the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 12 Sep 2025
Geospatial Analysis of Fault–Epicenter Dynamics in Bangladesh and Adjacent Regions Using Remote Sensing and Statistical Modeling
Abstract. Bangladesh and its adjacent regions are situated at the junction of several tectonic plates and are hence highly susceptible to earthquakes. This study investigates the spatial dynamics between fault lines, their classifications, and earthquake epicenters in Bangladesh and its neighboring countries. With a high-density population, absence of urban planning, and inter-border seismic hazards, identifying the way fault types interact with seismic activity is crucial for an effective estimation of the hazard in this area.
Landsat 8 Thermal Infrared Sensor (Band 10) satellite imagery was used for the detection of faults, followed by the extraction of lineaments using PCI Geomatica and spatial analysis in ArcMap 10.8. Fault lines were identified as four principal types: normal, reverse, left-lateral, and right-lateral, based on geometric and spatial features. Earthquake epicenter data between 1924 and 2024 were derived from the USGS Earthquake Catalog. Spatial autocorrelation analysis (Moran's I), Kruskal-Wallis test, Dunn's test, and multinomial logistic regression were used to examine fault-epicenter relationships.
Approximately 40,000 fault lineaments were identified. Moran's I index (0.298, p<0.000001) confirmed significant spatial clustering of epicenters and fault lines. Dunn's test demonstrated that reverse faults significantly differ from the others in terms of proximity to epicenters. Multinomial logistic regression revealed that earthquakes tend to be closer to normal (p = 0.042) and left-lateral faults (p = 0.016), whereas reverse faults (p = 0.676) did not exhibit significant differentiation based on proximity.
This work highlights the crucial need to incorporate fault types and epicenter spatial relationships into seismic hazard models. The results offer practical insight into regional earthquake risk mitigation, infrastructure design, and transboundary disaster preparedness in Bangladesh and adjacent regions.
- Preprint
(1885 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-3774', Anonymous Referee #1, 28 Nov 2025
-
AC1: 'Reply on RC1', Md. Abu Dardha Limon, 18 May 2026
Author comment in response to Referee #1
Referee comment:
This study investigates the spatial relationships between fault and earthquake locations in Bangladesh. First, faults are mapped using satellite imagery and an automated workflow that extracts lineaments. Earthquake locations are taken from the USGS catalog. Statistical analysis are then performed, which indicate that earthquake locations in Bangladesh are not randomly distributed but tend to cluster near faults.
The statistical analysis is impressive, and the study is well written. However, as outlined below, I have significant reservations about the underlying data used in this analysis. In particular, the robustness of fault mapping and whether their automated workflow is reliably identifying active faults and correctly classifying their kinematics. In addition, the high (5-10 km) earthquake location uncertainties in Bangladesh severely limit the spatial analysis, as does the fact that it is performed in 2D (i.e.., it doesn’t consider the depths at which the earthquakes occur, or the down-dip projection of faults).
I appreciate the author’s extensive work analysing fault-earthquake spatial relationships in Bangladesh. However, given the above points, my recommendation is that they instead just focus on using the satellite imagery to develop a robust active fault map for Bangladesh. I have provided some ideas for how this could be achieved below. As the authors correctly point out, Bangladesh is highly vulnerable to future earthquakes, and I think a study that provides new insights into the distribution of its active faults would form an important and interesting article in Solid Earth.
Author comment:
We sincerely thank the referee for carefully reading our manuscript and for the constructive, detailed comments. We are grateful for the positive assessment of the statistical analysis and writing quality, as well as for the valuable suggestions regarding the robustness of the fault mapping, fault-kinematic interpretation, earthquake-location uncertainty, and the limitations of the two-dimensional spatial analysis. We agree that these issues need to be addressed more carefully to strengthen the manuscript.
In response to the referee’s recommendation, we will substantially revise the manuscript by improving the validation of the mapped fault and lineament datasets, clearly distinguishing active faults from extracted lineaments, revising the fault-classification approach, and more explicitly discussing the limitations related to earthquake-location uncertainty and 2D analysis. We will also strengthen the manuscript’s focus on developing a robust active-fault and fault-related lineament map for Bangladesh and adjacent regions, while treating the earthquake–fault spatial analysis more cautiously.
Major comments
Referee comment:
1.) Fault mapping: Section 2.3 describes how faults in Bangladesh were mapped from lineaments identified in satellite imagery using PCI Geomatica’s lineament extraction software. Given the implementation of an automated workflow to map faults, I strongly recommend some testing for whether the extracted lineaments do represent faults and/or more documentation is provided for why these lineaments are interpreted as active faults (see for example Scott et al 2025). For example, are the lineaments consistent with faults identified in geologic maps, hillshade rendering of digital elevation models, and other active fault compilations in Bangladesh (e.g., Hossain et al 2020, Styron and Pagani 2020)? Is it possible that some of the lineaments are other geomorphic or structural features (e.g., joints?) The very high number of short (<10 km) diffuse faults in Figures 3 and 4 is unlike most other fault compilations in that: (1) the distribution of fault lengths tends to follow a power-law distribution with an exponent of 2 (Zou and Fialko 2024), and (2) most active fault traces (or more specifically, earthquake surface ruptures) are >5 km long (Christophersen et al 2015). Some of these comments are addressed in the discussion (Lines 276-284), but additional work is needed to demonstrate that these lineaments represent active faults.
