler waves. Geophysical Research Letters, 49, e2022GL097770. https://doi.org/10.1029/2022GL097770
Line 251: write Appleton-Hartree.
Line 257: introduce q and m.
Equation 2, line 264: I don't understand the simplification you use to go from X / (1-YL) to X/(1-Y). Please explain?
Line 262: we can ignore collisions when h > 90 km i.e. when γ the << w. Doesn't this have any implication on your calculations? Say why.
Line 285: what is Δt?
Line 293: don't you mean equation (7) instead?
Lines 299 and 300: are we really talking about the same H here? In one case it is presented as an operator in the other it is the vertical height of the lightning.
Lines 328 to 330: is there a reference to justify this?
Figure 6 is not cited in the text.
Line 347: how do you verify that it is really related to a lightning flash besides the fact that it is a whistler? A whistler can be linked to other emissions coming from the magnetosphere (hiss, chorus, ...)? Didn't you try to look for a coincidence with lightning detected on the ground?
Caption of Figure 6: add details around the figure as measured by the ZH-1 satellite, the place and time of the observation, etc...
Caption of Figure 7: add that this spectrogram is relative to the signal measured in Figure 6.
Line 352: neither t1 nor t2, neither f1 nor f2 were presented in section 2.
Caption of Figure 8: Complete the caption. How is it that an ionogram can give information beyond the maximum height of the F2 region? Are you using a model? If so, please tell us.
Line 371: I don't really understand. In fact, why not calculate the electron density deviation by altitude range? For example, here Ne = 4 m-3 corresponds as well to an altitude below 200 km as to more than 400 km what is the sense of mixing the deviation for these two altitudes?
Figure 9 could also be represented in log.
Line 425 discussion: the influence of the electrical signal strength, or in other words its SNR, has not been evaluated. What is it? What is also the importance of the impact of several flashes arriving at the same time on the satellite?
Lines 426 to 433: I had not understood before arriving here that the choice of frequencies was taken at random. Better to insist in the methodology.
Line 446 447: the width of the frequency gap below has an undeniable impact on the evaluation of the electron density, when it is less than 1 kHz. Nevertheless, is this impact the same when the chosen frequencies are low (near 1 kHz) or high (near 10 kHz)?
Figures 13 and 15: I count 8 points on the curve whereas I should have 4 if I read correctly what is said on lines 434 or 439. Explain this difference.
Author Response
Please see the attachment
Author Response File: Author Response.docx
Round 2
Reviewer 2 Report
The manuscript entitled “Assimilation and inversion of ionospheric electron density data using lightning whistlers” by Xiang et al. [Manuscript ID: remotesensing-2401505] presents a new assimilation algorithm to obtain the ionospheric electron density profile using the time-frequency dispersion spectrum of lightning whistlers observed by satellite.
The authors proposed own inversion method to calculate the ionospheric electron density on the basis of the time-frequency information in the dispersion spectrum of lightning whistlers received by the ZH-1 satellite (or The China Seismo-Electromagnetic Satellite, CSES) as the observed value, whereas the International Reference Ionosphere (IRI) model was used as the background model. Then they adopted the Kalman filtering algorithm to assimilate the observed and background values of electron density along the propagation path of lightning whistlers and compared obtained results with the electron density profile calculated by a data from digisonde stations (for example, the Wuhan Zuoling digisonde station) as the true values for the effectiveness evaluation. The authors got good agreement between the ionospheric electron density after assimilation with the true value, and showed effective implementation of assimilation algorithm to correct the background value using observed values.
The topic of the study is within the scope of Remote Sensing, and the used data and methods are suitable to investigate the research question.
The authors provided detailed reply due to reviewer’s comments and added some papers on data assimilation in space weather study (Bust 2020, Ercha Aa 2021 and Qiao 2022), detection algorithms for lightning whistlers recorded by satellite (Dharma 2014, Ahmad 2019 and Konan 2020) and development of ground-based ELF/VLF receiver system for similar task (Chen 2016 and Cohen 2009) in the text of the manuscript. In the second revised manuscript they have incorporated reviewer’s comments appropriately.
I therefore recommend the manuscript to be accepted for publication in MDPI Remote Sensing after correction of minor technical errors.
Minor technical errors
1) For the convenience of readers, please put in order some citations for references 37-40 which are mentioned ahead of their order and ref.35 which is mentioned after of their order in the text of the manuscript.
2) There are additional symbols for references 5, 10, 15, 16, 17, 18, 20, 21, 22, 25, 26, 34 in the section “Reference” of the manuscript.
