A Study of Small-scale Landslides Susceptibility Dynamic Assessment in South Chian: Merging SBAS-InSAR and Machine Learning

Liang, Zhu; Hou, Jingxin; Liu, Yang; Wang, Ting; Liu, Guochao; Si, Chunshuai; Wu, Jun

doi:10.5194/egusphere-2026-218

Preprints

https://doi.org/10.5194/egusphere-2026-218

Preprints

23 Feb 2026

| 23 Feb 2026

A Study of Small-scale Landslides Susceptibility Dynamic Assessment in South Chian: Merging SBAS-InSAR and Machine Learning

Zhu Liang, Jingxin Hou, Yang Liu, Ting Wang, Guochao Liu, Chunshuai Si, and Jun Wu

Abstract. Shallow landslides, triggered by factors such as heavy rainfall and human engineering activities, are characterized by their sudden occurrence, wide distribution, and small scale, which pose significant threats to human life, especially in southern China. Landslide susceptibility assessment (LSA) is crucial for disaster prevention, mitigation, and land-use planning. Traditional assessment methods, such as field surveys and statistical models, predominantly depend on static geological environment factors, which exhibit inherent limitations in capturing spatiotemporal information on dynamic surface deformation. This deficiency directly leads to a lack of timeliness in hazard assessment, which is particularly important for sudden disasters. In this study, small Baseline Subset (SBAS – InSAR) and machine learning were married to explore a dynamic method in LSA, and the Conghua district in Guangzhou was selected as the study area. First, a dataset consisting of 326 historical landslides and 10 static factors, and 2 dynamic factors in the area was prepared. Then, the dataset was divided into two parts, one for modeling training (80 %) and the other for testing (20 %). Third, factors and samples were involved in the modeling of Random Forest (RF), Light Gradient Boosting Machine (LightGBM), and Extreme Gradient Boosting (XGBoost). Finally, the performance of these models was validated through the area under the curve (AUC) and compared with the models analyzed by 10 static factors only. The results show that the XGBoost model exhibits the best performance, with an AUC of 0.933, and the models considered dynamic factors all performed better than that of static factors only. This study demonstrates that merging SBAS-InSAR and machine learning can provide a reliable technical approach for dynamic assessment of small-scale LSA, which is of great significance for targeted disaster prevention and mitigation in rural and hilly areas of South China.

Received: 16 Jan 2026 – Discussion started: 23 Feb 2026

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.

Download & links

Zhu Liang, Jingxin Hou, Yang Liu, Ting Wang, Guochao Liu, Chunshuai Si, and Jun Wu

Status: final response (author comments only)

RC1:
'Comment on egusphere-2026-218', Anonymous Referee #1, 15 Mar 2026
This manuscript presents an approach to landslide susceptibility assessment that integrates SBAS-InSAR-derived surface deformation data with tree-based ensemble machine learning models (RF, LightGBM, XGBoost) in the Conghua district of Guangzhou, southern China. The central premise that dynamic deformation factors can improve upon static-factor-only susceptibility models is reasonable and addresses a recognized gap in the landslide susceptibility literature. However, the manuscript suffers from several methodological shortcomings and presentation issues that substantially weaken the reliability of the reported findings. Major revisions are required before this work can be considered for publication.
Major Comments:
The manuscript provides no information whatsoever on how non-landslide samples were generated. This is a critical omission. The selection of negative samples is known to significantly influence the performance and spatial predictions of machine learning-based susceptibility models (e.g., Dou et al., 2023, which the authors themselves cite; and https://doi.org/10.1007/s10346-020-01473-9; https://doi.org/10.3390/rs15123200). Were non-landslide points selected randomly across the study area? Were buffer zones applied around known landslide locations to avoid spatial contamination? Was any consideration given to ensuring geomorphological plausibility of the negative samples? Without this information, it is impossible to evaluate whether the reported AUC values and accuracy metrics are meaningful or artificially inflated. For instance, if non-landslide points were drawn predominantly from flat, low-elevation terrain far from faults and roads, the models would trivially learn to separate landslide from non-landslide locations based on topographic setting alone, inflating performance metrics without reflecting genuine predictive skill. The authors must clearly describe the negative sampling protocol and ideally test sensitivity to alternative sampling strategies.

