REVIEWER #1

This study presents results of an ambient seismic noise imaging in the northwest of Iberia. Based on the higher observed discrepancy between the Rayleigh and Love-wave group velocity and corresponding shear velocity models, a radial anisotropic model is also presented. The manuscript ends with discussion trying to interpret both the isotropic shear velocity model and the observed radial anisotropy in details and in relation with internal and external structures of the area. The manuscript is well written and structured in which careful prepossessing of data and selecting cross-correlations and dispersion measurements has been done. In general, publication of the manuscript is recommended after revision according to the following points and comments.

Major comments:

1.- It’s not clear that from which inter-component the Rayleigh-wave dispersions were obtained. Only from the ZZ? or RR? or both? as shown in Figure 3. In Figure 4, the Rayleigh group velocities out of ZZ are shown. How different or similar are the Rayleigh group velocities out of the RR, and in comparison with the Love-wave velocities? Would it result in the similar Rayleigh Love wave discrepancy as the ZZ? In this study, ambient noise data from 3-components seismic stations was used. It seems the data quality of the horizontal components were good enough to reconstruct the Love-wave velocities (from TT). I am curious why the authors haven’t used the great opportunity to benefit all of the possible inter-components containing the Rayleigh-wave, which are ZZ, RZ, ZR, and RR. Therefore, my recommendation (unless there’s been a serious issue with the data!) is to use all of these four inter-components cross-correlations and to obtain the dispersions of Rayleigh-wave as a production of these four inter-components (e.g a production by logarithmic stacking in the period–group velocity domain introduced by Campillo et al., 1996; see for instance Zigone et al., 2015). This would provide more reliable group velocity of the Rayleigh-wave, and therefore to improve the reliability of the Rayleigh-Love discrepancy.

The Rayleigh-wave dispersion measurements that were used in this study were extracted exclusively from vertical-component (ZZ) cross-correlations. Rayleigh group velocities from radial components (RR) were also calculated, and their average dispersion curve seems quite similar to the ZZ average (Fig. R1):

Figure R1. Rayleigh-wave group velocity dispersion curves obtained from vertical (left panel, ZZ) and radial (right panel, RR) component cross-correlations. Top panels show the number of velocity determinations as a function of the period.

However, we decided not to include dispersion measurements from RR cross-correlations in further processing steps due to three reasons:

• RR derived velocities display higher uncertainty than ZZ velocities. The stability and error analysis of the dispersion curves in Fig. S2 (supplement) shows that RR measurements exhibit uncertainties up to 4%, while ZZ uncertainties are well below 1%. Consequently, this is also noticeable in Fig. R1. In that figure, it can be seen that the range of group velocities is wider for the R-R dispersion measurements, specially at longer periods (> 8s).
• The signal-to-noise ratio (SNR) of the ZZ cross-correlations is, on average, 3 times higher than the RR cross-correlations SNR.
• Some RR dispersion curves cannot be calculated due to low-quality empirical Green’s functions or higher-mode contamination. This fact leads to a lower number of group velocity estimations (Fig.R2) and the loss of some important interstation paths.

Figure R2. Vertical (left panel) and radial (right panel) component MFA surfaces derived from the same inter-station path (CAST-DEGA). Note the lower quality of the radial MFA surface.

Considering these three points, we believe that adding Rayleigh-wave group velocities from RR, RZ or ZR cross-correlations to the calculations is not likely to benefit or enhance the reliability of the results. On the contrary, it will mean to include higher uncertainty data in our processing. As stated in Campillo et al., (1996), the logarithmic stacking of the ZZ, RZ, ZR and RR inter-components is useful when the quality of the seismic dataset makes difficult to define dispersion curves with confidence. In our case, we demonstrated that the ZZ dispersion measurements and the subsequent results are reliable. In fact, most of the radial anisotropy studies use Rayleigh dispersion measurements obtained only from ZZ cross-correlations (e.g., Shrizard et al., 2017, Dreiling et al., 2018; Movaghari et al., 2021; Alder et al., 2021).

We have added a few lines in the manuscript (265-268) to clarify that Rayleigh wave dispersion measurements come from ZZ cross-correlations and briefly commenting the reasons why RR cross-correlations were discarded. We have also included RR examples to Fig. S2 in the supplemental material.

