the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 22 Mar 2023
G&M3D 1.0: an Interactive framework for 3D Model Construction and Forward Calculation of Potential Fields
Abstract. Building source models and performing forward calculations are the basis for data processing, analysis, and interpretation of geophysical data. However, open-source tools for flexibly constructing source models and forward modelling of the potential fields are still lacking. This paper developed a new MATLAB-based software – G&M3D to fill this gap. The software has two main functions: (1) constructing 3-D gravity and magnetic source models and (2) calculating and visualizing their gravity and magnetic fields. In the 3D-Modeling Module, rectangular prisms are used to approximate anomalous geologic bodies. Users can flexibly construct 3-D regular-shaped models with variable densities or magnetic parameters using the Sphere, Cylinder, and Cube tools, or build irregular-shaped models using the Irregular (Layer-Building) tool. On the other hand, the gravity anomalies, gravity gradients, total magnetic intensity, and magnetic gradients generated by the created 3-D sources can be rapidly calculated, visualized, and saved in the Forward-Modelling Module of G&M3D. In order to improve the efficiency of the gravity and magnetic forward calculations, the 2-D discrete convolution algorithm is improved and applied in the software for the forward modelling of the gravity and magnetic fields. Finally, we use G&M3D for the forward gravity modelling over a salt dome in Vinton Dome, southern Louisiana, U.S., which verifies its correctness and practicality.
This preprint has been withdrawn.
-
Withdrawal notice
This preprint has been withdrawn.
-
Preprint
(3496 KB)
Interactive discussion
Status: closed
-
RC1: 'Comment on gmd-2022-314', Anonymous Referee #1, 10 May 2023
Summary:
The paper discusses the development of a new MATLAB-based software called G&M3D 1.0, an open-source tool for constructing source models and forward modelling of potential fields.
The software is able to construct 3-D gravity and magnetic source models composed of rectangular prisms and calculate and visualize the corresponding gravity and magnetic fields.
The software is tested on synthetic data and is applied to forward modelling over a salt dome in Louisiana, US.Â
Practical applications demonstrate how the software can be used for research.Strengths:
Authors describe in details the basic methodology behind the software, the modelling workflow, and explain how the software can be used to build a salt dome model and calculate a gravity response from it.
Papers suggests four additional modelling strategies to improve the existing block circulant extension method (BCE), which is the main theoretical and practical achievement.Limitations:
Modelling with rectangular prisms can be considered as limitation, since majority of complex models require triangulated modelling domain. An attempt to represent such a domain with rectangular prisms leads to requirement of a much higher resolution and hence growing computational time, as well as required physical memory resources.
In my opinion, this limits applicability to real complex model geometries despite the provided example with a salt dome.
In practice, modellers have to deal with numerous layer intersections, e.g. with faults, where rectangular simplification is not practical.
It is unclear how four strategies suggested my the authors are improving the modelling, since there are no tests showing the results with and without application of those.Implications:
The open-source tool like the one suggested in the paper is valuable for researchers or teachers who have access to MATLAB software and can be used for quick testing of model geometries, as well as for simple visualization of the modelling process.Major comments:
1. P1 L28-29: I do not agree with the statement that there are lacking open-source tools for model construction/forward modelling of potential fields.Â
Later in the manuscript authors cite several open-source tools in P2 L60-61, P3 L66-68.
Apart from those, there more open-source tools.
In particular:
- Fatiando a Terra (Uieda et al., 2013), a set of Python libraries for data processing, modelling, and inversion of potential fields.
- pyGIMLi (Rücker et al., 2017) - an open-source software package for geophysical inversion and modelling for triangulated models.
Also, IGMAS+ (e.g., Anikiev et al. 2020) is available - a free (although not open source) practical tool that allows interactive construction of triangulated 3D models.2. It is unclear from P3 L69-75, what is the main goal of the paper. If it is creation of a software, more tests again other existing tools are needed, in terms of speed, accuracy, memory consumption.