The authors rely on a single random train-test split (stated as 80/20 in the abstract but 70/30 in Section 3.2.1, see inconsistency comment below) to evaluate model performance. A single hold-out partition is highly sensitive to the particular random split and does not provide a reliable estimate of model generalizability. The landslide susceptibility modeling community has broadly adopted k-fold cross-validation (or spatial cross-validation) as a minimum standard for performance evaluation. The absence of any cross-validation scheme means the reported accuracy metrics may not be reproducible under alternative data partitions. Furthermore, spatial autocorrelation between training and testing samples is not addressed. If landslide and non-landslide points in the training and test sets are spatially proximate, the models may benefit from information leakage rather than learning generalizable patterns. The authors should implement at a minimum stratified k-fold cross-validation (e.g., 5-fold or 10-fold), and ideally spatial cross-validation to account for spatial dependence.

The abstract states the dataset was divided into 80% training and 20% testing, whereas Section 3.2.1 states 70% training and 30% validation. The authors must clarify and reconcile this discrepancy.

The abstract and Section 3.2.2 refer to 326 historical landslides and 10 static factors, while Section 2.1 reports 361 landslides and Section 3.2.1 also states 361 samples. The number of static factors listed in Table 3 is 9 (slope, aspect, elevation, distance to faults, distance to rivers, distance to roads, profile curvature, maximum elevation difference, NDVI), not 10 as claimed. These numerical discrepancies must be resolved.

The manuscript provides no information on the hyperparameters used for RF, LightGBM, or XGBoost, nor on whether any hyperparameter optimization was performed (e.g., grid search, random search, Bayesian optimization). Machine learning model performance is highly sensitive to hyperparameter settings, and the absence of this information undermines reproducibility. The authors should report all key hyperparameters and describe the tuning procedure.

The authors frame their approach as a "dynamic" susceptibility assessment method. However, the SBAS-InSAR deformation data are summarized as a single cumulative deformation value and a single average deformation rate over the entire two-year monitoring period (Jan 2024–Dec 2025). This is, in effect, a static representation of a dynamic process. A truly dynamic assessment would involve time-varying susceptibility maps that update as new deformation data become available. For instance, producing monthly or seasonal susceptibility maps that reflect evolving deformation patterns. As currently implemented, the InSAR-derived variables are simply two additional static predictors appended to the existing feature set. The authors should moderate their claims accordingly or demonstrate a genuinely time-varying assessment framework.

The InSAR methodology section (3.1) presents textbook equations but omits essential processing parameters: coherence thresholds, unwrapping algorithm and parameters, atmospheric phase screen correction method, reference point selection, multi-looking factors, and filtering approach. These details are critical for evaluating the quality and reliability of the derived deformation products, particularly in a heavily vegetated subtropical environment where decorrelation is a major concern.

The SBAS-InSAR processing in Section 3.1 extracts only line-of-sight deformation, which represents a one-dimensional projection of the true three-dimensional displacement vector. The manuscript does not specify whether ascending, descending, or both orbital geometries were used. This is a significant limitation, particularly given the authors' own observation (Section 4.3) that high-deformation zones do not fully correspond to high-susceptibility areas and that some highly unstable areas occur in residential plains, a pattern likely reflecting anthropogenic vertical subsidence rather than slope-related displacement. Three-dimensional displacement decomposition, achievable by fusing ascending and descending InSAR observations, would help disentangle landslide-related motion from unrelated ground deformation processes such as consolidation or groundwater extraction. Moreover, decomposed horizontal and vertical displacement components could serve as more physically meaningful and discriminative conditioning factors for the machine learning models than a single LOS measurement, potentially reducing the misclassification issues evident in the current results. Since Sentinel-1 data are freely available in both orbital geometries over the study area, the authors should at minimum acknowledge the limitations of single-geometry LOS measurements and discuss how three-dimensional displacement retrieval could strengthen the proposed framework. Relevant references include, for instance, doi:10.1016/j.earscirev.2014.02.005, doi:10.3390/rs11030241, and doi:10.1016/j.geoai.2026.100061.

Minor Comments:
The title contains a typo: "South Chian" should be "South China."

Table 5 labels the last row as "ROC" rather than "AUC." The ROC is the curve itself; AUC is the scalar metric.

The claim that overall accuracy improved by 15–20% relative to traditional static methods (line 385) is not supported by the reported results. The largest AUC difference between a static-only and SAR-integrated model on the validation set is 0.933 − 0.882 = 0.051 (XGBoost), which corresponds to roughly a 5 percentage-point improvement, not 15–20%.