2.- The authors have mentioned the significance of the azimuthal anisotropy resulted in their previous study (Acevedo et al., 2020). Anisotropic structures in general might not be well explained neither by radial anisotropy alone nor by azimuthal anisotropy. Since results of azimuthal anisotropy (fast orientation and delay time) in the study area is present, and also in order to have a better picture of the anisotropic structures, both radial and azimuthal anisotropy should be discussed. More particularly, the distribution of fast-polarization orientation over depth (at each period) in comparison with Radial anisotropy could be discussed in terms of deformation history and regional tectonic regimes of the area associated with late Variscan shear zone and /or Alpine convergence.  However, there is little discussion in this matter (line 434-435). I therefore suggest including the depth variation of the fast orientations superimposed on the radial anisotropy pattern at each depth (Figure 7) and add more discussion based on such comparison. The author could try to address questions, for instance, how the fast-polarization orientation varies from surface to depth while radial anisotropy is increasing? Is there a relation between radial and azimuthal anisotropy at the region where there is a change from negative to positive RA (contrasts observed under GTOMZ, CIZ and CZ; line 421-422)? In lines 463-465 as the authors discussed the cause of positive radial anisotropy by the horizontally sheared fabrics, here the depth pattern of fast orientations could help in better understanding such effect.

We agree that a more direct comparison between radial and azimuthal anisotropy will help to better understand the anisotropic structure because, as you have rightly pointed out, it might not be well explained neither by radial nor by azimuthal anisotropy alone. We have rewritten and extended the discussion on this matter (from lines 456 to 492), and we have added a new figure (Fig. 9) that compares directly the radial and the azimuthal anisotropy magnitude (Acevedo et al., 2020) in the CZ and shows how fast azimuthal directions vary with depth. In addition, a panel showing the variation of the seismic anisotropy with lithostatic pressure measured in rocks from NW Spain (Brown et al., 2009) is included.

The new discussion tries to answer the interesting questions raised but there are some limitations that are inherent to the techniques used in Acevedo et al., (2020). On the one hand, seismic anisotropy derived from shear wave splitting represents the average anisotropy in the layer causing it, but it is not constrained at depth. On the other hand, ambient noise-based azimuthal anisotropy measurements are dependent on their period, and they can be associated to their approximate depth by analizing the sensitivity kernels, but they are only representative of the CZ region. That is why we can only compare directly radial and azimuthal anisotropy in the CZ. Nevertheless, some striking features can be observed. The minimum magnitude of both the radial and the azimuthal anisotropy coincides with the theoretical depth at which cracks are closed by the lithostatic pressure (we have used the laboratory data in Brown et al., (2009) because those rocks are from near the study area, but in fact, the effect of crack closure is commonly observed in crystalline rocks at increasing confining pressures). This also coincides with the depth of the transition between the Alpine-reworked Variscan cover and the pre-Variscan basement. The shift is also reflected in the pattern at depth of the fast-azimuthal directions, and it has important implications in terms of the processes that control the anisotropy. These observations are now described in the discussion section. 

Other comments:

Introduction:

3.- Line 71; “The ANI reflects the variation of the seismic velocities of the bulk rock …” seems not a correct statement. Other passive seismic imaging techniques that use earthquake data can also provide images of velocity variation of subsurface structures.

The sentence has been changed to clarify that other passive seismic imaging techniques can provide information about the variation of seismic wave velocities.

4.- Line 72; “…, which is controlled by their elastic parameters”. How about temperature? Please modify or ignore this statement.

The incorrect statement has been removed.

5.- Line 73-75; Earthquake-based tomography can also provide images of upper crustal structure. A main advantage of ANI is data availability, particularly in regions with low seismicity -where not enough regional/local earthquake occurring- and also as an alternative to high cost active survey.

Lines 79 and 82-83 have been modified to explicit the mentioned advantages of the ANI method.

6.- Please provide an earlier study as reference in line 27-28 (e.g. Silver, 1996). The same in line 29; here you could use e.g. Mainprice et al., 2000.

