← Back to Sessions

Uses and Misuses of the Frequency-magnitude Distribution of Earthquakes

Description: 

The frequency-magnitude distribution shows that, on average, the numbers of earthquakes (N) increases by a factor of 10 for each step of one unit of decreasing magnitude (M). Thus, b=1 in log N = a – bM (equation 1, with a a regional constant). Co-authors and I have shown that b varies as function of local stress levels, with low b correlating with high stress. Proofs of this are provided by the variation of b as a function of focal mechanisms, and by decrease with hypocentral depth. This enables mapping of asperities along active faults and locations of magma chambers of volcanoes. In contrast to these useful facts, many believe falsely that the return time (Tr) of large earthquakes can be estimated from local a and b-values, and base probabilistic seismic hazard estimates (PSHA) on this erroneous assumption, worldwide. This error is anchored in the guidelines by ICOLD (International Committee on Large Dam) of how to estimate seismic hazard. Several authors have pointed out that a relation between equation 1 and Tr has never been proven to exist, but has been disproven for about 200 active faults. Some of the clearest examples that the PSHA as advocated by ICOLD yields false result, comes from paleoseismology and crustal deformation rates along the San Andreas (SA) fault’s two M8 class earthquakes of 1857 and 1906. Along the 300 km segment of the SA that ruptured in 1857, the ICOLD recommendation of using PSHA requires that about 10,000 earthquakes of M ≥ 3.7 should have occurred, since recording of this size earthquake has been complete, but in fact only 7 have been recorded. Investigations of quality data have shown that the simple PSHA assumption advocated by ICOLD is incorrect in 99% of cases. This wrong assumption of PASHA should be replaced by deterministic seismic hazard estimates, based on the existence of active faults and their likely rupture lengths, which allows an estimate of their Mmax, and hence the regional expected maximum accelerations of probable future ruptures.

Session: Advances in Probabilistic Seismic Hazard Analysis and Applications

Type: Oral

Date: 4/19/2023

Presentation Time: 02:45 PM (local time)

Presenting Author: Max Wyss

Student Presenter: No

Invited Presentation: 

Authors