Variability in the Statistical Properties of Continuous Seismic Records on a Network of Stations and Strong Earthquakes: A Case Study from the Kamchatka Peninsula, 2011–2021
Round 1
Reviewer 1 Report
=================
Section 1
=================
- The authors state the possibility of identifying areas of strong earthquakes with a lead time of months (line 67). Please elaborate on this, explaining the contribution of each reference that was cited.
- Please state the advantages and advantages of using the method developed by A.A. Lyubushin.
- Are there other authors or research groups working recently in the same research, application of a similar approach? Is the method used in the manuscript comparable to other ones in the literature?
_ It is not clear if the method was previously used in other areas. In case positive, what are the findings or recommendations?
=================
Section 2
=================
- Please format Table 3, the caption characters are misconfigured
- High-order polynomial can overfit some curve-fitting, depending on the data sparsity (line 233). Are there alternatives to the 8th order polynomial to get the estimates of the multi-fractal features?
=================
Section 3
=================
The following statement seems too vague and needs deeper discussion and a strong argument (associated with the results): "Thus, a certain correspondence of the distinguished areas of strong earthquakes danger with occurred seismic events give reason to believe the processes of strong earthquakes preparation are reflected in regular behavior of seismic noise parameters."
=================
Section 4
=================
- Could the authors discuss this research's application of the conceptual model to other areas? This suggestion appears in the abstract, but it needs to go deeper to consolidate the discussion on further research.
- Please make sure your conclusions section underscores the scientific value added to your paper, and the applicability of your findings, as indicated previously. Please revise your conclusion part in more detail, and enhance your contributions and limitations.
Author Response
We express our gratitude to reviewer 1 for a careful acquaintance with our work and a number of useful comments aimed at improving the content of the paper. Below is a double numbering of the reviewer's comments by section and our responses.
1-1. The authors state the possibility of identifying areas of strong earthquakes with a lead time of months (line 67). Please elaborate on this, explaining the contribution of each reference that was cited.
Corrections made “The possibility of the method for identifying areas of strong earthquakes in the Northwestern part of the Pacific seismic belt, North America and the whole world with a lead time of months – a few years was demonstrated in [5, 13, 14, 16].” Please see lines 65-66 in the new version of the manuscript.
1-2.- Please state the advantages and advantages of using the method developed by A.A. Lyubushin.
The modification of the Lyubushin method used for processing data from continuous recording of seismic signals at a network of stations on the Kamchatka Peninsula has not been previously used. The advantages of the method used are related to the deep formalization of all procedures for processing the initial data, which is due to their large volume. All implemented data processing procedures are described in sufficient detail in the article and previous publications listed in the References.
1-3. - Are there other authors or research groups working recently in the same research, application of a similar approach? Is the method used in the manuscript comparable to other ones in the literature?
In this paper, to study the spatio-temporal properties of continuous seismic noise at the network of stations on the Kamchatka Peninsula, an original set of methods and software for multivariate statistical analysis of time series. Moreover, each of the individual methods used, in particular, wavelet analysis, multifractal analysis, multidimensional spectral-temporal analysis, are traditional in geophysics and seismology. Relevant references to primary sources that describe individual statistical methods are given in the References (No 40, 41, 42).
1-4 _ It is not clear if the method was previously used in other areas. In case positive, what are the findings or recommendations?
The method previously was used for investigation of seismic noise properties in Japan, South California and for the whole world, please look articles in References:
No 15: Lyubushin A. (2020) Global Seismic Noise Entropy // Frontiers in Earth Science, 8:611663. https://doi.org/10.3389/feart.2020.611663
No 16: Lyubushin A.A. (2021) Seismic Noise Wavelet-Based Entropy in Southern California // Journal of Seismology, First online: 21 August 2020, 25:25-39 (2021), https://doi.org/10.1007/s10950-020-09950-3
No 17: Lyubushin A. (2021) Low-Frequency Seismic Noise Properties in the Japanese Islands // Entropy 2021, 23, 474. https://doi.org/10.3390/e23040474 ,
These works show the prospects of the developed methodological approach for spatiotemporal diagnosing modern geodynamic processes, including the preparation of strong earthquakes. These works consider examples of synchronization of seismic noise parameters before earthquakes in the northwestern part of the Pacific seismic belt - the Kronotsky earthquake on December 5, 1997, М=7.8; Simushir earthquake on November 15, 2006, M=8.3; earthquakes in the area of Japan - Hokkaido on September 25, 2005, M=8.3, Tahoku 2011, M=9.1 and others.