Both studies have been added to the manuscript.

7.- Line 30; please remove Shapiro et al., 2004 as it’s not a proper reference for anisotropy.

The reference has been removed

8.- Line 61; please provide reference for “Overall, the part of the Variscan belt that crops out in the Cantabrian Mountains (CM) represents one of the most complete sections of this orogen in Europe, …”

The assertion was extracted from a chapter of the book: Spanish Geological Frameworks and Geosites: An approach to Spanish Geological Heritage of International Relevance, edited by García-Cortés, A. The corresponding reference has been added to the manuscript (line 66).  

Geological setting:

9.- Line 135; “Overall, the part of the Variscan Belt that crops out in the CM represents one of the most complete sections of this orogen in Europe …” This is a repeated sentence. Please remove it.

The repeated sentence has been removed.

10.- Section 2.3 seems a bit long and include a lot of geology. Particularly, explaining the four distinguished domains from line 156 till line 179; Are all of these are necessary for the discussion? Could you make it shorter?

A few lines have been removed from section 2.3 to make it more concise. However, we think that it is a brief geological summary that helps to interpret the results from the tomographic maps and the radial anisotropy.

Data and Method:

11.- It’s not clear how many stations in total were used. 13 temporary (portable) but two of them were permanent?! In case they were all temporary, you may remove the word “permanent” in line 196.

The GEOCSN network was composed of 11 portable stations, but we also had access to the data that was simultaneously acquired by two permanent stations of the Spanish Seismological Network (SSN). That is the reason why we said that the network had 13 stations, 11 temporary stations plus 2 permanent stations. Please see next point for more details.

12.- Have you used all stations between June 2019 and February 2020?  I suggest making the Sec 3.1 more clear and including information about how many stations in total, from which network in details, and in which the time period the continues seismic recordings were used.

The GEOCSN network was active between June 2019 and February 2020. As we explained in the previous point, it was formed by 11 portable stations, and it was complemented with data from two permanent stations of the SSN. The other experiment that provided data for this study was the IberArray seismic network. It was active between 2011 and 2013, but we have processed only 12 months of data from that network (year 2012) because the temporary stations had different installation and removal dates, but they were all active during the year 2012. In total, we used data from 7 portable stations of this network. Like in the case of the GEOCSN network, the IberArray dataset was augmented with seismic data from one station of the SSN. We have rearranged Section 3.1 and we have added some explanations to make these points clearer. We have also included the position of the seismic stations in Fig. 1 (see point 20).

13.- How you assessed any seasonal effect in the cross-correlation functions? Perhaps you could assess it in a way in Figure S1, and to add a sentence about it to the text.

Seasonal effects are the result of the variation in the position and intensity of the noise sources. Theoretically, noise sources must be fully equipartitioned to obtain reliable dispersion measurements, but this requirement is not usually fulfilled. The f-k analysis performed in the study area by Olivar-Castaño et al., (2020) demonstrated that, although the most intense seismic sources are located to the N and the NW of the area, a sufficient level of noise arises from all azimuths. Nevertheless, the stacking of noise-cross correlations over long time spans, as we do in this study, diminishes the impact of potential seasonal effects in the measurements (Corela et al., 2017). Anyway, the non-stationary/non-isotropic nature of the seismic noise wavefield does not alter substantially the measured velocities of surface waves and is not a critical issue (e.g., Froment et al., 2010). A brief paragraph on this matter was added to Section 3.2 (lines 269-273).

14.- Please provide more example of inter-station dispersion curve (as in Figure 2d) in Supplementary material.

A new figure showing more inter-station dispersion curves has been included in the Supplemental material (Fig. S1).

15.- I could not quite understand “… defining a constant S-wave velocity…” in line 272. Please make in a clearer way.

The statement has been changed to make it clearer. We tried to express that we used and initial model with a constant S-wave velocity of 3.35 km/s.

Discussion:

16.- In line 353, “…both the surface and shear-wave ones…”, do you mean both the surface-wave group and shear velocities?

That is right, the sentence has been corrected.

17.- What does the “RA style” refer to? in line 404 and in a couple of other lines. It’s confusing.