Secondly, this compilation is specifically for active fault traces (e.g., Line 234). Hence, it is necessary that this study provides details on how active and inactive faults are distinguished. This is important as there is no universal definition for what constitutes an ‘active’ fault (see for example, Styron and Pagani 2020, Williams et al 2022), and without these details, I cannot be confident that inactive faults are being excluded in the statistical analysis of earthquake-fault relationships.Author comment:
We sincerely thank the referee for this detailed and constructive comment. We agree that the submitted manuscript did not provide sufficient documentation of the validation procedure used to assess whether the automatically extracted lineaments represent fault-related structures. In Section 2.3, we described the PCI Geomatica lineament extraction workflow, but we did not adequately explain how the extracted lineaments were checked against independent geological, geomorphic, or active-fault datasets. We also acknowledge that the wording in section 3.1, particularly the statement that the red linear features represent “active faults and seismogenic lineaments,” may have created ambiguity regarding whether all extracted lineaments were interpreted as active faults.We would like to clarify that we did not intend to imply that all automatically extracted lineaments are active faults. The active fault traces were obtained from the Global Active Earthquake Faults dataset available through ArcGIS Online, whereas PCI Geomatica was used to extract additional lineaments from satellite imagery. However, this distinction was not clearly documented in the submitted manuscript. In the revised manuscript, we will explicitly separate these datasets and use different terminology and symbology for: (i) previously mapped active faults, (ii) extracted lineaments that are possibly fault-related, and (iii) uncertain or non-validated lineaments.
We agree with the referee that some extracted lineaments may represent other geomorphic or structural features, such as joints, fractures, lithological boundaries, drainage alignments, or image-processing artefacts, rather than active faults. Although the discussion briefly noted that lineaments do not necessarily all correspond to currently active faults, this limitation was not sufficiently incorporated into the methodology or statistical analysis. In the revised manuscript, we will address this by validating the extracted lineaments against available geological maps, DEM or hillshade interpretations, published active-fault compilations, and regional seismicity patterns, including relevant datasets and studies such as Hossain et al. (2020), Styron and Pagani (2020), and other applicable active-fault references.
We also accept the referee’s concern regarding the very high number of short and diffuse lineaments shown in the original Figures 3 and 4. In the revised manuscript, short, diffuse, or unsupported lineaments will not be treated as confirmed active fault traces. We will revise the maps and the statistical analysis using a more conservative dataset composed of validated active or potentially active fault structures. Uncertain lineaments will either be excluded from the main earthquake–fault relationship analysis or discussed separately as possible structural features.
Finally, we agree that the manuscript needs to clearly define how active and inactive faults are distinguished. Since there is no universal definition of an active fault, the revised manuscript will provide an operational definition for this study. Faults will be considered active or potentially active only where they are supported by published active-fault datasets, mapped geological or geomorphic evidence, consistency with DEM or hillshade interpretation, association with known tectonic structures, or regional seismicity. Inactive or uncertain lineaments will not be treated as confirmed active faults in the main statistical analysis. Section 2.3 will therefore be rewritten to describe the validation procedure and classification criteria, Figures 3 and 4 will be revised to show active faults and extracted lineaments as separate layers, and the earthquake–fault statistical analysis will be updated accordingly.
Referee comment:
2.) Fault Classification: Following the identification of fault lineaments, a slip type is assigned to a fault to based ‘on the assessment of its length, curvature, intersection patterns, clustering, and orientation of the faults (Lines 135-137).’ This is highly unusual as fault kinematics should instead be defined by offset markers (e.g., geologic units, geomorphic features). Is there any indication from these features for what the kinematics of Bangladesh’s faults are?
Alternatively, it’s noted that fault orientations are used to infer kinematics, and this is defensible For example, if it was performed by applying the Andersonian theory of faulting to Bangladesh’s regional stress state and/or comparison to earthquake focal mechanisms in this region? (https://www.globalcmt.org/CMTsearch.html). In this context, it is worrying that Figure 4 indicates that there are reverse and normal faults adjacent to each other and left lateral faults that strike at 90º to each other, and the same for right-lateral faults. In addition, there is an along-strike sharp transition from right- to left-lateral faulting around Rangpur without any change in fault orientation. This implies very small-scale stress rotations. Do the authors think these are realistic?
Author comment:
We sincerely thank the referee for raising this important methodological point. We agree that the submitted manuscript did not provide sufficient tectonic and kinematic justification for assigning slip types to the extracted fault-related lineaments. In the submitted version, the classification was based primarily on geometric and spatial characteristics, including fault orientation, length, curvature, clustering, and intersection patterns, together with comparison to earlier tectonic studies from the Indo-Burma subduction zone and the United States Fault and Fold Database. However, we acknowledge that these criteria alone are insufficient for confidently determining fault kinematics, particularly without direct geologic or geomorphic offset markers.
We agree with the referee that fault kinematics are more reliably constrained using geologic or geomorphic offset markers, focal mechanism solutions, and regional stress-field interpretation. In the revised manuscript, we will therefore substantially revise the fault-classification methodology by incorporating available geologic or geomorphic offset evidence, known tectonic structures, tectonic stress information from the World Stress Map, and available earthquake focal mechanism data. We will also clarify that fault orientation was used only as a preliminary indicator of possible kinematics rather than as definitive evidence of fault slip type.
We further acknowledge the referee’s concern regarding the unrealistic juxtaposition of reverse and normal faults, orthogonal left-lateral or right-lateral structures, and the abrupt transition from right-lateral to left-lateral faulting near Rangpur without substantial orientation changes. We agree that these patterns likely reflect limitations of the original geometry-based classification approach and may imply unrealistically small-scale stress rotations. In the revised manuscript, these interpretations will be reassessed using available geologic or geomorphic offset evidence, known tectonic structures, the regional tectonic stress field, and published focal mechanism information. Fault segments lacking sufficient tectonic or kinematic support will either be reclassified more conservatively or described as uncertain fault-related lineaments rather than confidently assigned slip types.
In the revised manuscript, Section 2.4 will therefore be substantially rewritten to provide clearer and more rigorous criteria for fault-type classification. Figure 4 and related interpretations will be revised, and the fault-type-based statistical analysis will be recalculated according to the revised classification.
