In this manuscript, the dynamic process of a rockslide was evaluated by three-dimensional discrete element method (DEM) combined the discrete fracture network (DFN). This paper is well written. I think it can be accepted via minor revision. I have some comments which may be useful during the revision:

1. The integration of field investigations and numerical simulations is a highly effective research approach. However, it is noted that the submission seems lacking comprehensive comparative analysis between the results obtained from field investigations and numerical simulations.

Response: As the reviewers noted, this study adopts an integrated approach combining field investigations with numerical simulations. Comparative analysis of UAV aerial surveys and historical satellite imagery enables the acquisition of topographic changes and deposit extent before and after the failure of the YXP rockslide. Field-documented fracture characteristics provide critical parameters for numerical modeling, while boundary measurements constrain simulation parameters for failure volume, final deposition range, and travel distance.

Given our objective to determine the landslide's maximum potential runout distance, the initial simulations deliberately excluded existing mitigation structures. However, as demonstrated in Figure 14, subsequent simulations incorporating these engineering measures show strong agreement with field observations.

2.Lines 140~183: it is not necessary to introduce the computing process of the DEM, instead, you may cite related papers.

Response: We appreciate the reviewer's suggestion. We will streamline the manuscript by removing redundant introductory content and supplementing the discussion with relevant literature on DEM.

3.The arrangement of Figure 6 should be rearranged to enhance the clarity of the numbers in Figure 6c.

Response: Thank you for your suggestion. We will list Figure 6c separately to make it clearer.

General comments：

In this submission, on-site inquiries and aerial photography were firstly utilized to study the basic characteristics of rockslides, and then the dynamic process of rockslide and the influences of fractures density and friction angle on kinetic characteristics were studied using the 3D discrete element method with the discrete fracture network. Generally speaking, this submission provides valuable information for assessing the dynamic process from on-site inquiries and numerical simulations, but there are still some issues that need to be revised to be published.

1. In the Introduction section, although there have been significant advancements in field investigations, theoretical models, and numerical simulations related to rockslides, the author should to highlight the innovative and necessary contributions of this submission's research.

Response: Thank you for your suggestion. We will emphasize the innovation and necessary contribution of this study at the end of the introduction.

2. Section 3.2, the bedrock and sliding rock mass were considered to be separate in the initial stage? Is the contact between the bedrock and sliding rock mass limited to frictional forces only? The authors should provide a detailed explanation.

Response: We sincerely appreciate the reviewer’s insightful question, which raises a critical point. We will address this issue from the following perspectives:

• Fixed Bedrock Assumption in DEM Simulation

In the discrete element method (DEM) simulation, all rock masses are inherently discretized. However, since our primary focus is on the kinematic behavior of the sliding mass, we applied a "fix" command to the underlying bedrock to constrain its movement in all directions. This treatment is consistent with real-world conditions, as bedrock typically remains stable during slope failures.

• Mechanical Contrast Between Bedrock and Sliding Mass

The bedrock and sliding mass exhibit intrinsic differences in mechanical properties, with the latter having significantly lower shear strength. This strength disparity naturally exists between the two materials.

• Structural Plane Parameters at the Interface

A structural plane (discontinuity) exists between the sliding mass and the slip surface. The stability of the overlying geomaterials is predominantly governed by the interface parameters, including Cohesion (c), Friction angle (φ), Normal stiffness (Kₙ), Shear stiffness (Kₛ).

These parameters collectively determine the sliding behavior, meaning that friction alone is not the sole controlling factor.

The study investigates the development of rock fractures on mountain ridges in meizoseismal areas, which can lead to fatal rockslides based on site inquiries and aerial photography. A three-dimensional discrete element method (DEM) combined with a discrete fracture network (DFN) was employed to assess the dynamic process of the rockslide. The study analyzed the effects of fracture density and friction angle on the kinetic characteristics of the rockslide. This research provides valuable insights into the kinetic processes and prevention strategies for post-earthquake rockslides in meizoseismal areas. But several misunderstanding are encountered in this paper:

1, In Section Introduction, please focus on scientific issues clearly and work to solve these problems well. This section needs definitely gives the research gap and objective about your work.

Response: We have reorganized the scientific question of this article, mainly on how to combine discrete element method with discrete fracture network for accurate simulation of rock landslides. At the same time, the objectives of this work have been clarified, mainly including the following four aspects: (1) To investigate the failure mechanism and kinematic characteristics of the YXP rockslide;(2) To develop an accurate back-analysis approach for fractured rock slopes by integrating the discrete element method (DEM) with discrete fracture network (DFN) modeling;(3) To quantify the influence of fracture parameters on rockslide dynamic behavior;(4) To evaluate the feasibility of using dam structures for mitigating such rock slope failures.

