Understanding and resolving the systematic differences observed for teleseismic body-wave magnitude measurements.

摘要

In the first of three projects, the procedure used by the Prototype International Data Centre (PIDC) to assign body-wave magnitude (m b) is compared to a traditional magnitude measurement, which I base on simulated Worldwide Standard Seismographic Network short-period (WWSSN SP) signals. This reference magnitude is necessary because many different magnitude scales are in use and because individual agencies have changed their procedures since seismic magnitudes were introduced in the 1930s. Based on an analysis of 10 large earthquakes and 13 nuclear explosions, the PIDC procedure is found to be adequate for explosions, but not for earthquakes.; In Chapter 2, I investigate why magnitudes at the U.S. Geological Survey (USGS) and PIDC are inconsistent. The mb procedures used by these two agencies are reproduced and then compared to a more traditional measurement based on simulated WWSSN SP signals. The dataset consists of 2009 earthquakes from 1996 to 1999 listed in the PDE bulletin having m b between 5.0 and 5.5. Magnitudes based on USGS procedure are similar to those based on WWSSN SP instruments; however, magnitudes based on PIDC procedure are significantly lower than those based on simulated WWSSN SP records. Differences between the three mb scales arise from four factors: response function, length of time window, and corrections for event depth and epicentral distance.; In Chapter 3, I test the assumption that source scaling, which includes a relationship between seismic moment (M0) and corner frequency (fc), is constant across a given region. Two clusters of earthquakes are analyzed, one in Tibet and another in Tien Shan. Source scaling for the two clusters is similar and stress drop in the broad region containing both clusters is approximately constant. The relationship between mb and f c, on the other hand, indicates that for a fixed value of m b, earthquakes in Tibet are characterized by lower fc than those in Tien Shan. The scaling differences are attributed to a local low Q that exists underneath Tibet and reduces teleseismic short-period amplitudes and therefore m b as well.