Referee comment:
3.) Earthquake data: The analysis of earthquake locations in this study was conducted using M>3 events between 1924-2024 in the USGS earthquake catalog (Section 2.5). However, it should be noted that due to uncertainties in picking earthquake arrivals, sparse station spacing, and seismic velocity models, the earthquake locations in this catalog have a horizontal and depth location uncertainty of 5-10 km (and that’s only for events after 2014 when these uncertainties are reported).
Ideally, statistical analyses between earthquake locations and faults should be performed using high resolution earthquake catalogs (e.g. Hauksson et al 2012). If this is not possible for the Bangladesh earthquake catalogs, then I recommend a sensitivity analysis for whether the earthquake location uncertainties influence the earthquake-fault spatial analysis in Section 2.6 (notwithstanding my next comments below). For example, by repeating this analysis with randomly perturbed earthquake locations.
Author comment:
We sincerely thank the referee for this important and constructive comment. We agree that earthquake-location uncertainty is a critical limitation in fault–earthquake spatial analysis, particularly for historical earthquake records and regions with relatively sparse seismic-station coverage. In the submitted manuscript, we used M ≥ 3.0 earthquake events from the USGS Earthquake Catalog for 1924–2024, but we did not sufficiently discuss the potential effects of horizontal and depth-location uncertainties on the spatial analysis.
We also agree that a high-resolution relocated earthquake catalog, such as the waveform-relocated catalog of Hauksson et al. (2012), would be preferable for this type of analysis. For Bangladesh, we have identified several updated and unified earthquake catalogs that compile and homogenize events from multiple sources, including USGS, ISC, ISC-GEM, and related datasets. These catalogs are highly useful for improving regional earthquake data completeness and magnitude consistency. However, they are not directly equivalent to a dense-network waveform-relocated catalog such as that available for Southern California. Therefore, in the revised manuscript, we will supplement the USGS-only dataset with the most appropriate updated Bangladesh earthquake catalog, while explicitly acknowledging remaining location-uncertainty limitations.
To address the referee’s concern quantitatively, we will add a sensitivity analysis to the earthquake–fault proximity analysis. Specifically, earthquake epicenters will be randomly perturbed within plausible horizontal uncertainty ranges, such as 5–10 km, and the proximity analysis will be repeated to test whether the observed fault–earthquake spatial relationships remain stable. Section 2.5 will be revised to describe the earthquake-catalog selection and uncertainty limitations, while Section 2.6 will describe the sensitivity test because that section contains the proximity analysis between earthquake epicenters and fault traces.
The results and discussion will be updated according to the sensitivity test outcomes. If the perturbed-location analysis produces substantial variability, the interpretation of earthquake–fault spatial relationships will be made more cautious. If the spatial patterns remain broadly consistent, this will be reported as supporting evidence for the robustness of the observed relationships.
Referee comment:
4.) Earthquake-fault relationships: The statistical tests described in Section 2.6 between fault lines and earthquake epicentres is essentially a 2D analysis. It therefore neglects that earthquakes occur within 3D space and that faults are 2D planes that project down-dip through the crust. In other words, the Euclidean relationship between earthquake locations and faults should be considered in 3D (except in case of vertically dipping faults where this simplification is acceptable).
Author comment:
We appreciate the referee’s careful assessment of this issue. We agree that the submitted manuscript treated the earthquake–fault relationship as a two-dimensional surface-based relationship between earthquake epicenters and mapped fault traces. This approach does not fully account for the fact that earthquakes occur at depth and that faults are commonly dipping planes or zones that extend down-dip through the crust.We acknowledge that a full 3D analysis would ideally measure the distance between earthquake hypocenters and fault planes, rather than only the surface distance between epicenters and fault traces. Although the USGS earthquake catalog used in this study includes earthquake-depth information, depth data alone are not sufficient for a complete 3D fault-plane distance analysis because consistent fault-plane geometry is also required. Specifically, fault dip angle, dip direction, depth extent, and down-dip fault-plane projection are needed to calculate true hypocenter-to-fault-plane distances. These parameters cannot be derived reliably from the automatically extracted surface lineaments produced using PCI Geomatica, and they are also not provided consistently in the Global Active Earthquake Faults dataset used in this study.
In the revised manuscript, we will therefore clearly describe the analysis as a 2D epicentral proximity analysis, rather than a full 3D earthquake–fault distance analysis. We will incorporate the available earthquake-depth information to describe the hypocentral distribution of the earthquake dataset and to discuss whether the earthquake events are mainly shallow or deep in relation to the mapped surface fault traces. Where reliable published fault-geometry information is available for major structures, it will be discussed qualitatively. However, we will avoid over-interpreting the 2D proximity results as direct evidence of true 3D earthquake–fault association.
Section 2.6 will be revised to clarify that the distance calculation represents surface epicenter-to-fault-trace proximity. The Results and Discussion sections will also be revised to acknowledge the limitations caused by the absence of consistent fault-plane geometry, including fault dip, dip direction, depth extent, and down-dip projection, while making use of the available earthquake-depth information from the USGS catalog.
Minor Comments
Referee comment:
Lines 40-42: Suggest removing the reference that earthquakes “claimed an average of over 25,000 years lives annually” as this implies that there is some regularity to the number of earthquake casualties each year when this was not the case. Or in other words, most earthquake casualties in the 20th century can be attributed to a few destructive events, and so in most years there are relatively few earthquake casualties (see for example Figure 1 in Holzer and Savage 2013).
Author comment:
We thank the referee for this helpful comment. We agree that the phrase “claimed an average of over 25,000 lives annually” may incorrectly imply a regular annual pattern of earthquake casualties. In the revised manuscript, we will remove this phrase and reword the sentence to clarify that earthquake fatalities are often concentrated in a limited number of highly destructive events rather than occurring at a regular annual rate.
Referee comment:
Lines 65-68: Although fault damage zones undoubtedly influence earthquake rupture and seismic hazard (see also Biegel and Sammis 2004), I suggest remove these sentences from the introduction, as fault damage zones and their influence on earthquake hazards in Bangladesh are not discussed elsewhere in the manuscript.
