Using centroid time-delays to characterize source durations and identify earthquakes with unique characteristics

Rapidly Estimated Seismic Source Parameters for the 16 September 2015 Illapel, Chile M w 8.3 Earthquake

“…For source depths of 17.5-25.5 km this analysis yielded shallow-dipping (13.7°-17.4°) thrust mechanisms with moments of 3.5-3.4 9 10 21 Nm, and could be completed within five to ten minutes if operated in real time, as it requires only 200-300 s of recorded signal after the origin time. Centroid time shifts, t c , of 42-41 s were found, which are only moderately longer than the value of 36 s predicted based on the average relationship found by Duputel et al (2013) of t c = 2.59 9 10 -6 M 0 1/3 (for M 0 in Nm). This provides a rapid determination that the source was not a tsunami earthquake, as such events have significantly larger t c shifts relative to the average (DUPUTEL et al 2013), even while the seismic moment, mechanism, and location of the 2015 event do suggest significant tsunamigenesis.…”

“…Centroid time shifts, t c , of 42-41 s were found, which are only moderately longer than the value of 36 s predicted based on the average relationship found by Duputel et al (2013) of t c = 2.59 9 10 -6 M 0 1/3 (for M 0 in Nm). This provides a rapid determination that the source was not a tsunami earthquake, as such events have significantly larger t c shifts relative to the average (DUPUTEL et al 2013), even while the seismic moment, mechanism, and location of the 2015 event do suggest significant tsunamigenesis. An operational application of regional W-phase inversion was performed by the Chilean National Seismological Center of the University of Chile, and provided a solution within 5 min using data from 22 stations within 12°of the epicenter, with M W 8.1.…”

Evidence for a difference in rupture initiation between small and large earthquakes

“…By averaging P d among many stations, distributed over azimuth and distance, the resulting function approximates the MRF. Assuming a triangular shape for the MRF, the plateau of P d versus time curves is reached at the peak of the MRF, and the corresponding time is a proxy for the source moment and duration 22 . Our observations indicate that the initial shape of the P d versus time curves typically follows an exponential trend.…”

Evidence for non-self-similarity of microearthquakes recorded at a Taiwan borehole seismometer array

“…] = c which assumes that the corner frequency of the P wave, f cP , is the same as that of the S wave, f cS , but the observed ratio, f cP / f cS is commonly Duputel et al (2013). The right vertical axis gives the source dimension.…”

“…Allmann & Shearer 2009;Duputel et al 2013), it has also been a subject of debate, especially for small earthquakes (Abercrombie 1995;Ide et al 2004;Stork & Ito 2004;Oye et al 2005;Venkataraman et al 2006;Yamada et al 2007). Conventionally, the source duration of an event is determined from the spectral corner frequency by considering a frequency-dependent (Stork & Ito 2004;Oye et al 2005) or frequency-independent Q-model to remove path effects.…”