Seismic Attenuation Tomography of the Rupture Zone of the 2010 Maule, Chile, Earthquake

Abstract:

We used measurements of differential S to P seismic attenuation in the rupture zone of the 2010 M_w 8.8 Maule, Chile earthquake (33°S-38°S) to characterize the seismic attenuation structure of the South American crust and upper mantle wedge. We used data obtained from the IRIS CHAMP rapid-response temporary seismic network, filtered between 0.7-20 Hz. For events with large signal to noise ratios, we visually identified the P and S arrivals on the seismograms and used an evolving time window to determine 400 individual Q_s and t* values and their uncertainties using a spectral ratio method. Using a phase pair method allows us to neglect the source-time function and instrument response of each P-S phase pair. Assuming a constant Q_p/Q_s ratio for a given P-S phase pair, we evaluated the 400 spectral ratios and discarded portions of the evolving time window that incorporate multipathed phases. We recalculated the Q_s and standard deviation of the retained window and excluded measurements with standard deviations larger than half of the Q_s value. We also excluded measurements that span frequency windows longer than 10 Hz as they contain noise that contaminates Q_s measurements. We examined ~200 local events yielding a total of 1,076 path-integrated Q­_s measurements. Q_s values are low (100-400) for the majority of ray paths evaluated, however we observe a spatial distribution of low path-integrated Q_s values (100-300) in the northeastern portion of the rupture zone and higher values (300-600) in the southwest. We divided the rupture zone into cubes and implemented a bounded linear inequality least squares inversion (0<Q_s<1200) to solve for Q_s in individual model regions. We used a constant starting model of Q_s=450 and solved for perturbations to the model using our observed δt* values and a radial-earth velocity model (IASPI91). Our ray coverage provides good resolution extending to depths of ~75 km in the mantle wedge.