Author comment:
We thank the referee for this helpful suggestion. We agree that the discussion of fault damage zones is not directly developed elsewhere in the manuscript and may distract from the main focus of the study. In the revised manuscript, we will remove these sentences from the Introduction to make the background more concise and better aligned with the scope of the paper.
Referee comment:
Line 75: What does TEC mean?
Author comment:
We thank the referee for pointing this out. TEC refers to Total Electron Content. In the revised manuscript, we will define the abbreviation at its first occurrence to improve clarity. We will revise the sentence accordingly so that “TEC” is introduced as “Total Electron Content (TEC).”
Referee comment:
Figure 1: It would be helpful if tectonic plate boundaries around Bangladesh were depicted on this map
Author comment:
We thank the referee for this helpful suggestion. We agree that adding tectonic plate boundaries would improve the geological context of the study area map. In the revised manuscript, we will update Figure 1 by including the major tectonic plate boundaries around Bangladesh and adjacent regions. The figure title, legend, and caption will also be revised accordingly.Referee comment:
Line 117: Although a 30 m resolution image is acceptable for remotely mapping faults with prominent scarps (heights >5 m, see for example Hodge et al 2019), inevitably, some smaller faults would be missing from this mapping. I therefore recommend adding some commentary on how this bias would influence the completeness of their fault mapping.
Author comment:
We thank the referee for identifying this issue. We agree that the 30 m spatial resolution may introduce a mapping bias, as smaller or less prominent fault-related features may not be detectable, particularly where surface expression is weak, discontinuous, or obscured by vegetation, sediment cover, urbanization, or agricultural modification. In the revised manuscript, we will add commentary on this limitation and clarify that the remote-sensing-based mapping is more likely to identify larger and more prominent lineaments or fault-related features, while smaller-scale faults may be underrepresented.Referee comment:
Line 125: I recommend adding some references here for examples of where PCI’s Geomatica software lineament extraction analysis has been used to map geological features.
Author comment:
We thank the referee for this helpful recommendation. We agree that the statement regarding the application of PCI Geomatica for geological lineament extraction should be supported by appropriate references. In the revised manuscript, we will add relevant references showing previous applications of PCI Geomatica lineament extraction for mapping geological structures from satellite imagery. This will strengthen the methodological justification for using the software in the present study.Referee comment:
Lines 279-281: It would be helpful to elaborate here with examples of where lineaments are associated with earthquake clusters and/or lineaments are favourably oriented for reactivation in this region’s stress state. For example, what are the principal stress orientations in this region?
Author comment:
We thank the referee for pointing this out. We agree that the statement requires further support and should not imply reactivation of most lineaments without clear evidence. In the revised manuscript, we will elaborate this discussion by adding examples where mapped lineaments coincide with earthquake clusters and known tectonic structures in Bangladesh and adjacent regions. We will also incorporate regional stress-field information, including principal stress orientations from available stress datasets such as the World Stress Map, to assess whether selected lineaments are favourably oriented for possible reactivation. The wording will be revised to avoid overstatement and to ensure that any interpretation of possible lineament reactivation is supported by regional stress-field and seismicity evidence.Referee comment:
Lines 348-352: Lines 348-352: I disagree with the assertion that the correlation between earthquake locations and faults can be used to constrain a fault’s seismic hazard. Ultimately, the key factor that determines a fault’s earthquake potential is its slip rate and area, as these in turn, influence its seismic moment rate. Notably, there are several examples of ‘locked’ faults which have very little seismicity associated with them, but are inferred from geologic studies to have high slip rates and seismic hazard (e.g. New Zealand’s Alpine Fault, Norris and Toy 2014). Furthermore, it is expected that a considerable amount of seismicity occurs away from mapped faults (Zou and Fialko 2024).
Author comment:
We thank the referee for this important clarification. We agree that the correlation between earthquake locations and mapped faults should not be interpreted as a direct constraint on fault seismic hazard or earthquake potential, as these depend mainly on factors such as slip rate, fault area, recurrence behavior, and seismic moment rate. In the revised manuscript, we will remove or rephrase the statements suggesting that proximity-based correlation can directly indicate seismic hazard or enable probabilistic fault behavior modeling. The discussion will be revised to present the results more cautiously as a spatial association between mapped faults and earthquake epicenter distribution, rather than as a direct measure of fault activity or seismic risk.Citation: https://doi.org/10.5194/egusphere-2025-3774-AC1
-
AC1: 'Reply on RC1', Md. Abu Dardha Limon, 18 May 2026
-
RC2: 'Comment on egusphere-2025-3774', Anonymous Referee #2, 07 May 2026
This work investigated lineaments using satellite imagery to identify faults in Bangladesh and its neighboring regions. After classifying fault lines into four categories, earthquake data were compiled to measure the shortest distance from each epicenter to the mapped fault lines. The authors further applied several statistical tests to determine which types of faulting are spatially associated with earthquakes.This study and its results are significant for estimating the productivity of individual faults and quantifying nationwide seismic hazards. The workflows (Figure 2) are generally reasonable for analyzing the spatial proximity of normal and left-lateral faults. The topic and contents are suitable for publication in the journal Solid Earth; however, the article needs to refine its technical descriptions and conduct additional quality controls before publication.Major commentsIntroduction:The first and third paragraphs can be shortened or removed, as they contain overly general descriptions of seismic risks and fault zones. The second paragraph provides more specific justification for the present study.Materials and methods:In Section 2.4, please provide further details on how individual fault lines were classified into four types. For example, what criteria were used to identify normal faulting? Some fault lines may not be active under the current stress field. The authors could consider the optimal orientation of fault lines relative to the tectonic stress (e.g., the World Stress Map; https://www.world-stress-map.org/).In Section 2.5, the earthquake