Open AccessThe correlation analysis of spatial anomalies of computed and measured fields by the GIS-ENDDB system
Author(s) -
Anna Mikheeva
Publication year - 2020
Publication title -
rossijskij sejsmologičeskij žurnal
Language(s) - English
Resource type - Journals
eISSN - 2782-1609
pISSN - 2686-7907
DOI - 10.35540/2686-7907.2020.1.10
Subject(s) - histogram , correlation , spatial analysis , spatial distribution , field (mathematics) , geology , geophysics , mathematics , remote sensing , computer science , geometry , artificial intelligence , pure mathematics , image (mathematics)
The GIS-ENDDB expert-information system includes the global seismic catalogs and measurement data for various geophysical fields. Statistical processing of these data and visualization of the distributing characteristics of the seismic-geodynamic process in the form of isolines make it possible to reveal the spatial relationship between the seismicity distribution and anomalies of geophysical fields. The first attempts to identify this relationship were made with L.S. Sokolova in 2016 based on a visual comparison of maps obtained as a result of applying methods of various data analysis (in particular, the isolines maps of the maximum magnitude of recorded earth-quakes Mmax) for different samples of seismological catalogs on the one hand, and various modifications of the measured geophysical fields (in this case, the regional component of heat flow (HF) field) on the other hand. These results were more likely qualitative, but later a correlation analysis function was implemented into GIS-ENDDB for the quantitative comparison of the fields. The correlation analysis is carried out on a histogram of the distributing number of cells (nodes) NC from the difference ∆F of the reduced values of the fields F1 and F2 (relative to the arithmetic mean Fi.OCP) in the cells that are equal for the compared fields. To identify non-random variations, NC histograms (∆F) are given against the background of the normal distribution curve, which allows not only visually (by the deviation of the first columns from the norm), but also quantitatively to establish the degree of correlation. For the studied territories, histograms have obtained that show: 1. a moderate correlation of HF and Mmax fields in the North American continent (NEIC catalog, M≥5, cell 3×4); 2. no correlation across Asia - Far East region (COMPLEX, M≥4, H≥50 km, cell 6×8); and 3. high correlation in a rectangular region covering the Inside-Asian mountain belt (30-40N; 66-84E., COMPLEX, M≥5, H≥50 km, cell 1×1).