The reference list contains formatting inconsistencies, and some citations in the text do not match reference list entries. For example, "Breiman et al., 2021" is cited in the text (line 192), but the reference list entry is Breiman (2001). The references should be carefully checked for accuracy.
Citation: https://doi.org/10.5194/egusphere-2026-218-RC1
- AC1:
  'Reply on RC1', Zhu Liang, 07 May 2026
  This manuscript presents an approach to landslide susceptibility assessment that integrates SBAS-InSAR-derived surface deformation data with tree-based ensemble machine learning models (RF, LightGBM, XGBoost) in the Conghua district of Guangzhou, southern China. The central premise that dynamic deformation factors can improve upon static-factor-only susceptibility models is reasonable and addresses a recognized gap in the landslide susceptibility literature. However, the manuscript suffers from several methodological shortcomings and presentation issues that substantially weaken the reliability of the reported findings. Major revisions are required before this work can be considered for publication.
  Respond: Thank you for your kindly work and recognition for our study. We would reply to the comments one by one.
  Major Comments:
  The manuscript provides no information whatsoever on how non-landslide samples were generated. This is a critical omission. The selection of negative samples is known to significantly influence the performance and spatial predictions of machine learning-based susceptibility models (e.g., Dou et al., 2023, which the authors themselves cite; and https://doi.org/10.1007/s10346-020-01473-9; https://doi.org/10.3390/rs15123200). Were non-landslide points selected randomly across the study area? Were buffer zones applied around known landslide locations to avoid spatial contamination? Was any consideration given to ensuring geomorphological plausibility of the negative samples? Without this information, it is impossible to evaluate whether the reported AUC values and accuracy metrics are meaningful or artificially inflated. For instance, if non-landslide points were drawn predominantly from flat, low-elevation terrain far from faults and roads, the models would trivially learn to separate landslide from non-landslide locations based on topographic setting alone, inflating performance metrics without reflecting genuine predictive skill. The authors must clearly describe the negative sampling protocol and ideally test sensitivity to alternative sampling strategies.
  
  Respond: Thank you for raising this critical point, which is indeed central to the validity of landslide susceptibility modeling. We apologize for the lack of clarity in the initial manuscript and have now thoroughly revised the "Study Area and Data" section (Section 3.2, lines 145–160) to explicitly detail our negative sampling strategy.
  The authors rely on a single random train-test split (stated as 80/20 in the abstract but 70/30 in Section 3.2.1, see inconsistency comment below) to evaluate model performance. A single hold-out partition is highly sensitive to the particular random split and does not provide a reliable estimate of model generalizability. The landslide susceptibility modeling community has broadly adopted k-fold cross-validation (or spatial cross-validation) as a minimum standard for performance evaluation. The absence of any cross-validation scheme means the reported accuracy metrics may not be reproducible under alternative data partitions. Furthermore, spatial autocorrelation between training and testing samples is not addressed. If landslide and non-landslide points in the training and test sets are spatially proximate, the models may benefit from information leakage rather than learning generalizable patterns. The authors should implement at a minimum stratified k-fold cross-validation (e.g., 5-fold or 10-fold), and ideally spatial cross-validation to account for spatial dependence.
  
  Respond: It is true that a more reliable estimation of the model performance is essential. We have applied 10-fold cross-validation to account for spatial dependence, on page 7 line, page line
  The abstract states the dataset was divided into 80% training and 20% testing, whereas Section 3.2.1 states 70% training and 30% validation. The authors must clarify and reconcile this discrepancy.
  
  Respond: It should be 80% training and 20% testing. We have revised the manuscript on page 1, line
  The abstract and Section 3.2.2 refer to 326 historical landslides and 10 static factors, while Section 2.1 reports 361 landslides and Section 3.2.1 also states 361 samples. The number of static factors listed in Table 3 is 9 (slope, aspect, elevation, distance to faults, distance to rivers, distance to roads, profile curvature, maximum elevation difference, NDVI), not 10 as claimed. These numerical discrepancies must be resolved.
  
  Respond: It should 361 landslides and 9 static factors. We have revised it on page line
  The manuscript provides no information on the hyperparameters used for RF, LightGBM, or XGBoost, nor on whether any hyperparameter optimization was performed (e.g., grid search, random search, Bayesian optimization). Machine learning model performance is highly sensitive to hyperparameter settings, and the absence of this information undermines reproducibility. The authors should report all key hyperparameters and describe the tuning procedure.
  
