Geophysical research: article

Application of machine learning methods in tomography of receiving functions
Aleshin1,2
E.G. Kozlovskaya3
I.V. Malygin1
1 Schmidt Institute of Physics of the Earth, Russian Academy of Sciences, Moscow, Russia 2 Geophysical Center, Russian Academy of Sciences, Moscow, Russia 3 Oulu Mining School Faculty of Technology University of Oulu, Finland
Journal: Geophysical research
Tome: 23
Number: 1
Year: 2022
Pages: 49-61
UDK: 550.8.05+004.85
DOI: 10.21455/gr2022.1-4
Full text
Keywords: 3D image, receiving functions, machine learning method, nearest neighbor method, Fennoscandia, postglacial uplift, near-surface layer of low S-wave velocities
Аnnotation: A method for constructing a three-dimensional digital seismic model from a set of one-dimensional dependences of the elastic properties of the medium on depth is described. Such problem arises, for example, when interpreting interborehole measurements. A similar problem is also relevant when processing data from passive seismic experiments, when a significant number of seismic stations are installed close to each other with-in the studied region. If we use the method of receiving functions to process remote earthquakes recorded at sta-tions, then from the record at each of the stations we can obtain a velocity section – the dependence of the elastic properties of the medium under the station on depth. It is usually assumed that the medium directly under the station is laterally homogeneous. In this case, the dependence of the elastic parameters on the vertical coordinate can be parametrized by a set of flat layers. The presented method allows to build a three-dimensional seismic image of the region under study using a set of such models. The main challenge is the strong anisotropy of the spatial distribution of the input data. The distance between stations in experiments of this kind is determined by the first Fresnel zone of teleseismic phases and averages 50–100 km. At the same time, the distance between the values specified by the layered model is an order of magnitude less. It is shown that the problem can be solved by a scaling transformation of the horizontal coordinates. To determine the scale factor, methods developed in the theory of machine learning were used. This method is applicable when the elastic parameters of the medium change smoothly between stations. In other words, there are no sub-vertical faults between the stations. As an illustration, a three-dimensional digital model of the southern part of Fennoscandia was built according to the data of the European passive seismic experiment SVEKALAPKO. The obtained result made it possible to reveal some structural features of the upper crust in the area of contact between Archean and Proterozoic rocks caused by postglacial relaxation.
Bibliography: Aleshin I.M., Malygin I.V., Machine learning approach to inter-well radio wave survey data imaging, Russian Journal of Earth Sciences, 2019, vol. 19, no. 3, pp. ES30031-ES30036.

Aleshin I.M., Vaganova N.V., Kosarev G.L., Malygin I.V., The crust properties in Fennoscandia resulted from kNN-analysis of receiver function inversions, Geofizicheskie issledovaniya (Geophysical Research), 2019, vol. 20, no. 4, pp. 25-39. [In Russian].

Bannister S., Bryan C.J., Bibby H.M., Shear wave velocity variation across the Taupo Volcanic Zone, New Zealand, from receiver function inversion, Geophysical Journal International, 2004, vol. 159, no. 1,

pp. 291-310.



Bock G., Achauer U., Alinaghi A., Ansorge J., Bruneton M., Friederich W., Grad M., Guterch A., Hjelt S.-E., Hyvönen T., Ikonen J.-P., Kissling E., Komminaho K., Korja A., Heikkinen P., Kozlovskaya E., Nevsky M.V., Pavlenkova N., Pedersen H., Plomerová J., Raita T., Riznitchenko O.Yu., Roberts R.G., Sandoval S., Sanina I.A., Sharov N., Tiikkainen J., Volosov S.G., Wielandt E., Wylegalla K., Yliniemi J., Yurov Y., Seismic probing of Fennoscandian lithosphere, Eos, Transactions American Geophysical Un-ion, 2001, vol. 82, no. 50, pp. 621-629.

Beller S., Monteiller V., Combe L., Operto S., Nolet G., On the sensitivity of teleseismic full-waveform inversion to earth parametrization, initial model and acquisition design, Geophysical Journal International, 2018, vol. 212, no. 2, pp. 1344-1368.

Chiu P-W., Naim A.M., Lewis K.E., Bloebaum C.L., The hyper-radial visualisation method for multi-attribute decision-making under uncertainty, International Journal of Product Development, 2009, vol. 9, no. 1-3, pp. 4-31.

Grad M., Luosto U., Fracturing of the crystalline uppermost crust beneath the SVEKA profile in Central Finland, Geophysica, 1992, vol. 28, no. 1-2, pp. 53.