2, Page 1 line 23 geological disasters or geo-hazards? please check all the text.

Response: Thanks for your suggestions, we will unify the relevant vocabulary as geo-hazards throughout the text.

3, The phenomenon of cracks on the rear edge of the mountain before the landslide is common, whether it is rainfall, earthquakes, or gravity-induced landslide events, tends to form cracks at the trailing edge, but the formation of cracks is often a consequence rather than a cause. In other words, the formation of cracks is often the product of the beginning of the landslide, rather than the existing fracture and the landslide.

Response: Thank you for the reviewer's comments. As the research object of this manuscript is located in a strong earthquake prone mountainous area, the mountains in this region are extremely fragmented. Often, after a collapse occurs, there are still a large number of potentially unstable rock masses at the rear edge of these collapses, which still have the potential for sliding. The research object of this manuscript, YXP rockslide, is a typical case of such landslides.

4, In Section Introduction, the author mentions several times that the type of fracture is important for landslide evolution, and the detection of fractures is also important, but the author does not mention this work in the manuscript.

Response: For the detection of fractures, we mainly use on-site measurement methods to obtain more accurate data. We collected data from over 400 sets of fractures in order to reconstruct the true characteristics of rock fractures.

5, There are many words in the text that the author has not explained, and it is impossible to understand and read. For example, page3 line 83, the author mentions three major faults? What is the meaning?

Response: Thank you for the reviewer's suggestions. We will further standardize the relevant descriptions in this manuscript. The major thrust faults of Wenchuan earthquake include Wenchuan-Maoxian fault, Beichuan-Yingxiu fault and Anxian-Guanxian fault.

6, Figure 1 shows the YXP landslide, while in the text uses the YXP rockslide? Why?

Response: Thank you for the reviewer's reminder. According to the classification of landslide types by Hungr et al. (2014) (Hungr, O., Lerouil, S.,&Picarelli, L. (2014). The Varnes classification of landslide types, an update. Landslides, 11 (2), 167-194.), rockslide also belongs to landslide. In order to unify, the corresponding positions in the entire text have been changed to rockslide.

7, How was the data on the geohazards mentioned in Section 2.1 obtained? Or is it a reference to someone else's?

Response: Thank you for the reviewer's comments. The location, geology, rainfall, and other information in Section 2.1 were obtained through data research, and the evolution process and scale of the disaster were obtained from on-site investigations after the geological disaster occurred.

8, In chapter 2.2, the author directly gives the mechanism of the formation of the geohazard, which is only a simple description, without any data, how does the author make the mechanism clear?

Response: Thank you for the reviewer's comments. In this section, we mainly aim to explain the cause of the landslide through geological deduction methods. We have revised the title of this section to 'Formation mechanism'.

9, The method mentioned by the author in Chapter 3 is the DEM method, which the author uses directly in Rockslide. And how is this data obtained? What is the reason for this value?

Response: First, the acquisition of discrete elements includes obtaining terrain data and fracture parameters, with its core lying in fracture network construction. Through field investigations, we acquired 423 sets of fracture data and identified two dominant joint sets from this dataset via projection analysis. Subsequently, we introduced the Monte Carlo stochastic fracture network method and generated a random joint network primarily based on these two dominant joint sets. Then, we used the generated joint network to discretize the sliding mass, thereby establishing a 3D model that closely approximates the characteristics of real rock masses. Finally, employing the discrete element method (DEM), we defined models and assigned parameters separately for block units and joint units, conducting simulation studies according to different required working conditions.

10, There are two third parts of the text chapter? How to verify the reliability of the simulation, and how to use numerical simulation to reveal the gas process and prevent and control risks?

Response: The title numbering in the main text has been modified. Figure 14 verifies the reliability of the numerical simulation results. At the same time, the numerical simulation results can obtain the distance and range of the landslide movement, as well as the speed of the movement process. If prevention and control engineering is adopted, firstly, we can solve the layout position of the prevention and control engineering. Secondly, the simulation results of the speed of rockfall impact can provide a basis for the energy level of the prevention and control engineering, and the collapse range can provide a basis for the interception height of the prevention and control engineering.

General comments

Very good research on rock mechanics and discrete fracture network. Please, follow my specific comments to improve the manuscript.

Specific comments

1. Lines 64-66. “Discrete Fracture Network (DFN) models are primarily based on the establishment of joint and fracture networks... cannot be fully measured”. Insert recent review papers on the link between field surveys and modelling aspects of Discrete Fracture Network models:

- Medici G, Ling F, Shang J 2023. Review of discrete fracture network characterization for geothermal energy extraction. Frontiers in Earth Science 11, 1328397.

