Modeling the Impact of the Viscoelastic Layer Thickness and the Frictional Strength to the Lithosphere Deformation in a Strike-Slip Fault: Insight to the Seismicity Pattern along the Great Sumatran Fault
Round 1
Reviewer 1 Report
The buildup and release of strain, and the stress loading process in the earthquake cycle are important information for seismic hazard mitigation. Therefore, the impact parameters including lithospheric layer thickness and fault frictional strength, should be analyzed quantitatively to understand the above-mentioned process. The authors constructed simplified models referred to the Great Sumatran Fault using the finite-element model. In their model, the l elastic-brittle layer (upper crust) and the elastic-viscous layer (lower crust and upper mantle) make up the modeled lithosphere. They analyzed two main different phases including the interseismic and postseismic ones.
The results are very important not only for the knowledge of the Great Sumatran Fault seismicity but also for the option of modeling the deformation. Though the authors have shown important results, there are still some problems that need to be improved or provide more information as follows:
1) P3-L99, the authors set the research zone as a block with 16km×16km×6km, what is the downscale factor between the real Great Sumatran Fault zone and the block?
2) P4-L130, the authors selected 1cm/yr and 3cm/yr as the velocity boundary conditions in x and y directions, however, as I know from the previous parts, the authors downscale the geometry of the fault model for computational convenience. How did you obtain or determine these two velocity parameters? Did you consider the scale like the geometry downscale? Please provide more detailed information about them as they are the most important parameters for the simulation.
3) The authors simulated the earthquake cycle process, for each phase, in my opinion, a suitable way to determine each process duration by paleoseismic record or references. Please provide the reason for each time duration determination.
4) P1-L33-L35, please check the verb tenses.
5) P3-L80, I can’t find the reference Wang [2007] in the last reference part, same as the P3-L107 Turcotte and Schubert [2014], are they cited correctly?
6) P3-109, Table 1. The parameter name should be “Density”?
7) P4-L142, When the authors simulated the coseismic process, they set the steady lower creep rate as 1.5cm/yr. The lower steady creep of the fault causes stress loading or releasing, moreover, impacts the results in this paper, therefore, I suggest the authors give some reasons for it.
8) P5-P8, the titles of Fig4 and Fig5 are almost the same, please give a more detailed title for each. Same situations are in Fig6 and Fig7.
9) P5-P8, In Figs 4, 5, 8, the authors show the displacement magnitude or the slip magnitude without unit. The magnitude of displacement and slip are important, they will help decide whether the layer thickness and the frictional strength should be concerned in the models. If they are not significant in the simulation, then this paper will lose its real value. So, please give the correct unit for each figure.
10) P10-L279-284, check the words in “Based on our numerical results, we suspect that the stress accumulation in the central and southern is higher that the northern part”, and “The smaller stress accumulation in the northern part of GSF correlate with the absence of major events (M>7) in the northern part of GSF, even after the devastating M>9 event in 2004.”
Author Response
Response to Reviewer 1 Comments
Point 1: P3-L99, the authors set the research zone as a block with 16km×16km×6km, what is the downscale factor between the real Great Sumatran Fault zone and the block?
Response 1: The dimension of the model used in the paper was constructed based on the previous studies, i.e., Lynch & Richards, 2001; Natawidjaja & Triyoso, 2007; Burton & Hall, 2014, in which in average the segment length is 120 km with a thickness of 45 km (15 km and 30 km of brittle upper crust layer and lower crust and upper mantle, respectively. To simplify the calculation and reduce the computational time, we downscale the geometry of model by 7.5. Therefore, we end up with 16 km x 16 km x 6 km dimension of the fault segment length, width, and thickness, respectively. Previous studies, i.e., Li et al. (2009), Wang et al. (2017) and Meyer et al. (2017) have shown the resulted strain accumulation and release are also impacted by the thickness ratio between the elastic and viscoelastic layer. We also vary the thickness ratio between elastic upper crust and viscoelastic lower crust and upper mantle by 1:1, 1:2, and 2:1 to see the impact on the resulted strain accumulation and release.
Point 2: P4-L130, the authors selected 1cm/yr and 3cm/yr as the velocity boundary conditions in x and y directions, however, as I know from the previous parts, the authors downscale the geometry of the fault model for computational convenience. How did you obtain or determine these two velocity parameters? Did you consider the scale like the geometry downscale? Please provide more detailed information about them as they are the most important parameters for the simulation.
Response 2: It was decomposed from oblique plate convergence with northward motion at about 50 to 70 mm/yr between Eurasian continental lithosphere and Indian-Australian ocean lithosphere [Fitch, 1972; Jarrard, 1986; Bradley et al., 2017]. Giving the subduction angle is about 70°- 80° with respect to the fault plane, it gives us the average velocity ratio in x and y directions to 1:3 cm/year.
