Calibration of magnitude scales for earthquakes of the Mediterranean

Abstract

In order to provide the tools for uniform size determination for Mediterranean earthquakes over the last 50-year period of instrumental seismology, we have regressed the magnitude determinations for 220 earthquakes of the European-Mediterranean region over the 1977–1991 period, reported by three international centres, 11 national and regional networks and 101 individual stations and observatories, using seismic moments from the Harvard CMTs. We calibrate M(M0) regression curves for the magnitude scales commonly used for Mediterranean earthquakes (ML, MWA, mb, MS, MLH, MLV, MD, M); we also calibrate static corrections or specific regressions for individual observatories and we verify the reliability of the reports of different organizations and observatories. Our analysis shows that the teleseismic magnitudes (mb, MS) computed by international centers (ISC, NEIC) provide good measures of earthquake size, with low standard deviations (0.17–0.23), allowing one to regress stable regional calibrations with respect to the seismic moment and to correct systematic biases such as the hypocentral depth for MS and the radiation pattern for mb; while mb is commonly reputed to be an inadequate measure of earthquake size, we find that the ISC mb is still today the most precise measure to use to regress MW and M0 for earthquakes of the European-Mediterranean region; few individual observatories report teleseismic magnitudes requiring specific dynamic calibrations (BJI, MOS). Regional surface-wave magnitudes (MLV, MLH) reported in Eastern Europe generally provide reliable measures of earthquake size, with standard deviations often in the 0.25–0.35 range; the introduction of a small (± 0.1–0.2) static station correction is sometimes required. While the Richter magnitude ML is the measure of earthquake size most commonly reported in the press whenever an earthquake strikes, we find that ML has not been computed in the European-Mediterranean in the last 15 years; the reported local magnitudes MWA and ML do not conform to the Richter formula and are of poor quality and little use, with few exceptions requiring ad hoc calibrations similar to the MS regression (EMSC, ATH). The duration magnitude MD used by most seismic networks confirms that its use requires accurate station calibrations and should be restricted only to events with low seismic moments.