- Kolapo P, Ogunsola NO, Munemo P, Alewi D, Komolafe K, Giwa-Bioku A 2023. DFN: an emerging tool for stoc hastic modelling and geomechanical design. Eng 4(1), 174-205.

Response: Thank you for your suggestion. We will add relevant literature on DEM in the article.

1. Line 79. Specify the 3 to 4 specific objectives of your research by using numbers.

Response: We sincerely appreciate the reviewer's constructive suggestions. In response, we have thoroughly restructured the introduction section. The key research objectives of this study are now clearly outlined as follows:

(1) To investigate the failure mechanism and kinematic characteristics of the YXP rockslide;

(2) To develop an accurate back-analysis approach for fractured rock slopes by integrating the discrete element method (DEM) with discrete fracture network (DFN) modeling;

(3) To quantify the influence of fracture parameters on rockslide dynamic behavior;

(4) To evaluate the feasibility of using dam structures for mitigating such rock slope failures.

1. Line 83. “Three major faults”. Please, specify the type of faults. Normal, strike-slip?

Response: It should be thrust fault.

1. Lines 321-334. The bulletin points are 3 so the specific objectives should be 3 to match.

Response: We have revised the conclusion and aligned it with the specific objectives.

1. Lines 321-334. Add a “take home” message for the researchers working in your field.

Response: We have reorganized the conclusion section to make it more concise.

1. Lines 350-442. Please, integrate relevant literature on DFN.

Response: Based on the revised introduction, we will modify the corresponding references.

1. Figures and tables

Figure 1. Make the figure larger.

Figure 6c. The stereonet that shows the two sets of fractures should be much larger.

Figure 6c. Consider to make it a separate figure.

Figure 14. Fractures are very difficult to see. Increase the graphic resolution.

Figure 14. Make also the figure larger, there is room for this change.

Response: We have enlarged Figure 1 and created a separate image for Figure 6c. We have increased the resolution of Figure 14 and enlarged it to 12cm.

This manuscript describes field investigations and a numerical modeling study of a catastrophic post-earthquake rockslide in Wenchuan, China. On-site investigations and aerial imagery were utilized to characterize the landslide. A 3D discrete element method, integrated with a discrete fracture network, was employed to evaluate the dynamic processes of the rockslide. The research analyzed how fracture density and friction angle influenced the kinetic properties of the landslide.

Although this manuscript investigates an important landslide with advanced numerical modeling techniques, the scientific methods are not robust and the presented results and corresponding conclusions should be improved prior to publication. This paper provides some valuable insights into the parameterization of 3D discrete element models but should address several major issues as described below.

1. The main objectives of this manuscript should be clarified. The introduction begins by discussing the formation of ridge-top fractures during earthquakes as an important conditioning factor for subsequent slope stability. The authors don’t provide any background information or substantiation that supports the studied landslide being related to the formation of cracks during the Wenchuan Earthquake, which occurred six years prior to the landslide failure. This should be discussed in section 2.1 – Geological Setting. In addition, the modeling investigations described in this paper do not simulate such processes (e.g., by simulating earthquake-induced damage), and it is not clear how this study contributes to a better understanding of earthquake-induced slope damage and the potential for subsequent landslides.

Response: Thank you very much for the reviewers' comments. The primary objective of this manuscript is to employ a coupled discrete element and discrete fracture network method to more accurately invert and reveal the instability mechanism and dynamic motion process of the XGJ rockslide. Finally, the main factors influencing the post-failure movement characteristics of the landslide are discussed, including fracture strength, density, and obstacles. 

Regarding the geological background mentioned by the reviewers, firstly, this region is located in the Longmen Shan seismic zone on the eastern margin of the Tibetan Plateau. Additionally, we have multi-temporal historical images for comparison, which demonstrate that the failure area has been continuously expanding. Moreover, there are many similar landslides in this area triggered by the Wenchuan earthquake that can serve as references (e.g., Zhang, Y., Chen, G., Zheng, L., Li, Y., & Wu, J. (2013). Effects of near-fault seismic loadings on run-out of large-scale landslide: a case study. Engineering Geology, 166, 216-236.). Therefore, we believe that the rear edge of this rockslide was generated by seismic activity, and the fractures have continued to propagate backward after the earthquake.

1. Geologic Setting – The following points in this section should be clarified. Lines 83–88 please provide citations for descriptions of the earthquake, site geology, and regional climate. Please provide additional detail on how the total landslide area and thickness were measured. Is the reported thickness of 5.8 m in Line 96 measured from the thickness of depletion in the source area or thickness of deposition? Please provide evidence or include references to substantiate the statement in Line 105 that “the landslide area was seriously broken” during the earthquake. What evidence do you have that these fractures did not exist prior to the earthquake? Please clarify the description of the failure plane in Line 112. Does the use of “clear” mean fully evacuated? What does “smooth” refer to?