epicenter data need to be presented with their associated location uncertainties. To ensure more reliable results, the authors could perform sensitivity tests using various magnitude cutoffs (e.g., 3, 4, and 5). A bootstrap test could also be applied to ensure robustness against to epicenter uncertainties.In Section 2.6, some operational details regarding software usage can be omitted, whereas the mathematical formulation of Moran’s I needs to be explicitly specified. For example, what attribute values were used to calculate Moran’s I? How was the weight defined?In section 2.8.1, was the Shapiro-Wilk normality test applied to the dataset of each fault type individually? If so, why is there only one p-value reported?Results and DiscussionI recommend merging these two sections, as the results section currently includes figures and interpretive descriptions. For example, Section 4.1 can be merged with 3.1. The description in 3.1 can be moved into the caption of Figure 3.As all information in Figure 3 is included in Figure 4, Section 3.1, 3.2, 4.1, and 4.2 could be merged.In Figure 5, the color and size of earthquake symbols could be scaled according to magnitude.In the section of 4.5.1, the multinomial logistic regression results showed that the distances to left lateral and normal faults were shorter than those of right lateral faults (Table 2). However, the Dunn’s test showed that these distances were not significantly different (Table 1). Conversely, while the text (line 340) states that reverse faulting distances were not significantly different from right lateral ones, Table 1 shows a significant difference. Please clarify these discrepancies between the Dunn’s test and the multinomial logistic regression results.Does changing the reference category in the multinomial logistic regression yield different interpretations?To improve the visual presentation, I recommend adding (1) histograms showing the distribution of distances for all four faulting types , and (2) a combined map of Figures 4 and 5 where epicenter colors are scaled by their distance to the nearest fault.Line-by-line comments are available below:Line 75: What does TEC stand for?Line 87: Mark the Indo-Burman range and Himalayan tectonic regime in Figure 1.Lines 89 and 149: Add either bounding boxes or a custom boundary polygon for the regions described here in Figure 1. Illustrate the custom boundary polygon in Figure 1.Line 98: Mark the Himalayan Arc, the Shillong Plateau, and the Burmese Arc in Figure 1.Line 120: Provide a reference for PCI Geomatica.Line 132: Provide a reference for ArcMap 10.8.Lines 186-187: Does the “dataset” refer to the nearest distance values? Why is it assumed that the dataset should be normally distributed?Line 282: The description of the lineaments and faults in Quebec and Ontario does not seem relevant to this study.Line 302: Mark the location of the 1762 Arakan earthquake in Figure 5. Other significant historical events could be marked as well.Citation: https://doi.org/
10.5194/egusphere-2025-3774-RC2 -
AC2: 'Reply on RC2', Md. Abu Dardha Limon, 18 May 2026
Author comment in response to Referee #2
Referee comment:
This work investigated lineaments using satellite imagery to identify faults in Bangladesh and its neighboring regions. After classifying fault lines into four categories, earthquake data were compiled to measure the shortest distance from each epicenter to the mapped fault lines. The authors further applied several statistical tests to determine which types of faulting are spatially associated with earthquakes.
This study and its results are significant for estimating the productivity of individual faults and quantifying nationwide seismic hazards. The workflows (Figure 2) are generally reasonable for analyzing the spatial proximity of normal and left-lateral faults. The topic and contents are suitable for publication in the journal Solid Earth; however, the article needs to refine its technical descriptions and conduct additional quality controls before publication.
Author comment:
We sincerely thank the referee for the careful evaluation of our manuscript and for the positive assessment of the study topic, workflow, and relevance to the Solid Earth Journal. We also appreciate the constructive recommendation to improve the manuscript before publication. In the revised manuscript, we will refine the technical descriptions and conduct additional quality-control procedures to strengthen the reliability and clarity of the study.
Major comments
Introduction:
Referee comment:
The first and third paragraphs can be shortened or removed, as they contain overly general descriptions of seismic risks and fault zones. The second paragraph provides more specific justification for the present study.
Author comment:
We thank the referee for this helpful suggestion. We agree that the first and third paragraphs of the Introduction contain overly general background information. We will shorten these paragraphs in the revised manuscript and remove non-essential general descriptions of seismic risk and fault zones. In the revised manuscript, the Introduction will be refocused on the specific tectonic setting of Bangladesh and adjacent regions, while retaining the second paragraph as the main justification for the study.
Materials and methods:
Referee comment:
In Section 2.4, please provide further details on how individual fault lines were classified into four types. For example, what criteria were used to identify normal faulting? Some fault lines may not be active under the current stress field. The authors could consider the optimal orientation of fault lines relative to the tectonic stress (e.g., the World Stress Map; https://www.world-stress-map.org/).
Author comment:
We thank the referee for this important and constructive comment. We agree that Section 2.4 did not provide sufficient detail on how individual fault lines were classified into normal, reverse, left-lateral, and right-lateral types. In the submitted manuscript, the classification was described mainly in terms of fault length, curvature, intersection patterns, clustering, and orientation, but we acknowledge that these criteria alone are not sufficient to confidently assign fault kinematics.
In the revised manuscript, we will substantially expand Section 2.4 by providing clearer criteria for fault-type classification. Specifically, we will consider available geological or geomorphic offset evidence, published fault information, known regional tectonic structures, earthquake focal mechanism data where available, and the orientation of faults relative to the present-day regional stress field. We will also incorporate information from the World Stress Map to assess whether mapped fault orientations are compatible with the current tectonic stress regime.
We agree that some mapped fault lines or lineaments may not be active under the current stress field. Therefore, in the revised manuscript, we will distinguish more clearly between validated active or potentially active faults and uncertain lineaments. Faults lacking sufficient kinematic or tectonic evidence will not be assigned a definitive slip type. Figure 4 and the related fault-type-based statistical analysis will also be revised according to the updated classification.