The anisotropy style referred to its sign: positive or negative. We now use the term “sign” instead of “style” in the text.

18.- How did you obtain the angels 60-90 and 0-30 degree for dipping features (line 406)? Are these numbers estimated from the shape of the positive and negative anomalies in the cross-section (Figure 8c)? If so, I recommend avoiding giving quantity of the dipping structures because such image can differ by changing smoothing parameters of the inversion or even in plotting.

The mentioned angles are not a result from our study. In fact, those values were firstly proposed by Xie et al. (2013) and they were obtained from laboratory tests in rock samples. To our knowledge, they are still valid, and they have been featured in recent radial anisotropy studies (e.g. Dreilling et al., 2018).

Conclusion

19.- Line 472-474; “The observed discrepancy between the measurements from Rayleigh and Love waves impeded the possibility of performing a joint inversion …” not really clear what you mean.  It is possible to perform joint inversion using the dispersion measurements from Rayleigh and Love to derive both isotropic and anisotropic Vs models. It’s fine that joint inversion was not the purpose of this study. But it would not be the reason that why you have used Rayleigh and Love-based separately derived Vs model to infer Radial anisotropy. Please rephrase the sentence or might move it to the end of the conclusion and as suggestion for a later study.

The sentence has been removed, because it is completely right that it is possible to perform a joint inversion.

Figures

Figure2:

20.- Even though the Fig2a with its colors related to Fig2e and 2f is quite helpful, it is too small for representing the station location and geometry. Please provide a larger figure showing all the station’s location with different symbol color for each network. Such a figure could be added to Figure 1.

Seismic stations have been represented in Fig. 1. Symbol colors indicate the seismic network at which they belong, and stations names are depicted. 

Figure3:

21.- With regard to my comment No.1, the ZR and RZ Cross-correlations could to be shown in fig.3.

Please see point number 1.

Suggested references:

• Campillo, M., Singh, S., Shapiro, N., Pacheco, J., and Herrmann, R.: Crustal structure South of Mexican Volcanic belt based on group velocity dispersion, Geofis. Int., 35, 361–370, https://doi.org/10.1016/j.crte.2011.07.007, 1996.
• Mainprice, D., G. Barruol, and W. Ben Ismail (2000), The seismic anisotropy of the Earth's mantle: From single crystal to polycrystal, in Composition, Structure and Dynamics of the Lithosphere Asthenosphere System, Geophys. Monogr. Ser., vol. 117, edited by S. Karato, A. Forte, R. Liebermann, G. Masters, and L. Stixrude, p. 237, AGU.
• Silver, P. (1996), Seismic anisotropy beneath the continents: probing the depths of geology, Annu. Rev. Earth Planet. Sci., 24, 385, 432.
• Zigone, D., Ben-Zion, Y., Campillo, M., and Roux, P.: Seismic Tomography of the Southern California Plate Boundary Region from Noise-Based Rayleigh and Love Waves, Pure Appl. Geophys., 172, 1007–1032.

Most of the suggested references are now cited in the manuscript.

REFERENCES

Acevedo, J., Fernández-viejo, G., Llana-Fúnez, S., López-Fernández, C. and Olona, J.: Upper crustal seismic anisotropy in the Cantabrian Mountains (North Spain) from shear-wave splitting and ambient noise interferometry analysis, Seismol. Res. Lett., 92, 421-436, https://doi.org/10.1785/0220200103, 2020.

Alder, C., Debayle, E., Bodin, T., Paul, A., Stehly, L., Pedersen, H. and the AlpArray Working Group: Evidence for radial anisotropy in the lower crust of the Apennines from Bayesian ambient noise tomography in Europe, Geophysical Journal International, 226, 941–967, 2021.

Brown, D., Llana-Funez, S., Carbonell, R., Alvarez-Marron, J., Marti, D. and Salisbury, M.: Laboratory measurements of P-wave and S-wave velocities across a surface analog of the continental crust–mantle boundary: Cabo Ortegal, Spain. Earth Planet. Sc. Lett., 285, 27-38, https://doi.org/10.1016/j.epsl.2009.05.032, 2009.

