Seismic magnitude homogenization by unified regression method and its effect on seismicity parameters

• Published in: Journal of seismology Vol. 28; no. 2; pp. 417 - 437
• Main Authors: Patel, Rohit Singh, Sinha, Ajay Kumar
• Format: Journal Article
• Language: English
• Published: Dordrecht Springer Netherlands 01-04-2024; Springer Nature B.V
• Subjects: Catalogues; Earth and Environmental Science; Earth Sciences; Earthquakes; Empirical analysis; Error analysis; Geographical variations; Geological hazards; Geophysics/Geodesy; Geotechnical Engineering & Applied Earth Sciences; Hazard assessment; Homogenization; Hydrogeology; Independent variables; Mathematical models; Nonlinear systems; Nonlinearity; Parameters; Predation; Regression; Seismic activity; Seismic hazard; Seismicity; Seismology; Structural Geology; Tectonics; India; Ordinary least-square regression; Non-linear regression; Earthquake magnitude; Homogenous earthquake catalog; Chi-square regression; Seismicity parameters
• ISSN: 1383-4649; 1573-157X
• DOI: 10.1007/s10950-024-10197-5

A homogeneous earthquake catalogue is necessary to conduct a thorough seismological investigation and improve seismic hazard assessment. Therefore, the present study intends to determine the influence of magnitude homogenization based on different regression methods on the seismicity parameters. The earthquake catalog of the Bihar, India region has been used for the study. An events data catalogue consisting of 4001 events has been compiled over the period 1911–2023. The data is primarily sourced from international and local databases have been assembled to provide a comprehensive overview of seismic activity within the study area. Empirical relationships have been derived between moment magnitude and commonly used local and global magnitude scales by performing non-linear regressions using chi-square regression method (CSQ) and ordinary least-square regression method (OLS). Measurement error for both dependent and independent variables of various data sets was calculated for the CSQ process. In order to address the non-linear nature of the correlations, we used curvilinear models, namely, the exponential model (EXP), which was recently introduced by the authors of the global catalogue supported by the global earthquake model (GEM) foundation. The use of CSQ method reduces the bias created by OLS method for defining the nonlinearity of the relationship. Seismicity parameters, including the magnitude of completeness (
M
c
) and
a
and
b
values of the Gutenberg-Richter recurrence model, have been calculated for homogenised undeclustered and declustered catalogue spanning five specific time periods. Declustering has been performed to isolate the foreshocks and aftershocks from the mainshocks. The observed changes in the
M
c
and
b
values are found to be dependent on the introduction of more recent data over time. The study area has been divided into eight seismogenic zones based on the geographic variations in earthquake occurrence and the prevailing tectonic conditions. The estimation of
M
c
,
b
, and
a
values has been conducted for each zone, and the analysis has been performed to examine the variations in these parameters. Derived equations serve as valuable tools for assessing and analysing seismic hazards in the north India and the central Himalayas region.