Machine Learning-Based Mineral Quantification from Lower Cambrian Shale in the Sichuan Basin: Implications for Reservoir Quality

Comments 1: [add“and mineral mapping”.]

Response 1: The phrase “and mineral mapping” has been added to the appropriate location in the original text Thank you for pointing this out. We agree with this comment.

Comments 2: [add the reference of Honarmand et al., 2024.]

Response 2: We have added this reference in both the text and the list.

Comments 3: Modify Figure 3

Response 3: We have modified the Figure 3, add the space, and the word “assigned”

Comments 4: Please omit the highlight

Response 4: The highlight has been deleted.

Comments 5: add “core samples or”

Response 5: The phrase “core samples or” has been added to the appropriate location in the original text Thank you for pointing this out. We agree with this comment.

Comments 1: I recommend that the topic be revised. It should indicate the implications for mineral quantification. I believe it has to do with reservoir characterisation and hydraulic fracturing. Also, the topic must reflect the novelty - e.g " A Machine Learning approach for mineral quantification using major element log data from........: Implications for reservoir quality studies". Consider revising the title to reflect the aim and novelty of the study.

Response 1: The title has been modified to “Machine Learning-Based Mineral Quantification from Lower Cambrian Shale in Sichuan Basin: Implications for Reservoir Quality” Thank you for pointing this out. We agree with this comment.

Comments 2: Under "3.2.1 Data interpretation" - You have resorted to assigning 0 to negative values of elements. However, it is unclear how other values below the instrument detection limit were recorded. For example, a value of <0.1 or > 0.1. These refer to existing elemental contents that could not be detected with the instrumentation, but it does not mean they don't exist altogether to be assigned 0.

Response 2: Thank you for your question. In XRF analysis, the instrument provides a detection limit, representing the minimum element concentration that can be reliably measured. If the actual concentration of certain elements is below this detection limit, the XRF instrument may report a value below a specified threshold (e.g., <0.1%). For such values, we apply the following approach: if the value is >0.05%, we round it to 0.1%, and if it is <0.05%, we assign it a value of 0. For values in the range of 0.1% to 100%, the measured data is typically used as is. For values exceeding the maximum detection limit (e.g., >100%), remeasurement of the sample and appropriate dilution methods may be necessary to ensure the data falls within the instrument’s measurable range.

Comments 3: Please add two lowermost rows on Table 1 for the mean value and standard deviation for each element/mineral

Response 3: Thank you very much for your valuable suggestions. We have calculated the mean and standard deviation for all elements and mineral percentage contents, which are shown in the last two rows of Table 1. During model training, we will standardize all element data to have a mean of 0 and a standard deviation of 1, ensuring that different features have the same scale, thereby minimizing the impact of scale differences on model training.

Comments 4: Please refer to some comments embedded on the attached manuscript file

Response 4: we have made corresponding revisions in response to each of the inline comments you provided. All suggested modifications have been implemented as recommended. Please find the revised manuscript attached, where these changes are reflected.

1. Summary

Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding revisions in track changes in the re-submitted files.

2. Point-by-point response to Comments and Suggestions for Authors

Comments 1: I recommend expanding the references, especially in the Introduction section. For example: lines 38–40 ‘These fine-grained lithologies present inherent challenges: their homo-generous texture obscures fabric details in core samples, while subtle geochemical variations often remain undetectable through conventional macroscopic’, I suggest adding: Pszonka, J.; Götze, J. Quantitative estimate of interstitial clays in sandstones using Nomarski differential interference contrast (DIC) microscopy and image analysis. Journal of Petroleum Science and Engineering 2018, 161, 582-589. https://doi.org/10.1016/j.petrol.2017.11.069

lines: 54-55 ’Machine learning (ML) and artificial intelligence (AI) have demonstrated significant potential in geological applications’ I suggest adding: Pszonka, J.,Godlewski, P., Fheed, A., Dwornik, M., Schulz, B., Wendorff, M. (2024) Identificationand quantification of intergranular volume using SEM automated mineralogy.Marine and Petroleum Geology, 162, 106708. https://doi.org/10.1016/j.marpetgeo.2024.106708

Response 1: The references have been added to the appreciated places.

