Seismic moment assessment of earthquakes in stable continental regions—II. Historical seismicity

SUMMARY This Part I study, in conjunction with Part 11, develops a method to determine, within specified uncertainty bounds, the seismic moment, and thus moment magnitude, of all earthquakes of stable continental regions (SCR) for which instrumental or intensity data exist. Its basis is polynomial regression analysis using a database of SCR earthquakes with direct seismic moment determination. The independent variables include modern teleseismic magnitudes and regional magnitudes (Part I), and isoseismal areas or number of recording stations (Part 11). Part I11 is an application of the methodology of Parts I and I1 to several major historical earthquakes. All data used in the regressions are assigned individual uncertainties estimated from the literature or from experience; formal confidence limits (68 per cent or 95 per cent) on both the regression formulas and the predicted seismic moment values are then possible via error propagation analysis. The most complete development is for the teleseismic magnitudes M, and mb. For both, the final regression for log(M,) is a quadratic formula that closely emulates the relationship between amplitude magnitudes and log(M,) expected from dislocation theory and source-scaling arguments. Regressions are also derived for the regional magnitudes mLg and M,, because there are many SCR events, mainly pre- 1964, that have no teleseismic magnitudes. Prediction uncertainties from teleseismic magnitudes in moment magnitude units are in the AO.18-0.28 range, and from regional magnitudes in the kO.23-0.38 range over a wide magnitude band. Finally, the methodology developed here is generic, even though the database is specific. Application to plate-boundary, oceanic intraplate, or active continental intraplate regions should be straightforward.