The role of source mechanism geometry in biasing estimates of earthquake slowness

Slow earthquakes, characterized initially as ones in which magnitude increases with period, are often observed on oceanic transform faults. Previous work found that when compared to oceanic ridge and intraplate earthquakes, oceanic transform earthquakes often have large MS relative to mb and large MW relative to MS (Stein and Pelayo, 1991). This pattern suggests that seismic wave energy is relatively greater at longer periods for these events. The cause of this phenomenon can be attributed to differences in rupture velocity, stress drop, or the physical properties of the source region. An alternative explanation could be a mechanism bias resulting from the way energy radiates from strike-slip earthquakes on a vertically dipping fault. Energy arriving at teleseismic distances (> 30 degrees) will have a take-off angle near the nodal plane, which results in smaller amplitudes. Calculating body- and surface-wave magnitudes from synthetic seismograms show that most oceanic transform earthquakes are classified as slow due to mechanism bias. In addition to looking at magnitude ratios, we calculate the energy-to-moment ratio (Newman and Okal, 1998). This ratio is analogous to mb :MS as it compares high- and low-frequency measurements (Okal, 1998) and thus can distinguish slow events. This measurement does not find any slow events in our catalog of oceanic transform earthquakes. A dataset of continental events is also examined with both of the methods above to look for differences due to source properties. These results are similar to those for oceanic events, so source properties are not a cause for ''slow'' earthquakes. Last, source time functions are modeled and compared to seismic moment, M0, in order to estimate stress drop. The results show similar stress drops for oceanic ridge, transform and intraplate events. However, many of the events have a poor fit.