If it is introduction of the new strategies (see the next comment), then one should add an analysis of their application.3. P8 L179-196: The authors suggest four new strategies, however in the synthetic tests they do not show how exactly these strategies improve the. I would suggest to add specific tests with and without the novel strategies for comparison.
4. P17 L 312: It is not clear how the results presented in the paper are consistent with the results in the other paper cited (Ennen 2012). It is important to add a direct comparison (misfit plot) between the modelled data in both approaches, rather than providing only the results generated by the new software.
Apart from that, I would suggest the authors to add more tests with comparison of their modelling results with other existing open-source (or freely available) tools (see the first comment).5. P18 L 320: "independent desktop sofware" - I was not able to find how the provided software can be used independently of MATLAB. First, it is not explained in the software manual. Second, to my knowledge, to deploy the software in MATLAB in a form of independent installer one needs to have the MATLAB Compiler, which is not free for everybody. The installer then will have the MATLAB Runtime libraries integrated. I was able to find only the app installer, which obviously requires MATLAB to install/run it.Â
Therefore, I would remove the mentioned part of the sentence.Minor comment:
Please add the refence to G&M3D 1.0 as a proper reference to the software in the list of references.
It was hard to find the DOI in the text (it is only in code availability, which is the end of the manuscript).References:
- Uieda, L., V. C. Oliveira Jr, and V. C. F. Barbosa (2013), Modeling the Earth with Fatiando a Terra, Proceedings of the 12th Python in Science Conference, pp. 91-98. doi:10.25080/Majora-8b375195-010
- Rücker, C., Günther, T., Wagner, F.M., 2017. pyGIMLi: An open-source library for modelling and inversion in geophysics, Computers and Geosciences, 109, 106-123, doi: 10.1016/j.cageo.2017.07.011.
- Anikiev D., Götze, H.-J., Meeßen, C., Plonka, C., Scheck-Wenderoth, M., Schmidt, S. (2020), IGMAS+: Interactive Gravity and Magnetic Application System. V. 1.3. GFZ Data Services. https://doi.org/10.5880/GFZ.4.5.igmas.v.1.3Citation: https://doi.org/10.5194/gmd-2022-314-RC1 -
RC2: 'Comment on gmd-2022-314', Anonymous Referee #2, 25 May 2023
General comments
Over the years numerous open-source tools have been developed and made available to the community to not only forward model but also invert potential field data (e.g. SimPEG (https://simpeg.xyz/) and Fatiando (https://www.fatiando.org/)) This manuscript introduces G&M3D a MATLAB based open source implementation of a 3D gravity forward modelling code where the subsurface is represented using prisms. The forward problem is solved by implementing the 2-D discrete convolution algorithm introduced by Chen and Liu (2019) in MATLAB.
While the authors identify and employ four strategies to increase the efficiency of the forward calculation it is not clear to me what the impact is on the overall efficiency. If improved efficiency of the forward computation is the focus of the manuscript, then the implementation presented here could be easily compared with other open-source implementations. It is currently not possible to for me assess how the methodology introduced here represents an improvement over Chen and Liu (2019).
The second focus area of the paper is the description/introduction of a graphical user interface to construct models for potential field forward modelling. While this could be of value for practitioners; when compared to existing open-source potential field modelling packages the lack of an ability to invert the data and the dependency on MATLAB appear to be major limitations.
Manuscripts to be published in GMD are expected to represent a sufficiently substantial advance in modelling science. When compared to other open-source potential field forward modelling and inversion packages it is unfortunately not clear to me where and how this work sufficiently substantially advances potential field data analysis. Perhaps the value of what has been developed may lie in it being a teaching resource.
References
Chen, L., & Liu, L. (2019). Fast and accurate forward modelling of gravity field using prismatic grids. Geophysical Journal International, 216(2), 1062–1071. https://doi.org/10.1093/gji/ggy480
Citation: https://doi.org/10.5194/gmd-2022-314-RC2 -
EC1: 'Comment on gmd-2022-314', Thomas Poulet, 26 May 2023
Dear authors,
Given the reviewer’s comments regarding the scientific significance of this manuscript in its current form, note that I am discouraging the submission of a revised manuscript which would not clearly address all points satisfactorily within the limited time available for revisions. I think that the limited applicability to real complex model geometries due to quadrilaterals and the lack of comparison with existing software (e.g. from the work of Chen and Liu, 2019, https://doi.org/10.1093/gji/ggy480) might present some fundamental shortcomings which might make it unrealistic to provide a revised version addressing all points within a short timeframe. I would therefore recommend an alternative pathway for this study at this stage.
