Geophysical research: article

I. Aleshin1,2,3
N. Vaganova2
G. Kosarev1
I. Malygin1
1 Schmidt Institute of Physics of the Earth, Russian Academy of Sciences
2 Laverov Federal Center for Integrated Arctic Research, Russian Academy of Sciences
3 Geophysical Center, Russian Academy of Sciences
Journal: Geophysical research
Tome: 20
Number: 4
Year: 2019
Pages: 25-39
UDK: 550.3; 550.347.62; 004.85
DOI: 10.21455/gr2019.4-2
Full text
Keywords: kNN-алгоритм, machine learning, kNN-algorithm, receiver function, Fennoscandia, Baltic shield, Earth crust, Moho boundary shape
Аnnotation: Our goal is an analysis of the structure of the crust of the northern part of the Baltic shield. We used data obtained by receiver functions approach. We used information obtained in previous studies of this region, supplemented by new calculations and measurements. Data for more than 60 permanent and temporary stations was available. To build the spatial distributions of the quantities of interest, the machine learning approach was used. We utilized one of the most effective methods. It can be used to analyze continuous data. Two aspects of the problem are discussed. First, a digital model of the spatial distribution of the crustal thickness in Fennoscandia and adjacent regions was calculated. We defined the crustal thickness as a distance from the Earth surface to the boundary of Mohorovichich. The shape of this boundary is very complicated due to ancient geodynamic processes. Previous studies have shown the lack of correlation of the spatial structure of the Mohorovichich boundary with the boundary of geological structures of different ages. Our work confirms this fact, with the exception of a small area in the north of the border. This part of boundary corresponds to sudden changes in the thickness of the crust from 50 km to 40 km. There is practically no sediment in the region under study. In spite of this, there are areas covered by a layer with low values of the transverse seismic wave velocity V