Referee comment:
In Section 2.5, the earthquake epicenter data need to be presented with their associated location uncertainties. To ensure more reliable results, the authors could perform sensitivity tests using various magnitude cutoffs (e.g., 3, 4, and 5). A bootstrap test could also be applied to ensure robustness against to epicenter uncertainties.
Author comment:
We thank the referee for this constructive suggestion. We agree that the earthquake epicenter data should be presented with associated location uncertainties, as these uncertainties may influence the reliability of the fault–earthquake proximity analysis. In the revised manuscript, we will expand Section 2.5 by reporting the available location-uncertainty information from the earthquake catalog and by clearly acknowledging that uncertainty estimates may be incomplete or less reliable for older historical events. To test the robustness of the results, we will perform sensitivity analyses using different magnitude thresholds, including M ≥ 3, M ≥ 4, and M ≥ 5. We will also apply a bootstrap or random-perturbation sensitivity test to evaluate whether epicenter-location uncertainty affects the observed spatial relationships between earthquake epicenters and mapped faults. The results and discussion will be revised accordingly based on these robustness checks.
Referee comment:
In Section 2.6, some operational details regarding software usage can be omitted, whereas the mathematical formulation of Moran’s I needs to be explicitly specified. For example, what attribute values were used to calculate Moran’s I? How was the weight defined?
Author comment:
We thank the referee for this helpful recommendation. We agree that Section 2.6 contains unnecessary operational details about software workflow, while the statistical formulation of the spatial analysis requires a clearer explanation. In the revised manuscript, we will shorten the software-specific description and explicitly define the mathematical procedures used in the proximity and Moran’s I analysis. Specifically, we will clarify that the distance attribute used for the spatial autocorrelation analysis is the nearest-distance value between each earthquake epicenter and its closest mapped fault trace. We will also specify how the spatial weight matrix was defined, including the weighting method and distance criteria used. In addition, we will correct and clearly present the Euclidean distance formula used in the proximity calculation. This revision will make the spatial-statistical method more transparent and reproducible.
Referee comment:
In section 2.8.1, was the Shapiro-Wilk normality test applied to the dataset of each fault type individually? If so, why is there only one p-value reported?
Author comment:
We thank the referee for pointing this out. We clarify that the Shapiro–Wilk normality test reported in the submitted manuscript was applied to the overall nearest-distance dataset, not to each fault-type group separately; therefore, only one W statistic and one p-value were reported. However, we agree that this should be stated more clearly. In the revised manuscript, we will clarify the scope of the normality test and, where appropriate, also report group-wise normality checks for each fault-type category to better justify the use of the Kruskal–Wallis and Dunn’s post-hoc tests.Results and Discussion
Referee comment:
I recommend merging these two sections, as the results section currently includes figures and interpretive descriptions. For example, Section 4.1 can be merged with 3.1. The description in 3.1 can be moved into the caption of Figure 3.
Author comment:
We thank the referee for this helpful suggestion. We agree that the current Results and Discussion sections contain overlap, particularly where figure descriptions and interpretations are presented separately. In the revised manuscript, Section 4.1 will be merged with Section 3.1 to improve readability and reduce duplication. The descriptive information currently included in Section 3.1 will also be moved into the caption of Figure 3.
Referee comment:
As all information in Figure 3 is included in Figure 4, Section 3.1, 3.2, 4.1, and 4.2 could be merged.
Author comment:
We thank the referee for this helpful recommendation. We agree that Sections 3.1, 3.2, 4.1, and 4.2 contain repeated descriptions of the mapped fault patterns and fault-type interpretations. In the revised manuscript, these sections will be merged to present the fault and fault-type distribution more coherently and concisely. This revision will avoid repeated descriptions of the same spatial patterns.Referee comment:
In Figure 5, the color and size of earthquake symbols could be scaled according to magnitude.
Author comment:
We thank the referee for this helpful suggestion. We agree that scaling the color and size of earthquake symbols according to magnitude would improve the visual interpretation of Figure 5. In the revised manuscript, we will update Figure 5 so that earthquake magnitudes are represented more clearly through symbol size and color variation.
Referee comment:
In the section of 4.5.1, the multinomial logistic regression results showed that the distances to left lateral and normal faults were shorter than those of right lateral faults (Table 2). However, the Dunn’s test showed that these distances were not significantly different (Table 1). Conversely, while the text (line 340) states that reverse faulting distances were not significantly different from right lateral ones, Table 1 shows a significant difference. Please clarify these discrepancies between the Dunn’s test and the multinomial logistic regression results.
Author comment:
We thank the referee for carefully identifying this important inconsistency. We agree that the relationship between the Dunn’s test results and the multinomial logistic regression results was not explained clearly enough in the submitted manuscript. The Dunn’s test compares pairwise differences in the distributions of nearest epicentral distances among fault-type groups, whereas the multinomial logistic regression estimates fault-type membership relative to a selected reference category as a function of distance. Therefore, the two analyses address related but not identical statistical questions.We acknowledge that some interpretations in Section 4.5.1 were overstated or inconsistently expressed. In particular, the statement regarding reverse faults will be corrected, because Table 1 shows a significant difference between reverse and right-lateral faults. We will also revise the interpretation of left-lateral and normal faults to avoid implying that they are definitively more seismically productive or more likely to generate earthquakes based only on proximity statistics. The text in Section 4.5.1 will be revised accordingly, and overstated phrases such as “fault types more likely to trigger nearby earthquakes,” “seismic productivity,” and “probabilistic fault behavior modeling” will be removed or rephrased more cautiously as evidence of spatial association rather than direct evidence of fault activity or seismic hazard.
Referee comment:
Does changing the reference category in the multinomial logistic regression yield different interpretations?