Campillo, M., Singh, S., Shapiro, N., Pacheco, J., and Herrmann, R.: Crustal structure South of Mexican Volcanic belt based on group velocity dispersion, Geofisica Internacional, 35, 361–370, 1996.

Corela, C., Silveira, G., Matias, L., Schimmel, M. and Geissler, W. H.: Ambient seismic noise tomography of SW Iberia integrating seafloor – and land-based data, Tectonophysics, 700, 131-149, 2017.

Dreiling, J., Tilmann, F., Yuan, X., Giese, J., Rindraharisaona, E. J., Rümpker, G. and Wysession, M. E.: Crustal radial anisotropy and linkage to geodynamic processes: a study based on seismic ambient noise in southern Madagascar, J. Geophys. Res.-Sol. Ea., 123(6), 5130-5146, https://doi.org/10.1029/2017JB015273, 2018.

Froment, B., Campillo, M., Roux, P., Gouédard, P., Verdel, A., and Weaver, R.L.: Estimation of the effect of nonisotropically distributed energy on the apparent arrival time in correlations, Geophysics, 75, 85-93, 2010.

Martínez-Catalán, J.R., Aller, J., Alonso, J.L. and Bastida, F.: The Iberian Variscan Orogen, in: Spanish Geological Frameworks and Geosites: An approach to Spanish Geological Heritage of International Relevance, edited by: García-Cortés, A., IGME, Madrid, Spain, 13-30, 2009.

Movaghari, R., JavanDoloei, G., Yang, Y., Tatar, M., and Sadidkhouy, A.: Crustal radial anisotropy of the Iran Plateau inferred from ambient noise tomography. Journal of Geophysical Research: Solid Earth, 126, e2020JB020236, 2021.

Olivar‐Castaño, A., Pilz, M., Pedreira, D., Pulgar, J. A., Díaz‐González, A. and González‐Cortina, J. M.: Regional Crustal Imaging by Inversion of Multimode Rayleigh Wave Dispersion Curves Measured from Seismic Noise: Application to the Basque‐Cantabrian Zone (N Spain), J. Geophys. Res.-Sol. Ea., 125(12), e2020JB019559, https://doi.org/10.1029/2020JB019559, 2020.

Shirzad, T., Shomali, Z. H., Riahi, M. A. and Jarrahi, M.: Near surface radial anisotropy in the Rigan area/SE Iran, Tectonophysics, 694, 23-34, http://doi.org/10.1016/j.tecto.2016.11.036, 2017.

Xie, J., Ritzwoller, M. H., Shen, W., Yang, Y., Zheng, Y. and Zhou, L.: Crustal radial anisotropy across eastern Tibet and the western Yangtze craton, J. Geophys. Res.-Sol. Ea., 118(8), 4226-4252, https://doi.org/10.1002/jgrb.50296, 2013.

REVIEWER #2

The paper entitled “Radial anisotropy and S-wave velocity depict the internal to external zones transition within the Variscan orogen (NW Iberia)” by Acevedo et al. conducted ambient noise tomography using recently deployed seismic arrays to constrain the velocity and radial anisotropy of the upper crust in NW Iberia.  The resulting seismic image shows a good correlation with major geological domains and known structural variation in the Variscan orogen.  Interestingly, the seismic model shows a clear structural transition from the hinterland and external zones of the Variscan orogeny.  This radial anisotropy model provides new seismic constraints to the study region and adds knowledge to the deformation processes in orogens.  The topic is a good fit for the journal of Solid Earth.  The manuscript is well structured and is generally well written.  I think that this manuscript is suitable for publication after some minor revisions.  I summarize my main concerns below, which are about the resolution analysis and the interpretation of anisotropic structures in regions with a suboptimal resolution, and hope these are helpful to further clarify some points and strengthen the paper. 

1.- The actual inversion of group velocity used a grid size of 0.1 degree. In the checkerboard test shown in the supplementary material, the size of the grid to construct the anomaly seems to be quite big.  Was the checkerboard test also using 0.1 gird, or the inversion gird was set to the same size as the anomaly?  This needs some clarifications.