I am looking forward to hearing your comments, feel free to contact me if you have any question.
Best regards,
Thomas Poulet.
Citation: https://doi.org/10.5194/gmd-2022-314-EC1
Interactive discussion
Status: closed
-
RC1: 'Comment on gmd-2022-314', Anonymous Referee #1, 10 May 2023
Summary:
The paper discusses the development of a new MATLAB-based software called G&M3D 1.0, an open-source tool for constructing source models and forward modelling of potential fields.
The software is able to construct 3-D gravity and magnetic source models composed of rectangular prisms and calculate and visualize the corresponding gravity and magnetic fields.
The software is tested on synthetic data and is applied to forward modelling over a salt dome in Louisiana, US.Â
Practical applications demonstrate how the software can be used for research.Strengths:
Authors describe in details the basic methodology behind the software, the modelling workflow, and explain how the software can be used to build a salt dome model and calculate a gravity response from it.
Papers suggests four additional modelling strategies to improve the existing block circulant extension method (BCE), which is the main theoretical and practical achievement.Limitations:
Modelling with rectangular prisms can be considered as limitation, since majority of complex models require triangulated modelling domain. An attempt to represent such a domain with rectangular prisms leads to requirement of a much higher resolution and hence growing computational time, as well as required physical memory resources.
In my opinion, this limits applicability to real complex model geometries despite the provided example with a salt dome.
In practice, modellers have to deal with numerous layer intersections, e.g. with faults, where rectangular simplification is not practical.
It is unclear how four strategies suggested my the authors are improving the modelling, since there are no tests showing the results with and without application of those.Implications:
The open-source tool like the one suggested in the paper is valuable for researchers or teachers who have access to MATLAB software and can be used for quick testing of model geometries, as well as for simple visualization of the modelling process.Major comments:
1. P1 L28-29: I do not agree with the statement that there are lacking open-source tools for model construction/forward modelling of potential fields.Â
Later in the manuscript authors cite several open-source tools in P2 L60-61, P3 L66-68.
Apart from those, there more open-source tools.
In particular:
- Fatiando a Terra (Uieda et al., 2013), a set of Python libraries for data processing, modelling, and inversion of potential fields.
- pyGIMLi (Rücker et al., 2017) - an open-source software package for geophysical inversion and modelling for triangulated models.
Also, IGMAS+ (e.g., Anikiev et al. 2020) is available - a free (although not open source) practical tool that allows interactive construction of triangulated 3D models.2. It is unclear from P3 L69-75, what is the main goal of the paper. If it is creation of a software, more tests again other existing tools are needed, in terms of speed, accuracy, memory consumption.
If it is introduction of the new strategies (see the next comment), then one should add an analysis of their application.3. P8 L179-196: The authors suggest four new strategies, however in the synthetic tests they do not show how exactly these strategies improve the. I would suggest to add specific tests with and without the novel strategies for comparison.
4. P17 L 312: It is not clear how the results presented in the paper are consistent with the results in the other paper cited (Ennen 2012). It is important to add a direct comparison (misfit plot) between the modelled data in both approaches, rather than providing only the results generated by the new software.
Apart from that, I would suggest the authors to add more tests with comparison of their modelling results with other existing open-source (or freely available) tools (see the first comment).5. P18 L 320: "independent desktop sofware" - I was not able to find how the provided software can be used independently of MATLAB. First, it is not explained in the software manual. Second, to my knowledge, to deploy the software in MATLAB in a form of independent installer one needs to have the MATLAB Compiler, which is not free for everybody. The installer then will have the MATLAB Runtime libraries integrated. I was able to find only the app installer, which obviously requires MATLAB to install/run it.Â
Therefore, I would remove the mentioned part of the sentence.Minor comment:
Please add the refence to G&M3D 1.0 as a proper reference to the software in the list of references.
