Near-surface structure of the Sodankylä area in Finland, obtained by passive seismic interferometry

Afonin, Nikita; Kozlovskaya, Elena; Heinonen, Suvi; Buske, Stefan

doi:https://doi.org/10.5194/se-12-1563-2021

Articles | Volume 12, issue 7

https://doi.org/10.5194/se-12-1563-2021

© Author(s) 2021. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/se-12-1563-2021

© Author(s) 2021. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 12, issue 7

Research article

|

07 Jul 2021

Research article |

| 07 Jul 2021

Near-surface structure of the Sodankylä area in Finland, obtained by passive seismic interferometry

Nikita Afonin, Elena Kozlovskaya, Suvi Heinonen, and Stefan Buske

Download

Final revised paper (published on 07 Jul 2021)
Supplement to the final revised paper
Preprint (discussion started on 05 Oct 2020)

Interactive discussion

Status: closed

AC: Author comment | RC: Referee comment | SC: Short comment | EC: Editor comment

- Printer-friendly version

- Supplement

RC1: 'suggested comments', Yunhuo Zhang, 17 Nov 2020
- AC1: 'Reply on RC1', Nikita Afonin, 14 Jan 2021
RC2: 'Review of "Near surface structure of Sodankylä area in Finland, advanced method of passive seismic interferometry" by Afonin et al.', Anonymous Referee #2, 27 Nov 2020
- AC2: 'Reply on RC2', Nikita Afonin, 14 Jan 2021

Peer-review completion

AR: Author's response | RR: Referee report | ED: Editor decision

AR by Nikita Afonin on behalf of the Authors (14 Jan 2021)

ED: Referee Nomination & Report Request started (20 Jan 2021) by Ulrike Werban

RR by Anonymous Referee #2 (19 Feb 2021)

RR by Meysam Rezaeifar (10 Mar 2021)

Suggestions for revision or reasons for rejection

Review report on « Near surface structure of Sodankylä area in Finland, obtained by passive seismic interferometry " by Nikita Afonin et. al.

The deduced results are interesting and can be considered for publication. provided some areas can be improved. The authors applied seismic interferometry to passive seismic data to retrieving surface-wave and to image the near subsurface structures related to mineral exploration. They used SNRS as part of the processing procedure to retrieve the green function. However, it is not clear to me how and why they used this algorithm. They also performed synthetic tests which the goal is unclear to me. Although the resulted 2D sections of shear-wave velocity models from inverting of the dispersion curves and the geological interpretations are very interesting, there are some points that need consideration.

My major concerns are the following:

1- In general the paper has a good structure and is well written but there are some (unnecessary) statements in the abstract, introduction, and also other sections that need to be rephrased on removed to improve the manuscript (ms). Also, there are some typos that need to be checked.

2- Lines 116-117: I can’t see how they concluded about the distribution of the sources from PSD of the signals in Fig. 2. Usually, for different frequency band that will be used in a study, we apply a beamforming or FK analysis to locate the main source but I don’t see such an analysis and there are just some statement in the text that is not enough in my opinion.

3- Line 170: I can’t see any spectral peaks at those frequencies in Fig. 2. I think the figures cross-referencing in the whole ms need some improvement as there are some sentences without (correct) cross-referencing to the corresponding figure.

4- Line 178-187: I do not understand how the numerical simulations results address any of the problems and how it helped the authors as the results demonstrations in Figs 6 and 8 is different. They didn’t show the cross-correlation results for the single source. And to me, it’s not clear how this helped the authors as they didn’t apply their processing technique to the simulated data. And there is no analysis of how applying SNRS improves the results. I understand it has been explained in another earlier paper but it would be useful, if possible, to do the comparison for this simulated data as well.

5- Line 200: In the processing, you mentioned 1-100Hz bandpass filtering but in all the figures for the dispersion analysis you only used frequencies <50, is there any specific reason to use 1-100 and not 50Hz for filtering? And in lines 245, what did you use to eliminate the surface-wave? Is it a notch filter?

