A Comparison of Long-Period SKS Datasets And What They Reveal About 1D Outer Core Structure

Seismology is the most direct tool for documenting the presences or absence of outer core stratification. The outermost core is most effectively sampled by SKS, S2KS, S3KS, S4KS, etc.) which have bottoming depths at the top of the outermost core. In order to incorporate modern data sets (e.g., USArray, Europe, China, etc), we need to sift through massive amounts of seismic data to identify the smaller portion of quality signals in a time-efficient manner. We evaluate the application of a cluster analysis technique (Houser et al., 2008) toward identifying and evaluating the SKS phases that traverse the outer core. Cluster analysis is a semi-automated method for interrogating large datasets by processing all the data for an earthquake while allowing the user to graphically interact with the data to remove low quality records. The Houser et al. (2008) cluster analysis method has already been applied to diffracted S waves (Manners et al., 2004), and here we will expand the cluster analysis to the radial component core phases. These newly measured SKS arrival times will be compared with SKS arrival time measurements used in previous mantle tomographic models, namely, S20RTS (Ritsema and van Heist, 2002) using a purely automated method and TXBW (Grand, 2002) using a purely manual method. We find that the arrival times collected by the three methods (automated, clustered, and manual) during overlapping time frames are in agreement within the measurement error bars. Therefore, the SKS data from these studies can be combined to constrain the radial structure of the outermost core. Thus, cluster analysis is an ideal tool for developing a large compilation of SKS arrival times from modern global seismic data, while simultaneously providing a measure of data quality.