Point 3: The authors simulated the earthquake cycle process, for each phase, in my opinion, a suitable way to determine each process duration by paleoseismic record or references. Please provide the reason for each time duration determination.
Response 3: The determination of the earthquake recurrent interval in this study is taken from probability seismic hazard analysis implemented in Indonesia and other regions. In which several seismic hazard maps are generated by taking into accounts earthquakes with recurrence interval up to 500 and 2500 years (PUSGEN, 2017). We take the 500 years recurrence interval to be simulated in this study and the strain accumulation will analyzed in each half recurrence interval before and after the earthquake, e.g., simulation year 250 and 750, respectively.
Point 4: P1-L33-L35, please check the verb tenses.
Response 4: We have revised the verb tense in the updated manuscript.
“When the frictional strength of the fault is exceeded, the accumulated strain will be released and deforms instantaneously during the coseismic phase [Reid, 1910; Thatcher, 1975, Koseluk and Bischke, 1981; Ziv and Schmittbuhl 2003]”
Point 5: P3-L80, I can’t find the reference Wang [2007] in the last reference part, same as the P3-L107 Turcotte and Schubert [2014], are they cited correctly?
Response 5: Thank you for the correction. We made a mistake in the reference; it should be Wang et al. [2003] and Turcotte and Schubert [2002]
Point 6: P3-109, Table 1. The parameter name should be “Density”?
Response 6: Revised in the updated manuscript.
Point 7: P4-L142, When the authors simulated the coseismic process, they set the steady lower creep rate as 1.5cm/yr. The lower steady creep of the fault causes stress loading or releasing, moreover, impacts the results in this paper, therefore, I suggest the authors give some reasons for it.
Response 7: For the strain release model, the lower portion of the fault slips at a steady rate of 1.5 cm/yr and fully release the accumulated strain at year 500 in the form of the earthquake rupture by 7.5 m slips. In addition, we also ran the model to accommodate partially release the accumulated strain by 80%, 60% and 40% which is represented by the 6m, 4.5m and 3m slips, respectively.
Point 8: P5-P8, the titles of Fig4 and Fig5 are almost the same, please give a more detailed title for each. Same situations are in Fig6 and Fig7.
Response 8: We thanks the reviewer for the comments. We have updated the titles of Fig 5, Fig 6 and Fig7.
Point 9: P5-P8, In Figs 4, 5, 8, the authors show the displacement magnitude or the slip magnitude without unit. The magnitude of displacement and slip are important, they will help decide whether the layer thickness and the frictional strength should be concerned in the models. If they are not significant in the simulation, then this paper will lose its real value. So, please give the correct unit for each figure.
Response 9: Revised in the updated manuscript.
Point 10: P10-L279-284, check the words in “Based on our numerical results, we suspect that the stress accumulation in the central and southern is higher that the northern part”, and “The smaller stress accumulation in the northern part of GSF correlate with the absence of major events (M>7) in the northern part of GSF, even after the devastating M>9 event in 2004.”
Response 10: We have revised the paragraph as
Based on our numerical results, we suspect that the stress accumulation in the central and southern is higher than in the northern part. Therefore, more major events were observed in the central and southern GSF. The small stress accumulation in the northern part of GSF deduced from this study might explain the absence of major events (M>7) in the northern part of GSF, even after the devastating M>9 events in 2004.
Reviewer 2 Report
Dear Authors,
I appreciate the opportunity to revise this manuscript, which I found very well written and partially well illustrated. Nevertheless, before its publication in GeoHazards, the manuscript needs some minor improvements in text and moderate improvements in figures in its present form. The study presents a good and relevant problem well contextualized in the introduction. The scientific gap (Lines 80-82) is entirely relevant and addressed in the manuscript. However, the scientific gap needs to be addressed better. For instance, the scientific gap is expressed in just one sentence (Lines 80-82), which can be improved. I recommend adding a sentence highlighting the importance of considering stress build-up and release with various lithosphere thicknesses and the effect of fault friction force on the lithosphere. That is why the scientific gap in this study is relevant and how it can contribute. Furthermore, at the end of the Introduction Section, I missed a sentence on how the results were reached/addressed (methodology); two sentences about the main results found; and a sentence on the implications of this study.
Some features are called out in the text and given specific importance, but they are not illustrated. Examples of this can be seen in Lines 54-55, 58, 59, and 265-266. It may be obvious to you where these features lie geographically. However, for most readers, this is not obvious and generates uncertainties that make it difficult to understand the results. I recommend that you illustrate such features.
Figures 1 and 10 need some attention. First, decreasing the amount of text in the figure would be interesting. Leave only the points with the offset values. Second, it is critical to insert a more regional location map of this area and add north and scale, as well as geographic quadrants in the coordinates. Finally, insert colors to help the reader better understand the figure. Perhaps using a satellite map, such as those provided by GeoMapApp, will help with the aesthetics of this figure.
