The local magnitude distance correction term, log A0(Δ), has been estimated for paths within the Great Basin by converting seismograms of earthquakes in the vicinity of Mammoth Lakes, California, into equivalent Wood-Anderson seismograms and measuring the decay of peak amplitude with distance. The shape of the log A0(Δ) curve which results suggests more attenuation of energy in the Wood-Anderson passband than in southern California. For Great Basin paths we find, for events with ML ≦ 5.5, that
where Δ is epicentral distance, is hypocentral distance, and h is focal depth. Our data also suggest a magnitude bias in the correction term, in that for events above ML ≈ 5.5, peak amplitudes recorded at epicentral distances below 20 km decay less rapidly than those for smaller events recorded at the same sites. This observation is consistent with the hypothesis that there is a near-source saturation of ML due to the effects of a finite source size for large earthquakes. Using the revised magnitude scale and seismic moments, M0, estimated from spectral analysis, the data are well fit by the straight line
for 1 ≦ ML ≦ 6. M0 versus ML values from model calculations which assume constant stress drop predict curved moment-magnitude plots over this magnitude range. The fact that we observe a linear relationship suggests that the stress drop of these events is not constant but rather increases with ML, in particular for events above ML ≈ 5.5.