3) typo:
Lines 559-560: Reaserch --> Research
4) wrong DOI:
Lines 614-615, ref. 23: 10.1080/00966665.1952.10467555
Comments for author File: 
Comments.pdf
Quality of English Language of the manuscript is improved. There are some minor technical errors.
Author Response
Please see the attachment.
Author Response File: Author Response.docx
Reviewer 3 Report
Thank you for your clear and precise answers. The corrections made to the text make the description of the method and results more comprehensible. Nevertheless, I regret that some very interesting answers have not been integrated into the text. I list them below. In addition, there are still some misunderstandings on figures 13 and 15 that need to be clarified. I therefore recommend a minor revision.
To be added in the text:
- Explicit description to be added in the caption of what the red curve in Figure 3 is.
- Add in the caption to Figure 5 that the model is 2D.
- Add after equation (2) the justification for saying that Y = YL. “In the whistler theory, the approximate model of quasi-longitudinal propagation is satisfied, the partial energy of the VLF/ELF electromagnetic wave radiated by lightning in nature propagates along the geomagnetic field line in the ionosphere in the form of a whistler. So j≈0 and YL≈Y.”
- Your justification for ignoring collisions is very good.
“The collision affects the strength of the signal, but has little effect on the propagation time of the lightning whistlers. Based on the time, geography and corresponding ionospheric state information in Figure 7, the propagation time of the lightning whistler with a frequency of 5.5 kHz from the bottom of the ionosphere to the altitude of the satellite is calculated with and without collision respectively. The time difference is 6.8 us, which is far less than the millisecond order of the propagation time in Figure 7. Therefore, it is reasonable to ignore the collision. The formula of the collision frequency is nu = 1.82 x 1011 exp(-0.15h) ( the references are as follows), where h represents the height and its unit is km.”
Add it on line 308 when you mention it.
- The definition of Δt in equation line 331 is not written in the text. I'm fine with your definition in the answer: “Δt is the propagation time difference of two whistler signals with different frequencies in a mesh.”
- You can add in the caption to Figure 6 that this whistler corresponds to a negative ground cloud flash (-CG) measured by the Institute of Electrical Engineering of the Chinese Academy of Science at 06 :43 :50.521 UT of -24.1 kA located at 30.536°N and 112.316°E i.e. ~300 km from the satellite's ground position .
- You haven't added a reference to Figure 6 in the legend to Figure 7. Please, do it.
- You can add to the discussion part the (very interesting) answer on the influence of SNR and incidentally of the choice of frequency range that you give in the answers.
“We determine the corresponding frequency according to the position with the maximum power spectral density at a specific time point. Therefore, the accuracy of frequency selection will be affected where the electric field intensity of lightning whistlers is weak, that is, where the SNR is low. Inaccurate frequency selection will have a negative impact on the assimilation results. Multiple dispersion spectrum may overlap when several lightning arrive simultaneously, and the accuracy of frequency selection will also be affected. Most of the energy of lightning whistlers is concentrated in the range of 500-5500Hz (Fiser et al., 2010). Based on extensive observations of lightning whistlers received by the ZH-1 satellite electric field detector, we found that most of the energy of lightning whistlers received by the ZH-1 satellite is concentrated in the range of 1-10kHz. In order to reduce the error of the frequency selection, we selected non overlapping dispersion spectrum of lightning whistlers, and then the time-frequency points were selected as the observed values in the range of 1~10kH with large electric field intensity.”
Reference:
Fiser J, Chum J, Diendorfer G, et al. Whistler intensities above thunderstorms[J]. Annales Geophysicae, 2010, 28(1), doi: 10.5194/angeo-28-37-2010.
Note that in the legend to Figure 8, there's an extra "t" at the end of the sentence.
Concerning figures 13 and 15, I still don't understand despite your answer.
I find 9 points in figure 13 with x-axis: 0.4, 0.8, 1.5, 3.1, 4.7, 3.9, 4.7, 5.5 and 5.9 kHz whereas in the text and in figure 12 you only indicate 0.4, 0.8, 3.9, 5.5 kHz. Where do these 5 additional frequency shifts come from?
Similarly, on figure 15, I find 8 points with x-axis: 0.4, 1.2, 2.0, 2.8, 3.5, 5.1 and 5.9 kHz, whereas in the text and on figure 13 you only indicate 0.4, 0.8, 2.8, 5.5 kHz. So I can't find the 0.8 kHz and 5.5 kHz shifts? Where do these 7 additional frequency shifts come from?
Author Response
Please see the attachment.
Author Response File: Author Response.docx