The study has very clear and relevant implications. However, a simplification was elaborated in the downscale of the fault segments. I provoke the authors to question whether it would not be important to discuss this point since how this "simplification" might affect the results. I understand this would be understandable, aiming to make the simulations more operational and mitigate the variables addressed in this study. However, this is an important limitation as different fault segments behave differently from a single large segment. Thus, I question the authors if it is not relevant to discuss this and even consider a future work analyzing this point.
The manuscript presents exciting results well explained in the methodology sections. The interpretations of the results are very interesting and supported by the results. Comparisons with previously published data are equally good and undoubtedly denote the contribution of this study. Finally, the conclusions are well-written and fully supported by the results of this study. Overall, these are my general reviews, but I have attached the detailed reviews in a PDF file. Therefore, because of all comments, I recommend some modifications, which can be described as MINOR REVISIONS. Certainly, a relatively large audience can be obtained from it, consistently with the GeoHazards readers.
Author Response
Response to Reviewer 2 Comments
Point 1: The scientific gap (Lines 80-82) needs to be addressed better. For instance, the scientific gap is expressed in just one sentence (Lines 80-82), which can be improved. It also need to add a sentence on how the results were reached/addressed (methodology); two sentences about the main results found; and a sentence on the implications of this study.
Response 1: We have revised the scientific gap in the updated manuscript.
“Given all the consequences of major earthquakes along the shallow strike-slip fault, modeling the lithosphere deformation model is crucial as a proxy to estimate the earthquake occurrence. A few studies of the deformation model have been conducted, such as Savage & Prescott [1978], Wang [2003], and Li et al [2009]. However, these studies have not reviewed the strain buildup and release with various lithosphere thicknesses and the effect of fault frictional strength on the lithosphere. Savage & Prescott [1978] studied a simple two-dimensional earthquake cycle model (strain accumulation, strain release, and postseismic readjustment) in an elastic lithosphere overlying a viscoelastic asthenosphere. Meanwhile, Wang [2003] modeled a subduction earthquake cycle by addressing the loading mechanism, such as the driving forces, the strength of plate boundary faults, and the coupling of long-term tectonic processes and short-term earthquake cycles. Hence, this study uses three-dimensional finite-element modeling to highlight the importance of stress buildup and release with various lithosphere thicknesses and the effect of fault friction force on the lithosphere by adding the velocity boundary conditions based on the decomposed oblique plate convergence. We also evaluate several parameters, such as material and boundary conditions, frictional strength of the fault to the strain buildup, and release along the fault, where the smaller differential stress showed the strain released on the elastic layer on the upper crust. The numerical modeling resulted in this study is used to understand the earthquake pattern observed along the GSF.”
Point 2: Lines 54-55, 58, 59 and 265-266 should be illustrated to explain the features geographically.
Response 2: We thanks the reviewer for the suggestion to add illustration to explain the features geographically. We have updated our Figure 1 and Figure 10 which also illustrated line 54-55, 58, 59 and 265-266.
Figure 1. Slip rates (mm/ yr) distribution across the Great Sumatran Fault (GSF) zone at each segment of the fault. The slip rates are inferred from GPS measurement of previous studies [Bradley et al., 2017; Ito et al., 2012; PUSGEN, 2017; Natawidjaja et al., 2017; Natawidjaja, 2018]. The fault segments refer to National Center for Earthquake Studies of Indonesia [PUSGEN, 2017].
Figure 10. The variation of crustal depth along GSF based on CRUST1.0 model and the M ≥ 7.0 earthquake history since 1892 [Hurukawa et al., 2014].
Point 3: Figures 1 and 10 need some attention. First, decreasing the amount of text in the figure would be interesting. Leave only the points with the offset values. Second, it is critical to insert a more regional location map of this area and add north and scale, as well as geographic quadrants in the coordinates. Finally, insert colors to help the reader better understand the figure. Perhaps using a satellite map, such as those provided by GeoMapApp, will help with the aesthetics of this figure.
Response 3: We have revised the Figure 1 and 10 by decreasing the amount of text, adding the regional location map, north direction, scale, and geographic quadrants in the coordinates, also adding colors to the map. The updated Figure 1 and 10 is shown in the response to the previous question from reviewer.
Point 4: The study has very clear and relevant implications. However, a simplification was elaborated in the downscale of the fault segments. I provoke the authors to question whether it would not be important to discuss this point since how this "simplification" might affect the results. I understand this would be understandable, aiming to make the simulations more operational and mitigate the variables addressed in this study. However, this is an important limitation as different fault segments behave differently from a single large segment. Thus, I question the authors if it is not relevant to discuss this and even consider a future work analyzing this point.
Response 4: Aside from reducing the computational time, the simplification we did for our study was to better understand the lithosphere deformation behavior in general towards the viscoelastic material thickness & frictional strength used in the model. This understanding is also carried out by providing examples of variations in crustal thickness in Sumatra, which are related to the seismic pattern in Sumatra.
