Suitability of rapid energy magnitude determinations for emergency response purposes

Abstract

SUMMARY It is common practice in the seismological community to use, especially for large earthquakes, the moment magnitude Mw as a unique magnitude parameter to evaluate the earthquake's damage potential. However, as a static measure of earthquake size, Mw does not provide direct information about the released seismic wave energy and its high frequency content, which is the more interesting information both for engineering purposes and for a rapid assessment of the earthquake's shaking potential. Therefore, we recommend to provide to disaster management organizations besides Mw also sufficiently accurate energy magnitude determinations as soon as possible after large earthquakes. We developed and extensively tested a rapid method for calculating the energy magnitude Me within about 10–15 min after an earthquake's occurrence. The method is based on pre-calculated spectral amplitude decay functions obtained from numerical simulations of Green's functions. After empirical validation, the procedure has been applied offline to a large data set of 767 shallow earthquakes that have been grouped according to their type of mechanism (strike-slip, normal faulting, thrust faulting, etc.). The suitability of the proposed approach is discussed by comparing our rapid Me estimates with Mw published by GCMT as well as with Mw and Me reported by the USGS. Mw is on average slightly larger than our Me for all types of mechanisms. No clear dependence on source mechanism is observed for our Me estimates. In contrast, Me from the USGS is generally larger than Mw for strike-slip earthquakes and generally smaller for the other source types. For ∼67 per cent of the event data set our Me differs ≤±0.3 magnitude units (m.u.) from the respective Me values published by the USGS. However, larger discrepancies (up to 0.8 m.u.) may occur for strike-slip events. A reason of that may be the overcorrection of the energy flux applied by the USGS for this type of earthquakes. We follow the original definition of magnitude scales, which does not apply a priori mechanism corrections to measured amplitudes, also since reliable fault-plane solutions are hardly available within 10–15 min after the earthquake origin time. Notable is that our uncorrected Me data show a better linear correlation and less scatter with respect to Mw than Me of the USGS. Finally, by analysing the recordings of representative recent pairs of strong and great earthquakes, we emphasize the importance of combining Mw and Me in the rapid characterization of the seismic source. They are related to different aspects of the source and may differ occasionally even more than 1 m.u. This highlights the usefulness and importance of providing these two magnitude estimates together for a better assessment of an earthquake's shaking potential and/or tsunamigenic potential.