DO RICHTER MAGNITUDES REVEAL SEISMIC SITE RESPONSE?

We have undertaken an assessment of earthquake magnitudes in the Pacific Northwest US. Historically, the Pacific Northwest Seismograph Network (PNSN) has utilized the coda duration magnitude (Mc) scale. The PNSN Mc scale was derived by Crosson in 1972 to match Local (ML or Richter) Magnitudes determined from amplitude measurements made on a Wood-Anderson seismograph located in Seattle. The Mc method of determining earthquake magnitude has the advantage that it can be used even with seismic records whose amplitudes are off-scale, as was common with early seismographs having low dynamic-range.

The introduction of wide dynamic-range digital seismographs and accelerometers has made possible the direct calculation of ML from the amplitudes of seismic waves. We have computed the ML of 76 earthquakes that occurred from 2000-2003 with 2.5<M<6.8 using observations from a total of 2812 seismograms. We utilized a method of Joint Magnitude Determination in which earthquake magnitudes and station corrections are determined simultaneously for all earthquakes. The computed magnitude corrections at the stations range between -0.4 to +0.4 magnitude-units.

On average, the magnitude corrections are correlated with, and can be predicted from, the NEHRP soil classification assigned to each site. Computed magnitudes tend to be higher at sites where the estimated surface shear wave velocity is low. The variations in computed magnitude do not seem to depend on the presence of deep-basin structures in the subsurface. On the other hand, such structures have been shown to affect the amplitudes of low-frequency seismic waves. This apparent discrepancy is probably due to the fact that Richter magnitudes are intrinsically sensitive to higher frequency ground motions. We conclude that variations in Richter magnitude can be used to detect shallow subsurface site conditions likely to affect earthquake shaking.