  Respond: We have supplemented the relevant explanations on page 7, line.
  The authors frame their approach as a "dynamic" susceptibility assessment method. However, the SBAS-InSAR deformation data are summarized as a single cumulative deformation value and a single average deformation rate over the entire two-year monitoring period (Jan 2024–Dec 2025). This is, in effect, a static representation of a dynamic process. A truly dynamic assessment would involve time-varying susceptibility maps that update as new deformation data become available. For instance, producing monthly or seasonal susceptibility maps that reflect evolving deformation patterns. As currently implemented, the InSAR-derived variables are simply two additional static predictors appended to the existing feature set. The authors should moderate their claims accordingly or demonstrate a genuinely time-varying assessment framework.
  
  Respond: We have adjusted the expression of “dynamic susceptibility assessment” throughout the manuscript to avoid overstatement. We now clearly describe our method as a hybrid static-dynamic evaluation framework that combines multi-temporal InSAR deformation features with traditional static geo-environmental factors. Besides, we have updated the InSAR data processing from single-period statistics to time-varying analysis on page , line
  The InSAR methodology section (3.1) presents textbook equations but omits essential processing parameters: coherence thresholds, unwrapping algorithm and parameters, atmospheric phase screen correction method, reference point selection, multi-looking factors, and filtering approach. These details are critical for evaluating the quality and reliability of the derived deformation products, particularly in a heavily vegetated subtropical environment where decorrelation is a major concern.
  
  Respond: We fully agree that comprehensive processing parameters are essential for ensuring reproducibility and evaluating the reliability of deformation results, especially in subtropical regions with dense vegetation cover. In response to this concern, we have substantially expanded Section 3.1 to include all critical SBAS-InSAR processing parameters. These parameters were carefully selected and optimized based on extensive preliminary experiments adapted to the environmental conditions of our study area (dense vegetation, high humidity, and complex topography). page, line
  The SBAS-InSAR processing in Section 3.1 extracts only line-of-sight deformation, which represents a one-dimensional projection of the true three-dimensional displacement vector. The manuscript does not specify whether ascending, descending, or both orbital geometries were used. This is a significant limitation, particularly given the authors' own observation (Section 4.3) that high-deformation zones do not fully correspond to high-susceptibility areas and that some highly unstable areas occur in residential plains, a pattern likely reflecting anthropogenic vertical subsidence rather than slope-related displacement. Three-dimensional displacement decomposition, achievable by fusing ascending and descending InSAR observations, would help disentangle landslide-related motion from unrelated ground deformation processes such as consolidation or groundwater extraction. Moreover, decomposed horizontal and vertical displacement components could serve as more physically meaningful and discriminative conditioning factors for the machine learning models than a single LOS measurement, potentially reducing the misclassification issues evident in the current results. Since Sentinel-1 data are freely available in both orbital geometries over the study area, the authors should at minimum acknowledge the limitations of single-geometry LOS measurements and discuss how three-dimensional displacement retrieval could strengthen the proposed framework. Relevant references include, for instance, doi:10.1016/j.earscirev.2014.02.005, doi:10.3390/rs11030241, and doi:10.1016/j.geoai.2026.100061.
  
  Respond: We agree that 3D displacement decomposition can provide more complete and physically meaningful information for landslide susceptibility assessment, and we greatly appreciate the valuable references suggested.
  Our response addresses three main aspects:
  Clarification of orbital geometry (Table 1):
  
  We have added a new row “Orbit: Ascending” in Table 1 to explicitly indicate that only Sentinel‑1 ascending‑orbit data were used in this study.
  Analysis of deformation-susceptibility mismatch (Section 4.3):
  
  We have supplemented the discussion in Section 4.3 to explain the observed inconsistency between high deformation zones and high landslide susceptibility areas. Page , line
  Acknowledgment of LOS limitations and future improvements (Section 5.2):
  
  We have supplemented a detailed discussion in Section 5.2 to acknowledge the limitations of single-orbit LOS measurements and the potential value of 3D displacement decomposition. The three relevant references suggested by the reviewer have also been incorporated. Page , line
  
  Minor Comments:
  The title contains a typo: "South Chian" should be "South China."
  
  Respond: We have revised the mistake.
  Table 5 labels the last row as "ROC" rather than "AUC." The ROC is the curve itself; AUC is the scalar metric.
  
