the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
The estimation of porosity in Japan Trench plate boundary using low-resolution X-ray computed tomography (XCT) images and laboratory measurements
Abstract. X-ray computed tomography (XCT) is an advanced imaging technique that has been increasingly used in the past years because it can provide valuable information on internal structures of a rock sample in a non-invasive manner. The maximum resolution of lab-based XCT facilities is ~0.5 μm, which might be sufficient to capture macropores in some rocks (i.e., sandstone), but will result in underestimation of porosity in clay-rich sediments containing micro-and nano-scale pores. Furthermore, such high-resolution XCT facilities are quite expensive and not ubiquitous. In this study, we introduce a new methodology based on the K-means clustering algorithm to process of low-resolution XCT images, illustrating its capability through porosity analysis of drillcores obtained during Integrated Ocean Drilling Program (IODP) expedition 343. The cation exchange capacity (CEC) of the squeezed samples of the same cores was also measured and used to correct shipboard measurements of Moisture and Density (MAD) porosity for the effect of the water bound in the interlayer clay particles, thereby calculating interstitial porosity. The results indicate that the porosities estimated by our method are in agreement with these MAD_derived interstitial porosities in several cores acquired from the overthrusted sediments above the Japan trench plate boundary. Also, considering interstitial porosity as a realistic measurement of porosity, the results show that our semi-automatic method improves estimations compared with a manual thresholding segmentation, as the latter suffers from user subjectivity.
This preprint has been withdrawn.
-
Withdrawal notice
This preprint has been withdrawn.
-
Preprint
(1701 KB)
Interactive discussion
Status: closed
-
RC1: 'Comment on se-2021-150', Anonymous Referee #1, 24 Feb 2022
While the premise behind the MS (to provide better ways to quantify features in XCT data when the resolution of the image is too low to do so easily) is good, the MS as a whole isn't easy to follow (especially the technical aspects of the method being applied) and the mix of method development and science driven results doesn't improve the discussion about the suitability. The suitability and effectiveness of the method is limited and under tested against both the decisions taken in the MS (including key issues of the impact of filters on outcome, impact of final subjective step, and how does this method perform on this dataset further downsampled) and other published methods (comparing any new method against a badly chosen global threshold will suggest the new method is better). This means several of the key statements are unsupported by the data presented. Full comments are included on the pdf attached.
-
RC2: 'Comment on se-2021-150', Anonymous Referee #2, 31 Mar 2022
The authors have provided an interesting study using different methods to enhance the accuracy of the segmentation in CT images. The idea is very good but I think there are a few significant aspects which should be considered.
- The manuscript only talks about porosity and uses this as a reference for the CT segmentation. However, I think this is not enough. At least pore size distribution (PSD) measured by other lab techniques should be provided to compare with the PSD extracted from CT images. The author mentioned that there are many micropores and mesopores in the sample. The definition was not mentioned. I assume the author refers to the IUPAC classification, i.e. micropores < 2 nm and mesopores 2-50nm. The pixel size is 0.188mm, which is 3760 to over 100, 000 times larger than the pores. As the author mentioned, the grey value of one pixel is an average value of all the compositions. The compositions are very complicated. Even if a tiny change in the compositions will significantly change the average grey values. It is hard to believe the changes in grey value are caused by pores, which are 1/100, 000 of a pixel (although they might be many pores; this is not mentioned at all in this manuscript), not the mineral compositions.
- In clayey rocks, although the micropores and mesopores are the major components (in number or in frequency) in the pore system, they are not the major contributions to the total porosity. See Figure 2C in this paper: https://www.nature.com/articles/s41598-018-30153-x This means that the total volume of the pores below the resolution might be even smaller than the errors you have shown in your data.
- The authors mentioned a few times that the best resolution of CT is 0.5 um, but this is not true. It can achieve 50 nm. For example: 1) 1016/j.marpetgeo.2016.02.008 2) 10.1021/acs.energyfuels.0c03225 3) https://doi.org/10.1039/D0EE03651J
- Many of basic concepts, for example, CT (line 90-110), segmentation (line125-134), do not needed to be detailed explain in this manuscripts. There are many papers talking about the details already. Proper citations are enough.