All the chequerboard resolution tests were performed using the same grid size (0.1º x 0.1º) and regularization parameters (damping = 0.001, smoothing = 0.1) of the group velocity inversions, in order to ensure the representativeness of the tests. The grid size and the inversion parameters do not change with the cell size of the initial chequerboard model. This information has now been added to the captions of Fig. S3 and Fig. S4.

2.- Shear velocity structures of Vsh and Vsv are inverted separately. The results look reasonable, but could the authors elaborate on how the inversion parameters were properly chosen to ensure the same degree of amplitude recovery between the two models? In other words, how could you make sure that the velocities obtained from two separate inversions are comparable between each other?

This is a good point because it is true that the sensitivity of Love waves decays more rapidly with depth than Rayleigh waves, and that can affect the resulting shear wave velocities. We are aware that some studies try to compensate the decrease on Love wave group velocity sensitivities by varying the inversion parameters at depth (e.g. Wang et al., 2020). However, the analysis of the Rayleigh and Love sensitivity kernels (Fig. S5) shows that, at the depth range that has been investigated in this study (1-12 km), Love and Rayleigh sensitivities are comparable, ensuring the extraction of comparable Vsh and Vsv velocities. For that reason, we have maintained identical inversion parameters for Vsv and Vsh velocities, like other radial anisotropy studies (e.g. Lynner et al., 2018).

3.- The resolution near the edges of the imaging area is really not ideal, and structures there may not be well constrained by the data. Therefore, I am a bit worried about the interpretation of small-scale anomalies in these regions of suboptimal resolution.  For example, on lines 414-415, a deeper transition depth of anisotropic structures beneath the CIZ-GTOMZ is used as an argument for the presence of a basal detachment fault.  This is a good observation, but I feel that this may run into the risk of over-interpreting structures that are not well constrained by the data.  Similarly, on lines 420-424, the resolution in CIZ and GTOMZ are relatively low, as also acknowledged by the authors, yet detailed interpretations are given here. Unless the authors can substantiate the robustness of these structures, I would suggest minimizing the discussion of structures with suboptimal resolution.

As reviewer #2 has pointed out, the resolution in the edges of the models is reduced due to the lack of interstation paths in those areas. As suggested, we have minimized the discussion about structures and anomalies in regions that were not well resolved, such as the CIZ-GTOMZ domain. For example, the interpretation of the transition between anisotropic structures as the result of the presence of a basal detachment can be supported with observations from the CZ, the best resolved area in our models (lines 440-441). The discussion about the CIZ-GTOMZ area has also been reduced and we now clearly state that the resolution in the area is not optimal and further investigation is recommended (lines 445-453).    

4.- In figure 6, I suggest using the same color range when plotting the two models. It is difficult to compare them.

The suggestion is interesting, and we have created a figure depicting both models with the same color range (Fig. R3). The main feature of the models, which is the velocity variation in the external-internal zones transition, is still visible. However, Love wave velocities are higher than Rayleigh wave velocities, and this results in a general color range that is too wide to display some velocity anomalies in the Vsv maps, specially at higher periods. Considering that these velocity variations may have important implications from a geological perspective, we believe that using two different color ranges for the Vsv and the Vsh models renders more information to the reader. Nonetheless, we have modified slightly the Vsv and the Vsh color scales. The previous ones were constructed by extracting the minimum and the maximum velocity values within all the inverted slices. Now, we have only considered the depth slices depicted in Fig. 6 to select these values, enhancing the visibility of the high velocity anomalies in the Vsv – 12 km map.

Figure R3.  Inverted Vsv (left panels) and Vsh (right panels) tomographic maps for depths of 3, 6, 9 and 12 km. A common color scale has been used in all the maps.

5.- Figure 8, please label geographic locations such as CCB, NA, Allande and Vivero faults on cross-sections. Also, the top 1 km of the model is not shown, any reason for this?

The mentioned geographical locations and structures have been labelled in Fig. 8. Moreover, the trace at depth of the labelled faults has been represented in the cross-sections. In many ambient noise-based studies, the top kilometer of the models is not represented due to the difficulty of extracting dispersion measurements at low periods (< 2s in our case). The absence of high-frequency measurements leads to a lack of information in the shallowest part of the models that compromises their reliability near the surface.