In each of these works, the features of the computational procedures for seismic noise parameters and diagnostics of synchronization effects are used, showing the development of the method over time. The approaches used in our work, taking into account the characteristics of the region and the network of stations, are described in sufficient detail in Section 2.
2 – 1. Please format Table 3, the caption characters are misconfigured
Corrections have been made to the Table 3 header. Highlighted in green.
2-2. High-order polynomial can overfit some curve-fitting, depending on the data sparsity (line 233). Are there alternatives to the 8th order polynomial to get the estimates of the multi-fractal features?
The choice of 8th order polynomial for trend is following from the experience of testing different orders. This order provides elimination of trends due to tides and temperature influence. This question is discussed in details earlier in the article (see References, No. 16):
Lyubushin A.A. (2021) Seismic Noise Wavelet-Based Entropy in Southern California. Journal of Seismology, First online: 21 August 2020, 25:25-39 (2021), https://doi.org/10.1007/s10950-020-09950-3
3-1. The following statement seems too vague and needs deeper discussion and a strong argument (associated with the results): "Thus, a certain correspondence of the distinguished areas of strong earthquakes danger with occurred seismic events give reason to believe the processes of strong earthquakes preparation are reflected in regular behavior of seismic noise parameters."
4-1. Could the authors discuss this research's application of the conceptual model to other areas? This suggestion appears in the abstract, but it needs to go deeper to consolidate the discussion on further research.
4-2. Please make sure your conclusions section underscores the scientific value added to your paper, and the applicability of your findings, as indicated previously. Please revise your conclusion part in more detail, and enhance your contributions and limitations.
We have tried to respond to these comments in an updated version of Section 4 Discussion and Conclusions. Please look lines number 474-488 in new version of the paper.
Galina Kopylova
August 26, 2022
Author Response File: 
Author Response.docx
Reviewer 2 Report
This paper reports on a 10-year retrospective study of potential precursory anomalies. It studies variability and synchronicity of signals in seismic noise measured on a a seismic network in Kamchatka from 2011-2021. Time series are constructed four noise components: the generalized Hurst exponent, singularity support width, wavelet based spectral exponent and minimum normalized entropy of squared orthogonal wavelet coefficients. It was found that strong earthquakes occurred near areas of the minimum valuues of the four noise components.
I congratulate the authors for an excellent account of this work. This study is a good example of the kind of painstaking empirical work, related to conceptual physical theory, that needs to be carried out to develop well-defined forecasting models, test them in due course, and eventually advance the information value of medium-term earthquake forecasts in the future. The presentation of this work is mostly very clear and readable, even though a huge amount of technical work has gone into it.
In my opinion the paper can be published with only minor revisions. Here and there, the flow of sentences can be improved to a more idiomatic style of English, but I am sure this can be handled by the copy editors. I have detailed below a few minor wording suggestions and typos that came to my attention.
1. Line 215: "strive". I think a better word would be "force
2. Line 225: too many "not"s
3. Line 239-240. This sentence is not clear. Please reword.
4. Line 272: "suppress" -> "suppression"
5. Line 279: "each of noise statistics" -> "each noise statistic".
6. Line 284: "the features in noise parameters changes" -> "the features of noise parameter changes"
7. Line 475 "weakly" -> "weak"
Author Response
We express our gratitude to reviewer 2 for careful acquaintance with our work and positive assessment. All comments of the reviewer (1-7) are taken into account in the revised manuscript.
- Line 215: "strive". I think a better word would be "force»
Agreed and replaced.
- ine 225: too many "not"s
Edited: For some values a the set C(a) are not empty, that is, there are some minimum amin and maximum amax, such that only for amin < a < amax the set C(a) contain some elements.
- Line 239-240. This sentence is not clear. Please reword.
Edited: Let cj(k) – are the wavelet coefficients of the analyzed signal x(t), t = 1,…,L are a discrete indexes numbering successive values of the time series.
- Line 272: "suppress" -> "suppression"
Agreed and replaced.
- Line 279: "each of noise statistics" -> "each noise statistic".
Agreed and replaced.
- Line 284: "the features in noise parameters changes" -> "the features of noise parameter changes"
Agreed and replaced.
- Line 475 "weakly" -> "weak"
Agreed and replaced.
With good wishes,
Galina Kopylova
Author Response File: Author Response.docx
