Transition Zone Wave Propagation: Characterizing Travel-Time and Amplitude Information; Final rept. 23 Dec 2005-23 Dec 2007

摘要

We characterize transition-zone seismic wave propagation by mapping and calibrating the travel-time and amplitude behavior of P waves traveling through the transition zone at epicentral distances from 13 to 30 degrees and modeling the triplications resulting from the 410- and 660-km discontinuities. We have built an online database of waveforms from the IRIS FARM archive from 1990 to 2005 and process the data in order to compute source and station amplitude terms to correct for different magnitude sources and near-receiver site effects as well as errors in the instrument response functions. We use records from the full teleseismic P distance interval to estimate source-time- function envelopes and deconvolve them from the traces, discarding data from sources that last longer than 60 s and signals with low signal-to-noise ratios. The deconvolved traces are stacked into bins according to distance, providing an initial distance (or Earth-structure) wavefield term. Through several iterations we converge upon solutions for the event, station, and structure wavefield terms. This deconvolution technique is necessary to combine data from many different sources. We then compute both global and regional Earth- structure terms to obtain the average time-versus-distance amplitude of the wavefield, focusing on the 13 to 30 degree interval that is most sensitive to transition-zone structure. We model our data stacks using WKBJ synthetic seismograms and a niching genetic algorithm to explore the model space of different transition-zone velocity structures. We compare these results with long-wavelength models of 410- and 660-km discontinuity topography obtained from SS precursors and more detailed images beneath individual seismic stations derived from receiver functions.