Author comment:
We thank the referee for this helpful question. Yes, changing the reference category in the multinomial logistic regression changes the pairwise contrasts directly shown in the model output and can therefore lead to different apparent interpretations. In the revised manuscript, we will recheck the multinomial logistic regression using alternative reference categories and present the results more clearly. The revised interpretation will distinguish the pairwise distance differences shown by Dunn’s test from the reference-category-dependent contrasts shown by the multinomial logistic regression. The fault-type-based statistical analysis will also be updated if the revised fault classification changes the underlying dataset.Referee comment:
To improve the visual presentation, I recommend adding (1) histograms showing the distribution of distances for all four faulting types , and (2) a combined map of Figures 4 and 5 where epicenter colors are scaled by their distance to the nearest fault.
Author comment:
We thank the referee for this helpful suggestion. We agree that additional visualizations would improve the presentation and interpretation of the fault–earthquake proximity results. In the revised manuscript, we will add histograms showing the distribution of nearest epicentral distances for all four fault types. We will also prepare a combined map integrating the fault-type map and earthquake epicenters, with epicenter colors scaled according to their distance from the nearest fault. These additions will make the spatial and statistical patterns easier to interpret visually.
Line-by-line comments
Referee comment:
Line 75: What does TEC stand for?
Author comment:
We thank the referee for pointing this out. TEC stands for Total Electron Content. In the revised manuscript, we will define the abbreviation at its first occurrence as “Total Electron Content (TEC)” to improve clarity.Referee comment:
Line 87: Mark the Indo-Burman range and Himalayan tectonic regime in Figure 1.
Author comment:
We thank the referee for this helpful suggestion. In the revised manuscript, we will update Figure 1 by marking the Indo-Burman Range and the Himalayan tectonic regime to provide a clearer tectonic context for the study area.Referee comment:
Lines 89 and 149: Add either bounding boxes or a custom boundary polygon for the regions described here in Figure 1. Illustrate the custom boundary polygon in Figure 1.
Author comment:
We thank the referee for this helpful suggestion. We agree that the geographic extent described in lines 89 and 149 should be shown more clearly in Figure 1. In the revised manuscript, we will update Figure 1 by adding the custom boundary polygon used for earthquake-data filtering, covering Bangladesh, West Bengal, parts of Nepal and Bhutan, the Bay of Bengal, northeast India, including Assam, Meghalaya, Mizoram, and Tripura, and western Myanmar. We will also ensure that the Bangladesh national boundary and the broader study-area polygon are clearly distinguishable in the figure legend.Referee comment:
Line 98: Mark the Himalayan Arc, the Shillong Plateau, and the Burmese Arc in Figure 1.
Author comment:
We thank the referee for this helpful suggestion. In the revised manuscript, we will update Figure 1 by marking the Himalayan Arc, the Shillong Plateau, and the Burmese Arc to provide a clearer tectonic framework for the study area.Referee comment:
Line 120: Provide a reference for PCI Geomatica.
Author comment:
We thank the referee for this suggestion. In the revised manuscript, we will add an appropriate reference for PCI Geomatica.
Referee comment:
Line 132: Provide a reference for ArcMap 10.8.
Author comment:
We thank the referee for this suggestion. In the revised manuscript, we will add an appropriate reference for ArcMap 10.8.Referee comment:
Lines 186-187: Does the “dataset” refer to the nearest distance values? Why is it assumed that the dataset should be normally distributed?
Author comment:
We thank the referee for this helpful comment. We agree that the term “dataset” was unclear. It refers to the nearest-distance values between earthquake epicenters and the closest mapped fault lines, grouped by fault type. We did not assume that the data were normally distributed; rather, we tested whether the nearest-distance values were normally distributed or not in order to determine whether parametric or nonparametric statistical methods were appropriate. In the revised manuscript, we will revise this wording for clarity.Referee comment:
Line 282: The description of the lineaments and faults in Quebec and Ontario does not seem relevant to this study.
Author comment:
We thank the referee for this helpful comment. We agree that the example from Quebec and Ontario is not directly relevant to the tectonic setting of Bangladesh and adjacent regions. In the revised manuscript, we will remove this sentence or replace it with a more regionally relevant reference.Referee comment:
Line 302: Mark the location of the 1762 Arakan earthquake in Figure 5. Other significant historical events could be marked as well.
Author comment:
We thank the referee for this helpful suggestion. In the revised manuscript, we will update Figure 5 by marking the location of the 1762 Arakan earthquake. We will also consider marking other significant historical earthquakes in the study region where reliable location information is available.Citation: https://doi.org/10.5194/egusphere-2025-3774-AC2
-
AC2: 'Reply on RC2', Md. Abu Dardha Limon, 18 May 2026
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 6,733 | 1,110 | 140 | 7,983 | 272 | 348 |
- HTML: 6,733
- PDF: 1,110
- XML: 140
- Total: 7,983
- BibTeX: 272
- EndNote: 348
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This study investigates the spatial relationships between fault and earthquake locations in Bangladesh. First, faults are mapped using satellite imagery and an automated workflow that extracts lineaments. Earthquake locations are taken from the USGS catalog. Statistical analysis are then performed, which indicate that earthquake locations in Bangladesh are not randomly distributed but tend to cluster near faults.
The statistical analysis is impressive, and the study is well written. However, as outlined below, I have significant reservations about the underlying data used in this analysis. In particular, the robustness of fault mapping, and whether their automated workflow is reliably identifying active faults and correctly classifying their kinematics. In addition, the high (5-10 km) earthquake location uncertainties in Bangladesh severely limit the spatial analysis, as does the fact that it is performed in 2D (i.e.., it doesn’t consider the depths at which the earthquakes occur, or the down-dip projection of faults).
I appreciate the author’s extensive work analysing fault-earthquake spatial relationships in Bangladesh. However, given the above points, my recommendation is that they instead just focus on using the satellite imagery to develop a robust active fault map for Bangladesh. I have provided some ideas for how this could be achieved below. As the authors correctly point out, Bangladesh is highly vulnerable to future earthquakes, and I think a study that provides new insights into the distribution of its active faults would form an important and interesting article in Solid Earth.
