3D Inversion of Magnetic Gradient Tensor Data Based on Convolutional Neural Networks

Round 1

Reviewer 1 Report

Neural networks provide a success in geophysical inversion, but are used in less degree for the magnetic field tensor gradient inversion. The high-precision magnetic field vector detection has been widely used in the fields of celestial magnetic field detection, aeromagnetic detection, marine magnetic field detection and geomagnetic navigation. In this manuscript a 3D magnetic gradient tensor (MGT) inversion method by using the convolutional neural network (CNN) has been. developed. Authors have also investigated at first a method to identify the geometric shape and depth and then further identify the physical property parameters. This stepwise identification method can suppress the influence of uncertainties and improve the recognition rate. The feasibility and accuracy of the algorithm by two synthetic simulation models was validated. It is important that the developed method was applied to the inversion of MGT data from the Tallawang magnetite diorite in Australia.

The abstract clearly and concisely summarizes the paper and state the main results. The manuscript is written clearly and consistently. All parts of the text, references and figures are necessary for the new results and main points to be understood. The manuscript is well organized and written. It provides new results which are important for the high-precision magnetic field vector detection. The manuscript meets the criteria for the publication in a full measure.

            I recommend the manuscript for a publication in the present form.

Author Response

Response:

We thank the reviewer for reading our paper carefully and giving the above positive comments.

We have carefully considered all the comments from the reviewers and revised our manuscript accordingly.

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

Equation (1). There is no explanation about "the magnetic field is calculated as [35,36]". What does mean magnetic field U. What about units of U? In elctromagnetic field we distinguish two potentials one is a scalr potential and the second one is a vector potential. Probably it is the scalar potential, but haw it is defined (there are different definitions which one was used by the Authors?). All this should be explained to the readers.

Table 1 demands more descripion about Horizontal position, Depth (m) -> (-10, 10)??? Sketch the objects and the whole region of interest, please. It help the reader to understand the problem.

The paper is very interesting but it is very hard to read it (at least for me).

Author Response

Response:

We would like to thank you for your careful reading , helpful comments, and giving the above positive comments, which has significantly improved the presentation of our manuscript.

We have made correction according to the Reviewer's comments.

In our revisions, we paid specific attention to 1) We rewrote Equation 1 and added parameter descriptions,  2) The prediction range for each parameter in Table 1 refers to the possible theoretical range of the parameter itself, and we arbitrarily reduce the theoretical range of the parameter, such as the shallow surface depth range was set to 40~360; the theoretical range of D was -180~180, and we chose -10~10 in order to reduce the sample generation workload. 3)  We have corrected some spelling errors and change some of the text descriptions. 4) Redrawn some of the figures, 5) Two references were added.

We look forward to hearing from you regarding our submission. We would be glad to respond to any further questions and comments  that you may have.

Please see the attachment.

Author Response File: Author Response.docx