  Respond: We have revised the mistake.
  The claim that overall accuracy improved by 15–20% relative to traditional static methods (line 385) is not supported by the reported results. The largest AUC difference between a static-only and SAR-integrated model on the validation set is 0.933 − 0.882 = 0.051 (XGBoost), which corresponds to roughly a 5 percentage-point improvement, not 15–20%.
  
  Respond: We have revised the claim on page, line.
  The reference list contains formatting inconsistencies, and some citations in the text do not match reference list entries. For example, "Breiman et al., 2021" is cited in the text (line 192), but the reference list entry is Breiman (2001). The references should be carefully checked for accuracy.
  
  Respond: We have revised the reference list
  
  Citation: https://doi.org/10.5194/egusphere-2026-218-AC1
RC2:
'Comment on egusphere-2026-218', Anonymous Referee #2, 16 Apr 2026

Dear Authors,
I carefully read your manuscript on the use of InSAR data to improve landslide susceptibility assessment.
Although the idea is interesting, the topic should be addressed in a more rigorous and appropriate way.
You used Sentinel data, claiming that you can detect antecedent movements of slopes affected by small and shallow landslides, but you did not prove it. You used only two deformation maps calculated over a 2-year period, without any verification of the temporal correlation between slope deformations (if any) and landslide triggering. the deformation rate appears to be a relevant factor in your machine learning models, but no clear explanation is provided on how it is incorporated. Based solely on the manuscript, it is also possible that the models identify slow deformation rates as indicative of high susceptibility. Models set up is not well described and the reported AUC values can suggest some overfitting issues, furthermore, you presented only some statistics of the results, rather then the results themselves (what about the number of TP, TN, FP, FN?)
I am somehow doubtful on the possibility of a correlation between long-term slope deformation and shallow-landslide, since the latter are often triggered by out-of-scale events (extreme rainfall, earthquackes), that are not detectable by SAR data. Moreover, even if you recognized the influence of rainfall in the triggering of the landslide, you did not analyze it.
You did not describe the parameters to create the deformation maps: did you consider any coherence thresholds?
Overall,your claims are not supported by the results your presented and the method and data you used are not described in enoughn detail to sggest the manuscript for publication of further reviewing
Detaild comments are rpvided in the attached file.

Citation: https://doi.org/10.5194/egusphere-2026-218-RC2
- AC2: 'Reply on RC2', Zhu Liang, 07 May 2026
  
  Dear reviewers:
  Thank you for your kindly work of our study and the comment you made has been of great help to our research. We now reply the comment one by one:
  1.I carefully read your manuscript on the use of InSAR data to improve landslide susceptibility assessment. Although the idea is interesting, the topic should be addressed in a more rigorous and appropriate way.
  Respond: Thank you for your kindly work and recognition for our study. We would reply to the comments one by one.
  
  2.You used Sentinel data, claiming that you can detect antecedent movements of slopes affected by small and shallow landslides, but you did not prove it. You used only two deformation maps calculated over a 2-year period, without any verification of the temporal correlation between slope deformations (if any) and landslide triggering. the deformation rate appears to be a relevant factor in your machine learning models, but no clear explanation is provided on how it is incorporated. Based solely on the manuscript, it is also possible that the models identify slow deformation rates as indicative of high susceptibility.
  Respond: Thank you for this critical feedback, which has helped us significantly strengthen the methodological rigor and clarity of our manuscript. We apologize for the insufficient detail and validation regarding the Sentinel-derived deformation data and its integration into the machine learning (ML) model. We have thoroughly revised the relevant sections (page18, line367~390) to address your concerns.
  
  3.Models set up is not well described and the reported AUC values can suggest some overfitting issues, furthermore, you presented only some statistics of the results, rather then the results themselves (what about the number of TP, TN, FP, FN?)
  Respond: Thank you for this constructive feedback, which has helped us improve the transparency and rigor of our model evaluation. We have added relevant explanations on page 16, line 311~368.
  
  4.I am somehow doubtful on the possibility of a correlation between long-term slope deformation and shallow-landslide, since the latter are often triggered by out-of-scale events (extreme rainfall, earthquackes), that are not detectable by SAR data. Moreover, even if you recognized the influence of rainfall in the triggering of the landslide, you did not analyze it.
  Respond: We have adjusted the expression of “dynamic susceptibility assessment” throughout the manuscript to avoid overstatement. We now clearly describe our method as a hybrid static-dynamic evaluation framework that combines multi-temporal InSAR deformation features with traditional static geo-environmental factors. Besides, we have updated the InSAR data processing from single-period statistics to time-varying analysis on page 18~19 , line 366-389.
  