Guo Zh., Chen Y.J., Ning J., Feng Y., Grand S.P., Niu F., Kawakatsu H., Tanaka S., Obayashi M., Ni J., High resolution 3-D crustal structure beneath NE China from joint inversion of ambient noise and receiver functions using NECESSArray data, Earth and Planetary Science Letters, 2015, vol. 416, pp. 1-11.

Hastie T., Tibshirani R., Friedman J., The Elements of Statistical Learning: Data Mining, Inference, and Predic-tion, New York: Springer-Verlag, 2009, 746 p.

Kosarev G.L., Kind R., Sobolev S.V., Yuan X., Hanka W., Oreshin S.I., Seismic Evidence for a Detached Indian Lithospheric Mantle Beneath Tibet, Science, 1999, vol. 283, no. 5406, pp. 1306-1309.

Kozlovskaya E., Kosarev G.L., Aleshin I.M., Riznitchenko O.Yu., Sanina I.A., Structure and composition of the crust and upper mantle of the Archean-Proterozoic boundary in the Fennoscandian shield obtained by joint inversion of receiver function and surface wave phase velocity of recording of the SVEKALAPKO array, Geophysical Journal International, 2008, vol. 175, no. 1, pp. 135-152.

Laske G., Masters G., Ma Zh., Pasyanos M., Update on CRUST1.0 – A 1-degree global model of Earth’s crust, Geophys. Res. Abstr., 2013, vol. 15, pp. 2658.

Pedersen H., Campillo M., Depth dependance of Q beneath the Baltic Shield inferred from modeling of short pe-riod seismograms, Geophysical Research Letters, 1991, vol. 18, no. 9, pp. 1755-1758.

Ryberg T., Weber M., Receiver function arrays: a reflection seismic approach, Geophysical Journal Internation-al, 2000, vol. 141, no. 1, pp. 1-11.

Sibson R., A vector identity for the Dirichlet tessellation, Mathematical Proceedings of the Cambridge Philo-sophical Society, 1980, vol. 87, no. 1, pp. 151-155.

Sosa A., Thompson L., Velasco A.A., Romero R., Herrmann R.B., 3-D structure of the Rio Grande Rift from

1-D constrained joint inversion of receiver functions and surface wave dispersion, Earth and Planetary Science Letters, 2014, vol. 402, pp. 127-137.

Thurber C., Ritsema J., Theory and Observations – Seismic Tomography and Inverse Methods, Treatise on Geo-physics (Second Edition), 2015, vol. 1, pp. 307-337. https://doi.org/10.1016/B978-0-444-53802-4.00009-9

Vinnik L.P., Receiver Function Seismology, Izvestiya, Physics of the Solid Earth, 2019, vol. 55, no. 1, pp. 12-21.

Vinnik L.P., Deng Y., Kosarev G.L., Oreshin S.I., Zhang Zh., Makeyeva L.I., Sharpness of the 410-km disconti-nuity from the P410s and P2p410s seismic phases, Geophysical Journal International, 2020, vol. 220, no. 2, pp. 1208-1214.

Vinnik L.P., Reigber Ch., Aleshin I.M., Kosarev G.L., Kaban M.K., Oreshin S.I., Roecker S., Receiver function tomography of the central Tien Shan, Earth and Planetary Science Letters, 2004, vol. 225, no. 1-2, pp. 131-146.

Virieux J., Operto S., An overview of full-waveform inversion in exploration geophysics, Geophysics, 2009, vol. 74, no. 6, pp. WCC1-WCC26. DOI: 10.1190/1.3238367

Wang X., Wei S., Wang Y., Maung Maung P., Hubbard J., Banerjee P., Huang B.-S., Kyaw M.O., Bodin T., Foster A., Almeida R., A 3-D Shear Wave Velocity Model for Myanmar Region, Journal of Geophysical Research: Solid Earth, 2019, vol. 124, no. 1, pp. 504-526.

Ward K.M., Zandt G., Beck S.L., Douglas H.C., McFarlin H., Heather Seismic imaging of the magmatic under-pinnings beneath the Altiplano-Puna volcanic complex from the joint inversion of surface wave disper-sion and receiver functions, Earth and Planetary Science Letters, 2014, vol. 404, pp. 43-53.

Yanovskaya T.B., On the theory of the microseismic sounding method, Izvestiya, Physics of the Solid Earth, 2017, vol. 53, no. 6, pp. 819-824.

Yoo H.J., Herrmann R.B., Cho K.H., Lee K., Imaging the Three-Dimensional Crust of the Korean Peninsula by Joint Inversion of Surface-Wave Dispersion and Teleseismic Receiver Functions, Bulletin of the Seismo-logical Society of America, 2007, vol. 97, no. 3, pp. 1002-1011.