Major comments
1.) Fault mapping: section 2.3 describes how faults in Bangladesh were mapped from lineaments identified in satellite imagery using PCI Geomatica’s lineament extraction software. Given the implementation of an automated workflow to map faults, I strongly recommend some testing for whether the extracted lineaments do represent faults and/or more documentation is provided for why these lineaments are interpreted as active faults (see for example Scott et al 2025). For example, are the lineaments consistent with faults identified in geologic maps, hillshade rendering of digital elevation models, and other active fault compilations in Bangladesh (e.g., Hossain et al 2020, Styron and Pagani 2020)? Is it possible that some of the lineaments are other geomorphic or structural features (e.g., joints?) The very high number of short (<10 km) diffuse faults in Figures 3 and 4 is unlike most other fault compilations in that: (1) the distribution of fault lengths tends to follow a power-law distribution with an exponent of 2 (Zou and Fialko 2024), and (2) most active fault traces (or more specifically, earthquake surface ruptures) are >5 km long (Christophersen et al 2015). Some of these comments are addressed in the discussion (Lines 276-284), but additional work is needed to demonstrate that these lineaments represent active faults.
Secondly, this compilation is specifically for active fault traces (e.g., Line 234). Hence, it is necessary that this study provides details on how active and inactive faults are distinguished. This is important as there is no universal definition for what constitutes an ‘active’ fault (see for example, Styron and Pagani 2020, Williams et al 2022), and without these details, I cannot be confident that inactive faults are being excluded in the statistical analysis of earthquake-fault relationships.
2.) Fault Classification: Following the identification of fault lineaments, a slip type is assigned to a fault to based ‘on the assessment of its length, curvature, intersection patterns, clustering, and orientation of the faults (Lines 135-137).’ This is highly unusual as fault kinematics should instead be defined by offset markers (e.g., geologic units, geomorphic features). Is there any indication from these features for what the kinematics of Bangladesh’s faults are?
Alternatively, it’s noted that fault orientations are used to infer kinematics, and this is defensible For example, if it was performed by applying the Andersonian theory of faulting to Bangladesh’s regional stress state and/or comparison to earthquake focal mechanisms in this region? (https://www.globalcmt.org/CMTsearch.html). In this context, it is worrying that Figure 4 indicates that there are reverse and normal faults adjacent to each other and left lateral faults that strike at 90º to each other, and the same for right-lateral faults. In addition, there is an along-strike sharp transition from right- to left-lateral faulting around Rangpur without any change in fault orientation. This implies very small-scale stress rotations. Do the authors think these are realistic?
3.) Earthquake data: The analysis of earthquake locations in this study was conducted using M>3 events between 1924-2024 in the USGS earthquake catalog (Section 2.5). However, it should be noted that due to uncertainties in picking earthquake arrivals, sparse station spacing, and seismic velocity models, the earthquake locations in this catalog have a horizontal and depth location uncertainty of 5-10 km (and that’s only for events after 2014 when these uncertainties are reported).
Ideally, statistical analyses between earthquake locations and faults should be performed using high resolution earthquake catalogs (e.g. Hauksson et al 2012). If this is not possible for the Bangladesh earthquake catalogs, then I recommend a sensitivity analysis for whether the earthquake location uncertainties influence the earthquake-fault spatial analysis in Section 2.6 (notwithstanding my next comments below). For example, by repeating this analysis with randomly perturbed earthquake locations.
4.) Earthquake-fault relationships: The statistical tests described in Section 2.6 between fault lines and earthquake epicentres is essentially a 2D analysis. It therefore neglects that earthquakes occur within 3D space and that faults are 2D planes that project down-dip through the crust. In other words, the Euclidean relationship between earthquake locations and faults should be considered in 3D (except in case of vertically dipping faults where this simplification is acceptable).
Minor comments
Lines 40-42: Suggest removing the reference that earthquakes “claimed an average of over 25,000 years lives annually” as this implies that there is some regularity to the number of earthquake casualties each year when this was not the case. Or in other words, most earthquake casualties in the 20th century can be attributed to a few destructive events, and so in most years there are relatively few earthquake casualties (see for example Figure 1 in Holzer and Savage 2013).
Lines 65-68: Although fault damage zones undoubtedly influence earthquake rupture and seismic hazard (see also Biegel and Sammis 2004), I suggest remove these sentences from the introduction, as fault damage zones and their influence on earthquake hazards in Bangladesh are not discussed elsewhere in the manuscript.
Line 75: What does TEC mean?
Figure 1: It would he helpful if tectonic plate boundaries around Bangladesh were depicted on this map
Line 117: Although a 30 m resolution image is acceptable for remotely mapping faults with prominent scarps (heights >5 m, see for example Hodge et al 2019), inevitably, some smaller faults would be missing from this mapping. I therefore recommend adding some commentary on how this bias would influence the completeness of their fault mapping.
Line 125: I recommend adding some references here for examples of where PCI’s Geomatica software lineament extraction analysis has been used to map geological features.
Lines 279-281: It would be helpful to elaborate here with examples of where lineaments are associated with earthquake clusters and/or lineaments are favourably oriented for reactivation in this region’s stress state. For example, what are the principal stress orientations in this region?
Lines 348-352: I disagree with the assertion that the correlation between earthquake locations and faults can be used to constrain a fault’s seismic hazard. Ultimately, the key factor that determines a fault’s earthquake potential is its slip rate and area, as these in turn, influence its seismic moment rate. Notably, there are several examples of ‘locked’ faults which have very little seismicity associated with them, but are inferred from geologic studies to have high slip rates and seismic hazard (e.g. New Zealand’s Alpine Fault, Norris and Toy 2014).Furthermore, it is expected that a considerable amount of seismicity occurs away from mapped faults (Zou and Fialko 2024).
References