I also have some minor suggestions referring to the line number.

6.- Line 18, “orogenic grain” -> “orogenic belt”

The word has been changed

7.- The same line, “bulk properties of the rocks” sounds like the bulk composition of the rocks.  Please consider using antherword such as “elastic properties of the rocks”.

The word has been changed, elastic properties is more accurate.

8.- Line 24, “caused by” -> “which we attribute to”

Corrected

9.- Line 25, ‘the internal deformation of rocks either during the Variscan orogeny or prior to it’ -> “the pre- or syn-orogenic deformation associated with the Variscan orogenesis”.

Corrected

10.- Line 31, “shear waves” -> “shear wave velocities”

Corrected

11.- Line 37, “whose importance varies with depth” -> “that dominate different depth levels”

The sentence has been changed

12.- Line 42, “depth” -> “depths”

Corrected

13.- Line 45, “the features that govern the” -> “the governing features”

Corrected

14.- Line 50, “been” -> “been identified”.

Corrected

15.- Line 51, “ancient orogenic belts (Wang et al., 2020)”.  Although there are some disputes on the age of the initiation of Cordilleran orogenesis, it is certainly a Phanerozoic orogeny and is likely as old as the Variscan orogeny (Paleozoic).  So I would not use the word “ancient”, which more properly refer to orogenesis in Precambrian.

Many thanks for the clarification, it is true that the term “ancient” suggests a Precambrian origin. We have removed the word.

16.- Line 63, “an orogenic system, in the West, to the external zone, to the East” -> “an orogenic system (west) to the external zone (east)”.

Corrected

17.- Line 64, “in the area” -> “in this area”.

Corrected

18.- Line 67, “helped to broaden” -> “broadened”.

Corrected

19.- Line 68, “In order to increase the resolution in the structure of the crust” -> “To improve imaging resolution at crustal depths”.

Corrected

20.- Line 74, “for the unraveling of” -> “for unraveling”

Corrected

21.- Line 177, “that are mostly made of” -> “that they are mostly made of”.

Corrected

22.- Line 186, “It is only in the first of the domains defined, in the CZ, that it has been reported a …”, this can be simplified, “Earlier studies have been reported in the CZ domain a”

Corrected

23.- Line 194, ’11 stations’ -> “Among them, 11 stations”.

The entire section 3.1 (Seismic data) has been rearranged to make it clearer. This suggestion was useful to explicit that the GEOSN was composed by 11 portable stations, but we also have access to the data acquired by two permanent stations in the area. We have tried to explain this fact in a better way in the new text (please see Reviewer #1, point 11 for more details).

24.- Line 208, “26% of overlap” -> “26% overlap”.

Corrected

25.- Line 214 “with corner frequencies between 0.01-2.0 Hz”-> “with corner frequencies of 0.01 and 2.0 Hz”.

Corrected

26.- Line 292 “keep delineating a large high” -> “delineate a consistent large high”

Corrected

27.- Line 315, “higher depths” -> “greater depths”

Corrected

28.- Line 354, “element of our models, both the surface- and the shear-wave ones, is” -> “element in both group and shear velocity models is”

Corrected

29.- Line 390, highlight the velocity contour of 3.1 km/sec using a thick line or another color.

The 3.1 km/s velocity contour has been highlighted in grey in Fig. 8a. The figure caption has been changed accordingly.

30.- Line 455, citation format issue, remove the extra comma, “Chen et al. (2009) and Guo et al. (2012)”

Corrected

REFERENCES

Lynner, C., Beck, S. L., Zandt, G., Porritt, R. W., Lin, F. C. and Eilon, Z. C.: Midcrustal deformation in the Central Andes constrained by radial anisotropy, J. Geophys. Res.-Sol. Ea., 123(6), 4798-4813, https://doi.org/10.1029/2017JB014936, 2018.

Wang, J., Gu, Y. J. and Chen, Y.: Shear velocity and radial anisotropy beneath southwestern Canada: Evidence for crustal extension and thick‐skinned tectonics, J. Geophys. Res. - Sol. Ea., 125(2), e2019JB018310, https://doi.org/10.1029/2019JB018310, 2020.