  5.You did not describe the parameters to create the deformation maps: did you consider any coherence thresholds?
  Respond: We fully agree that comprehensive processing parameters are essential for ensuring reproducibility and evaluating the reliability of deformation results, especially in subtropical regions with dense vegetation cover. In response to this concern, we have substantially expanded Section 3.1 to include all critical SBAS-InSAR processing parameters. These parameters were carefully selected and optimized based on extensive preliminary experiments adapted to the environmental conditions of our study area (dense vegetation, high humidity, and complex topography). page 8-9, line 169-197.
  
  6.Detaild comments are provided in the attached file:
  (1) references needed
  Respond: We have added the references on page 2, line 45.
  (2) This is not a static influencing factor, moreover talking about influencing factors is misleading. In LSA you should distinguish between predisposing, preparatory and triggering factors. I suggest you a literature review on the topic. A few examples:
  https://doi.org/10.1016/j.jag.2025.104365
  https://doi.org/10.5194/nhess-23-2229-2023, 2023
  https://doi.org/10.3390/ijgi8020094
  Respond: It is true that rainfall should not be the static influencing factors. We have revised it.
  (3) small scale landslides usually do not have any acceleration phase in the previous months, since they are often linked to short and intense rainfalls (check the works from Guzzeti, Gariani, Peruccacci and others in the topic, you will see that many small and shallow landslide are triggerd by rainfall with durations of few hours). This sentence is also in cotrast with row 38, where you wrote that they are "highlty suddend and difficult to monitor". Do you believe that the landslides you showed in fig. 2 had some precursors in the months before? Finally, the reference is not relevant with your statement, since it describe an instruments and limited tests, cannot be used to generalize anything.
  Respond: We have replaced small scale landslides with landslides and we have adjusted the expression of “dynamic susceptibility assessment” throughout the manuscript to avoid overstatement.
  (4) a lot of confusion in here. InSAR decorrelation on vegetated areas depends upon multiple factors, such as vegetation type, density, height, season, SAR band, etc. Moreover you are comparing distributed scaterrers with permanent scatterers, which are differents, independently from the processing method. I suggest you to check the differences between PS and DS, and the processing method (SBAS, DS-InSAR, PS-InSAR, PSP, etc).
  Respond: We have revised it on page 2, line 64-70.
  (5) references needed.
  Respond: We have added the references on page 3, line 84.
  (6) if there are human causes, it is difficult to predict where a landslide can occur, unless you have access to data about future contruction projects
  Respond: Short-term heavy rainfall and prolonged drizzle were the major factors. page 3, line 84~85.
  (7) distance from road is a tricky paramenter, it ca add a bias to the results, since landslides along the roads are more easily mapped rather the landslides in remote areas. Did you verify the influence of this parameter in you results?
  Respond: It is true that landslides along the roads are more easily mapped. The samples were collected from the local geological disaster survey reports, field surveys, drone investigation and remote sensing image interpretation. We have also explored the relative importance of different factors.
  (8) you described the volume of the landslide, but it would be helpful to add also their area, since some of them are very small and I believe their size is lower than the resolution of Sentinel images.
  Respond: We have added information about the landslide area and the remote sensing interpretation images. page 3, line 86.
  
  (9) revisiting time of Sentinel is about 12 days, hence you can have missing data in the days before the landslides. how did you manage this issue?
  Respond: Regarding the time of the disaster, we obtained this information through on-site investigations and interviews. page 9, line 202.
  
  (10) 80-20 in the abstract
  Respond: We have revised it.
  (11) pleas add measurements units to the legends
  Respond: We have revised it. page 11, line 230.
  (12) usually a coherence threshold is applied to remove noisy data. What treshold did you use?
  Respond: We have added relevant explanations on page 8, line 171-184.
  