Response: The total area of the landslide was measured using satellite imagery, and the average thickness of approximately 5.8 m was obtained by performing a differential analysis between the UAV aerial elevation model and ALOS satellite elevation data. The failure conditions of the landslide area during the earthquake can be illustrated using historical images. The description of the failure plane in Line 112 has been modified to: "…the obvious sliding surface located at the rear edge of the YXP landslide was observed…"

1. Methods – The authors provide a detailed description of the equations used in the DEM formulation. However, there are several important aspects of the modeling methodology that are absent from the text. The authors choose to model bedrock and the moving landslide mass with different material properties without providing justification for this choice. In addition, the constitutive joint model is not specified, and the choice of material properties and joint strength properties are not substantiated. One reference is provided, but the assigned values are different than those used in this previous work. In addition, there is no reference or citation to the software package used when describing the equations used by the DEM software (e.g., Itasca Consulting Inc.), rather the authors cite their own prior work.

Response: Thank you for the reviewers' suggestions. In this study, certain simplifications were applied to the sliding body and sliding bed. Specifically, since the sliding bed remains stationary during motion and primarily interacts through its contact surface with the sliding body, we assigned a simple elastic model to the sliding bed in the modeling process. For the sliding body, however, we adopted the commonly used Mohr-Coulomb model. As for the internal fractures within the rock mass, the Mohr-Coulomb joint model was employed. This model is based on the Mohr-Coulomb strength criterion, enabling the assessment of joint shear failure and making it suitable for simulating joint slippage in rock mass stability analysis. Although we referenced parameters from similar projects, each project has its unique characteristics, making it impractical to adopt identical parameters outright. Therefore, a certain degree of trial calculation was necessary. Regarding literature citations, we will supplement the references with the DEM software manual and other relevant literature.

1. Results– The authors report findings from a field investigation and conduct numerical modeling of the landslide. However, model results are not contextualized or compared with landslide conditions (e.g., deposit extent or thickness) and therefore it is unclear whether the numerical results are able to accurately reproduce this event.

Response: Based on the suggestions of the reviewers, we will supplement the actual plan range figure and accumulation profile figure of the landslide after instability, and compare them with the actual situation on site to verify the accuracy of the numerical simulation results.

1. Discussion – As written, this section presents further results which describe a parameter study on the impact of friction and angle and joint density on the model results. Differences in each model run are presented, but it is not clear which set of parameters best replicated the observed landslide features. While this is an interesting experiment, it is not meaningful in the context of this case study without further explanation and a comparison to the actual site conditions. It is not clear if parametrization for any of the other input material properties was conducted or how changes to these properties may impact model results. In addition, the presentation of results related to a check dam are not substantiated.

Response: Thank the reviewers for their insightful suggestions. Indeed, in numerical simulations, many parameters may have varying degrees of impact on the simulation results, but these factors could be divided into controlling factors and general factors. Through literature research and preliminary simulation experiments, we believe that for rockslides with developed fractures, the controlling factors are the parameters related to the fracture surface, among which the inner friction angle of the fracture surface and the density of fracture are particularly important. Therefore, choosing these two factors for comparative analysis, further research is needed on other factors in the future. The discussion on the use of blocking dams for debris flow is based on the fact that conventional blocking measures are difficult to control such high-position rockslide disasters. Therefore, protection was chosen at the end of the movement path and compared with the actual situation (Figure 14c).

1. The conclusions presented are valid observations from the modeling study but are not substantial in relation to the studied landslide or implications for similar future hazards. Conclusion #1 regarding the failure for the landslide under high pore-water pressure is not substantiated by the results of this study.

Response: We have revised the conclusions of this paper, with particular emphasis on replacing the second and third conclusions: (2) By combining the discrete element method with discrete fracture networks, it is possible to accurately simulate and back-analyze the instability and movement processes of fractured rock landslides.  (3) The internal friction angle of fractures and fracture density significantly influence the movement velocity and deposit extent of rock landslides. A higher fracture density leads to faster instability of rock blocks, but it has no pronounced effect on travel distance.

1. There are many places throughout the text where additional references are needed (see comment #2) or where the cited references are not relevant, for example see references in Line 40 to Evans et al., 2001 and Lines 52–54 to Mauka et al., 2017, or are missing from the reference list (e.g., Line 108 to Huang et al., 2014).

Response: Thank you for the reviewer's reminder. We will carefully check all the references to ensure that there are no errors in citations or omissions.