Global regression relations for conversion of surface wave and body wave magnitudes to moment magnitude

A homogenous earthquake catalog is a basic input for seismic hazard estimation, and other seismicity studies. The preparation of a homogenous earthquake catalog for a seismic region needs regressed relations for conversion of different magnitudes types, e.g. m (b) , M (s) , to the unified moment magnitude M (w.) In case of small data sets for any seismic region, it is not possible to have reliable region specific conversion relations and alternatively appropriate global regression relations for the required magnitude ranges and focal depths can be utilized. In this study, we collected global events magnitude data from ISC, NEIC and GCMT databases for the period 1976 to May, 2007. Data for m(b) magnitudes for 3,48,423 events for ISC and 2,38,525 events for NEIC, M (s) magnitudes for 81,974 events from ISC and 16,019 events for NEIC along with 27,229 M (w) events data from GCMT has been considered. An epicentral plot for M (w) events considered in this study is also shown. M (s) determinations by ISC and NEIC, have been verified to be equivalent. Orthogonal Standard Regression (OSR) relations have been obtained between M (s) and M (w) for focal depths (h < 70 km) in the magnitude ranges 3.0 a parts per thousand currency sign M (s) a parts per thousand currency sign 6.1 and 6.2 a parts per thousand currency sign M (s) a parts per thousand currency sign 8.4, and for focal depths 70 km a parts per thousand currency sign h a parts per thousand currency sign 643 km in the magnitude range 3.3 a parts per thousand currency sign M (s) a parts per thousand currency sign 7.2. Standard and Inverted Standard Regression plots are also shown along with OSR to ascertain the validation of orthogonal regression for M (s) magnitudes. The OSR relations have smaller uncertainty compared to SR and ISR relations for M (s) conversions. ISR relations between m (b) and M (w) have been obtained for magnitude ranges 2.9 a parts per thousand currency sign m (b) a parts per thousand currency sign 6.5, for ISC events and 3.8 a parts per thousand currency sign m (b) a parts per thousand currency sign 6.5 for NEIC events. The regression relations derived in this study based on global data are useful empirical relations to develop homogenous earthquake catalogs in the absence of regional regression relations, as the events catalog for most seismic regions are heterogeneous in magnitude types.