  (13) subsidence along the slopes? can you confirm it?
  Respond: It should be slope deformation.
  (14) Subsidence is a very specific type of ground deformation. Did you verify it?
  Respond: It should be deformation.
  (15) hazard or risk?
  Respond: Risk.
  (16) you have to add also the total number of TP, FP, TN and FN, since a high number of TN can give good statistics, even with a low number of TP or a high number of FP
  Respond: We have added relevant explanations on page 16, line313~329.
  (17) can you better explain how you integrated SAR data in the LSA? Did you use the velocities as conditiong factors?
  Respond: 2 dynamic factors interpreted by SBAS – InSAR as conditioning factors.
  (17) these numbers suggest me the presence of some overfitting issues. How did you verified the absence of overfitting?
  Respond: A more reliable estimation of the model performance is essential. We have applied 5-fold cross-validation to account for spatial dependence, on page 9, line 203-207.
  (18) this is the fist time you write about slope units. How did you define them, which appraoch have you used? what is their size distribution? how did you assing the values of the conditioning factors to them? you need to explain these points
  Respond: We have added relevant explanations on page 9, line 215-219.
  (19) did you use mean values or something else? please clarify
  Respond: The mean value in the unit was counted as the representative value of the unit. on page 9, line 219.
  (20) add all the acronyms in table 3. what is DTS?
  Respond: We have added all the acronyms in table 3.
  (21) before you wrote that high deformation rate are located in plain areas. I suggest you to use XAI methods, like partial dependence plots, to verify the correlation between DR and LSI.
  Respond: We have revised it on page 18, line 364-375.
  (22) I suggest to replace with "actually moving". The presence of movement can be due also to other causes, like soil creep. To prove the effectiveness of SAR data with small landslides, you should at least try to correlate landslide occurences with SAR velocities
  Respond: We have replaced it and added relevant content.
  (23) UC increase (e.g. 0.882 → 0.933) does not necessarily correspond to a 15–20% improvement. You should use ap proper method to quantify this improvement
  Respond: We have revised it on page 22, line 458-460.
  (24) in oyur work you have simply verified that your models consider deformation rates somhow relevant for LSA, but without any explanation or verification. In teory, by your results solely, it is also possible that your models identify low deformation rates as conditioning factor for landslide identification
  Respond: We have adjusted the expression of “dynamic susceptibility assessment” throughout the manuscript to avoid overstatement. We now clearly describe our method as a hybrid static-dynamic evaluation framework that combines multi-temporal InSAR deformation features with traditional static geo-environmental factors. Besides, we have updated the InSAR data processing from single-period statistics to time-varying analysis on page 18~19 , line 366-389.
  (25) you did not demonstrate it.
  Respond: We have adjusted the expression of “dynamic susceptibility assessment” throughout the manuscript to avoid overstatement. We now clearly describe our method as a hybrid static-dynamic evaluation framework that combines multi-temporal InSAR deformation features with traditional static geo-environmental factors. Besides, we have updated the InSAR data processing from single-period statistics to time-varying analysis on page 18~19 , line 366-389.
  (26) all these methods use the same data, how can you extract different features?
  Respond: We have revised it on page 23, line 492-496.
  Thanks again.
  Yours,
  Zhu
  
  Citation: https://doi.org/10.5194/egusphere-2026-218-AC2

Zhu Liang, Jingxin Hou, Yang Liu, Ting Wang, Guochao Liu, Chunshuai Si, and Jun Wu

Viewed

Total article views: 1,262 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
828	362	72	1,262	74	72

HTML: 828
PDF: 362
XML: 72
Total: 1,262
BibTeX: 74
EndNote: 72

Views and downloads (calculated since 23 Feb 2026)

Month	HTML	PDF	XML	Total
Feb 2026	344	84	24	452
Mar 2026	386	195	43	624
Apr 2026	82	62	3	147
May 2026	16	21	2	39

Cumulative views and downloads (calculated since 23 Feb 2026)

Month	HTML	PDF	XML	Total
Feb 2026	344	84	24	452
Mar 2026	386	195	43	624
Apr 2026	82	62	3	147
May 2026	16	21	2	39

Viewed (geographical distribution)

Total article views: 1,262 (including HTML, PDF, and XML) Thereof 1,262 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 19 May 2026

Short summary

Small, sudden landslides are common in South China, often triggered by heavy rain. Traditional ways of assessing such risks rely on fixed land characteristics and lack the ability to track real-time ground changes. The study combines advanced satellite monitoring technology and intelligent data analysis to develop a dynamic risk assessment method. This method provides a reliable way to assess small landslide risks dynamically.


Total:	0
HTML:	0
PDF:	0
XML:	0