It was hard to find the DOI in the text (it is only in code availability, which is the end of the manuscript).References:
- Uieda, L., V. C. Oliveira Jr, and V. C. F. Barbosa (2013), Modeling the Earth with Fatiando a Terra, Proceedings of the 12th Python in Science Conference, pp. 91-98. doi:10.25080/Majora-8b375195-010
- Rücker, C., Günther, T., Wagner, F.M., 2017. pyGIMLi: An open-source library for modelling and inversion in geophysics, Computers and Geosciences, 109, 106-123, doi: 10.1016/j.cageo.2017.07.011.
- Anikiev D., Götze, H.-J., Meeßen, C., Plonka, C., Scheck-Wenderoth, M., Schmidt, S. (2020), IGMAS+: Interactive Gravity and Magnetic Application System. V. 1.3. GFZ Data Services. https://doi.org/10.5880/GFZ.4.5.igmas.v.1.3Citation: https://doi.org/10.5194/gmd-2022-314-RC1 -
RC2: 'Comment on gmd-2022-314', Anonymous Referee #2, 25 May 2023
General comments
Over the years numerous open-source tools have been developed and made available to the community to not only forward model but also invert potential field data (e.g. SimPEG (https://simpeg.xyz/) and Fatiando (https://www.fatiando.org/)) This manuscript introduces G&M3D a MATLAB based open source implementation of a 3D gravity forward modelling code where the subsurface is represented using prisms. The forward problem is solved by implementing the 2-D discrete convolution algorithm introduced by Chen and Liu (2019) in MATLAB.
While the authors identify and employ four strategies to increase the efficiency of the forward calculation it is not clear to me what the impact is on the overall efficiency. If improved efficiency of the forward computation is the focus of the manuscript, then the implementation presented here could be easily compared with other open-source implementations. It is currently not possible to for me assess how the methodology introduced here represents an improvement over Chen and Liu (2019).
The second focus area of the paper is the description/introduction of a graphical user interface to construct models for potential field forward modelling. While this could be of value for practitioners; when compared to existing open-source potential field modelling packages the lack of an ability to invert the data and the dependency on MATLAB appear to be major limitations.
Manuscripts to be published in GMD are expected to represent a sufficiently substantial advance in modelling science. When compared to other open-source potential field forward modelling and inversion packages it is unfortunately not clear to me where and how this work sufficiently substantially advances potential field data analysis. Perhaps the value of what has been developed may lie in it being a teaching resource.
References
Chen, L., & Liu, L. (2019). Fast and accurate forward modelling of gravity field using prismatic grids. Geophysical Journal International, 216(2), 1062–1071. https://doi.org/10.1093/gji/ggy480
Citation: https://doi.org/10.5194/gmd-2022-314-RC2 -
EC1: 'Comment on gmd-2022-314', Thomas Poulet, 26 May 2023
Dear authors,
Given the reviewer’s comments regarding the scientific significance of this manuscript in its current form, note that I am discouraging the submission of a revised manuscript which would not clearly address all points satisfactorily within the limited time available for revisions. I think that the limited applicability to real complex model geometries due to quadrilaterals and the lack of comparison with existing software (e.g. from the work of Chen and Liu, 2019, https://doi.org/10.1093/gji/ggy480) might present some fundamental shortcomings which might make it unrealistic to provide a revised version addressing all points within a short timeframe. I would therefore recommend an alternative pathway for this study at this stage.
I am looking forward to hearing your comments, feel free to contact me if you have any question.
Best regards,
Thomas Poulet.
Citation: https://doi.org/10.5194/gmd-2022-314-EC1
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 547 | 197 | 31 | 775 | 27 | 33 |
- HTML: 547
- PDF: 197
- XML: 31
- Total: 775
- BibTeX: 27
- EndNote: 33
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1