Comments 2: At the end of the Introduction section, clearly state the objective of your study and place it within a broader context.

Response 2: We have added a section at the end of the preface, clearly stating our research objectives and placing them in a broader context, as follows: This study specifically addresses two critical gaps in unconventional resource development: (1) the operational need for rapid mineralogical profiling during wellsite decision-making, and (2) the economic imperative to reduce reliance on specialized laboratory analyses. By integrating ANN-based interpretation of XRF measurements with established XRD benchmarks, we establish a workflow that transforms elemental abundance data into actionable mineralogical intelligence. The developed methodology enables generation of continuous mineral profiles at resolutions exceeding conventional laboratory sampling intervals (typically 30-100 cm), while maintaining 70-80% cost reduction compared to traditional XRD-based approaches. Within the broader context of energy transition technologies, this advancement supports more sustainable exploitation of unconventional reservoirs through precision targeting of lateral sections and minimization of unnecessary hydraulic fracturing stages. Furthermore, the machine learning framework demonstrates how legacy core datasets can be revitalized through digital transformation, creating new value from historical geological archives that were previously underutilized due to analytical cost constraints.

Comments 3: The Data and Methods section should be shortened. It seems that some of the results have been presented here. I recommend to change the title of this section: ‘Material and Methods’ instead of ‘Data and Methods’. The Materials and Methods section in a scientific article should provide a detailed and precise description of how the study was conducted. It should include information on the materials, equipment, sample collection, experimental design, and analytical techniques used. This section must be written with enough detail to allow other researchers to replicate the study. Any modifications to established methods should be clearly stated, and references to standard procedures should be provided where applicable. Transparency and accuracy in this section ensure the reliability and reproducibility of the research.

Response 3: The “Data and Methods” has been modified to “Materials and Methods”. We have enhanced the Materials and Methods section to improve methodological transparency.

Comments 4: I strongly recommend separating the Results and Discussion sections. The Results section should present the study’s findings objectively, without interpretation or discussion. It should include clear descriptions of data, supported by relevant figures, tables, and statistical analyses. The results should be structured logically, following the order of research objectives or hypotheses. This section should highlight key trends, relationships, and patterns observed in the data while avoiding redundancy. The goal is to provide a clear and concise presentation of the study’s empirical evidence, which will later be analyzed in the Discussion section.

Response 4: Results ow exclusively presents the empirical findings, including mineralogical quantification results (R² values, loss curves, and validation figures) and geological validation data (well S1 mineral profiles). All interpretative analyses and methodological discussions have been removed from this section to ensure objective presentation of the study's outcomes.

Discussion has been expanded to comprehensively address: The technical advantages of our XRF-XRD methodology (e.g., high-resolution logging, cost-efficiency, non-destructive sampling); Current limitations (e.g., small XRD dataset size, spatial sampling discrepancies); Challenges specific to quartz prediction (e.g., Si element interference in organic-rich shale systems); Future optimization strategies (data expansion, feature engineering, transfer learning).

Comments 5: Writing the Conclusions section in bullet points is highly recommended in scientific articles because it enhances clarity, readability, and impact. Bullet points allow for a concise and structured presentation of the key findings, making it easier for readers to quickly grasp the main takeaways of the study. This format helps avoid unnecessary repetition and ensures that each conclusion is clearly defined and distinct. Additionally, it improves accessibility for a wider audience, including researchers skimming through the paper. A well-structured conclusions section also facilitates citation and reference to specific findings, increasing the study’s visibility and influence within the scientific community.

Response 5: As recommended, we have revised the Conclusions section to adopt bullet-point formatting, which significantly enhances the presentation of key findings. This revision aligns with your emphasis on scientific communication best practices and ensures our conclusions are impactful and memorable. Please see the attachment manuscript.