6- Line 263: Where are the drilling locations? I couldn’t find any map about their location.

7- Lines 268-270: I don’t understand the meaning of this paragraph, usually the non-stationary phases increase the apparent velocity so there is no need for this statement.

8- Line 287: Again, I can’t see how you concluded if the noise sources are isotropic or not?

9- Could you please explain why you used different packages/codes for dispersion curve calculation for different parts of the data? Somewhere you used “Geopsy” and then changed to “MASW”? I know they both do the job but maybe it would be better to be consistent if you want to compare the results.

10- General suggestion on figures: Maybe you can combine Figures 6 and 8 for a better comparison. There is no need to plot the particle motions. Use a scalebar for some of the figures is possible.

Hope these questions and comments help in the improvement of ms.
Sincerely
M. Rezaeifar

Referee Report: PDF

Hide

RR by Anonymous Referee #4 (18 Mar 2021)

Suggestions for revision or reasons for rejection

Afonin et al conduct seismic interferometry on seismic data in Finland which was originally acquired for active source purposes. They produce velocity models of the shallow (<300m) surface and do numerical modelling of likely seismic sources in the area. The work is important to show how data acquired in non-optimal geometry and with a short recording time can be used for interferometry. However substantial details are missing that limit the conclusions that the authors draw - particularly details on the acquisition and modelling. The authors also do not compare their velocity models to those acquired with the active source, which should be a key part of their analysis. The text and figures also need improvements (particularly maps and poor colour scales).

Regarding the conclusions of the work - I agree that high velocity contrasts are needed to produce scattering but the authors use only 1 subsurface model so I don’t think they can claim they show this in their numerical modelling (they could also add reference to other modelling work). The other conclusion that the SNRS method is needed to produce EGFs is also not shown – can EGFs also be produced without this algorithm?

Major comments
Acquisition – the description of the acquisition set up is confusing and needs more details, a more detailed figure would help to clarify it. Did all 4 lines have the same set up? Were the profiles gathered along roads or footpaths? How many geophones / nodes were active at once and for how long / what was their max offset? Did the authors use the geophones that were deployed in the reflection survey (surely they were not active for very long)? We need more information on the recording time – what was the average recording time between station pairs, and the lowest and highest? Do the authors remove time periods when there are active sources?

Numerical modelling – From the modelling, the authors claim that surface waves are produced that can be used to derive EGFs. But what it is about their model that produces surface waves (e.g. from plane waves arriving from depth)? This could be more impactful if the authors used different velocity models and described how the presence of surface waves changed. E.g what is the effect of the shallow velocity layer and sharp velocity contrasts that are in the model? The authors should also try to describe why Rayleigh waves are not produced in some cases. What is also the effect of moving the source location from parallel to perpendicular to the line?
The authors should also place their modelling work in the context of previous work and compare it. I also question whether modelling the presence of directional Rayleigh waves is enough to conclude that the wavefield is diffuse (e.g. L178).

SNRS method – more details should be added about this method not just giving the reference. From what I understand, the method works by stacking green’s functions but only those that improve the signal to noise ratio. I am really concerned that this will not result in the ‘true’ green’s function but only that with the highest ‘signal’ by whatever way signal is measured.

Figures could be substantially improved:
Figure 1 – please add latitude, longitude. The aerial map is too low quality, I cannot see it clearly. In addition to the geological map, please add a similar sized topographical map showing roads, towns etc.
Figure 2 – the background map is too dark and low resolution. Please add a scale and north. Caption: four stations or 6?
Fig 9 need scale, is the view looking verticall down? Annotate virtual source locations.
The colour scales for velocity models in Fig 12, 14 , 18 are poor. They are not continuous and have been set in some way to make it look like blocks. It is difficult to assess the 2D velocity models with such colour scales. The colour scales are not the same between the models presented.
Add locations of 1D velocity profiles to the 2D velocity lines.

The writing could be improved – I have listed some cases below but there are many more examples.