Therefore, the research itself is interesting, but the results and the conclusions are not reliable owing to the unreliable validation. I suggest pore size distribution measured by other lab techniques (e.g. nitrogen adsorption) should be provided to compare with the PSD extracted from CT images. Also, SEM images or TEM images can be provided as evidence for the locations and distributions of the pores.
Citation: https://doi.org/10.5194/se-2021-150-RC2
Interactive discussion
Status: closed
-
RC1: 'Comment on se-2021-150', Anonymous Referee #1, 24 Feb 2022
While the premise behind the MS (to provide better ways to quantify features in XCT data when the resolution of the image is too low to do so easily) is good, the MS as a whole isn't easy to follow (especially the technical aspects of the method being applied) and the mix of method development and science driven results doesn't improve the discussion about the suitability. The suitability and effectiveness of the method is limited and under tested against both the decisions taken in the MS (including key issues of the impact of filters on outcome, impact of final subjective step, and how does this method perform on this dataset further downsampled) and other published methods (comparing any new method against a badly chosen global threshold will suggest the new method is better). This means several of the key statements are unsupported by the data presented. Full comments are included on the pdf attached.
-
RC2: 'Comment on se-2021-150', Anonymous Referee #2, 31 Mar 2022
The authors have provided an interesting study using different methods to enhance the accuracy of the segmentation in CT images. The idea is very good but I think there are a few significant aspects which should be considered.
- The manuscript only talks about porosity and uses this as a reference for the CT segmentation. However, I think this is not enough. At least pore size distribution (PSD) measured by other lab techniques should be provided to compare with the PSD extracted from CT images. The author mentioned that there are many micropores and mesopores in the sample. The definition was not mentioned. I assume the author refers to the IUPAC classification, i.e. micropores < 2 nm and mesopores 2-50nm. The pixel size is 0.188mm, which is 3760 to over 100, 000 times larger than the pores. As the author mentioned, the grey value of one pixel is an average value of all the compositions. The compositions are very complicated. Even if a tiny change in the compositions will significantly change the average grey values. It is hard to believe the changes in grey value are caused by pores, which are 1/100, 000 of a pixel (although they might be many pores; this is not mentioned at all in this manuscript), not the mineral compositions.
- In clayey rocks, although the micropores and mesopores are the major components (in number or in frequency) in the pore system, they are not the major contributions to the total porosity. See Figure 2C in this paper: https://www.nature.com/articles/s41598-018-30153-x This means that the total volume of the pores below the resolution might be even smaller than the errors you have shown in your data.
- The authors mentioned a few times that the best resolution of CT is 0.5 um, but this is not true. It can achieve 50 nm. For example: 1) 1016/j.marpetgeo.2016.02.008 2) 10.1021/acs.energyfuels.0c03225 3) https://doi.org/10.1039/D0EE03651J
- Many of basic concepts, for example, CT (line 90-110), segmentation (line125-134), do not needed to be detailed explain in this manuscripts. There are many papers talking about the details already. Proper citations are enough.
Therefore, the research itself is interesting, but the results and the conclusions are not reliable owing to the unreliable validation. I suggest pore size distribution measured by other lab techniques (e.g. nitrogen adsorption) should be provided to compare with the PSD extracted from CT images. Also, SEM images or TEM images can be provided as evidence for the locations and distributions of the pores.
Citation: https://doi.org/10.5194/se-2021-150-RC2
Viewed
HTML | XML | Total | BibTeX | EndNote | |
---|---|---|---|---|---|
574 | 189 | 37 | 800 | 29 | 39 |
- HTML: 574
- PDF: 189
- XML: 37
- Total: 800
- BibTeX: 29
- EndNote: 39
Viewed (geographical distribution)
Country | # | Views | % |
---|
Total: | 0 |
HTML: | 0 |
PDF: | 0 |
XML: | 0 |
- 1