Other comments
P2L34 ‘actual task’, please rephrase (important task?)
P2L36 remove ‘than earlier’
L49 remove ‘possibility’
L59 new paragraph
L66 I don’t know what the authors mean by ‘directional scatterer’?
L77 the authors should describe why this is an ‘advanced method’.
L105 how long and what time were the spectra calculated for?
Comparison of station spectra (fig 2) - The text implies that the spectra were calculated at different times so this doesn’t seem a fair comparison and could be removed. Or if the authors want to draw some other information out they should add what times the spectra are calculated for.
L131 would be useful to define early on what is meant by heterogeneity.
L137 please summarise the geology from the referenced paper and if there is evidence for mafic dykes as modelled.
L162 what is the orientation of the blast from the line? Why are Rayleigh waves not produced in this case?
L180 what do the authors mean by must be placed in first Fresnel zone area?
L195 is there a comparison to active source results?
L213 is the 1D model assumed to be from directly below the node? How were the 1D models interpolated to get 2D model?
L217 what layer thickness did Aberg etal find?
L225 – what is the depth sensitivity of 3-7 Hz Rayleigh waves?
L230 better to compare these models by adding a difference panel to Fig 14, difficult to compare the models as it is.
L254 is this still part of the vibroseis signals section? Are time periods of vibroseis removed or included?

Hide

ED: Reconsider after major revisions (26 Mar 2021) by Ulrike Werban

AR by Nikita Afonin on behalf of the Authors (05 May 2021) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (06 May 2021) by Ulrike Werban

RR by Anonymous Referee #4 (07 May 2021)

Suggestions for revision or reasons for rejection

"Near surface structure of Sodankylä area in Finland, obtained by passive seismic interferometry" by Afonin et al.

The paper is improved compared the earlier versions. Particularly the figures are much easier to read. I am pleased to see more details of the acquisition and the comparison of the SNRS method with another method for retreiving Green’s functions. The numerical modelling is better described but some questions remain, it might be worth describing in the abstract that it is not comprehensive but designed only to verify that their expected geological features could generate scattering. Also the authors should reduce claims about only the SNRS method working in this case, because they did not actually apply it to synthetic data.

It is interested to see the comparison between the SNRS and conventional method in Fig17. The authors should describe their ‘conventional’ workflow used.

I still don’t like the blocky velocity models. I recommend they are changed to continuous. The authors could read more into how to choose colour scales – e.g. https://www.nature.com/articles/s41467-020-19160-7

In the abstract the authors claim that ‘This scattered wavefield can be used to retrieve reliable Empirical Green Functions (EGFs) from short-term and non-stationary data, ifusing a special technique called “signal-to-noise ratio stacking” (SNRS) is applied.” This should be rephrased as it is misleading, since they did not actually show that only the SNRS method will work on the synthetic data.
Again this claim is made in the text at lines 224 – 227.
There are some sentences in the abstract not needed which would improve its focus.
I may have missed it – was the data acquired during the day and on weekdays?
L99 – what is an RAU?
Fig 14 – the difference colour scale is not good. Generally a difference plot should be something like white at 0 and then increase away from this point.
Fig 18 – colour scale.

Hide

RR by Meysam Rezaeifar (31 May 2021)

ED: Publish subject to minor revisions (review by editor) (31 May 2021) by Ulrike Werban

AR by Nikita Afonin on behalf of the Authors (04 Jun 2021) Author's response Author's tracked changes Manuscript

ED: Publish subject to technical corrections (07 Jun 2021) by Ulrike Werban

ED: Publish subject to technical corrections (08 Jun 2021) by CharLotte Krawczyk (Executive editor)

AR by Nikita Afonin on behalf of the Authors (08 Jun 2021) Manuscript

Short summary

In our study, we show the results of a passive seismic interferometry application for mapping the uppermost crust in the area of active mineral exploration in northern Finland. The obtained velocity models agree well with geological data and complement the results of reflection seismic data interpretation.