Assorted 2D Synthetics

Don V. Helmberger

Caltech
Email: helm@gps.caltech.edu

poster/oral: poster

Broadband record sections from arrays are compared with synthetic predictions generated from tomographic models. In general, such models require sharpening and enhancing to fit body waveforms. Models from the circum-Pacific fast belts can be modified by adding a temperature-dependent phase jump to explain the timing and strengths of the Scd phase occurring between S and ScS, and a global map of D" constructed. This model appears to explain the observed data from most fast (slabs) regions, although anisotropic features require more work. Beneath South Africa, the so-called African Plume appears to have the strongest slow anomalies. Delays of ScS relative to S of up to 10 sec at 90° are commonly observed. Such delays can be explained by long-period tomographic models with the addition of a slow D" structure (3 to 6% drops in shear velocities). However, these structures cannot explain the (SKS-S) differentials sampling the same region. To explain the (SKS-S) and (ScS-S) datasets simultaneously requires instead a large-scale ridge-like structure with a relatively uniform 3% reduction in shear velocity. The structure is about 800 km wide and extends at least 1000 km above D". It is orientated roughly NW-SE and leans towards the east at latitudes from 15° to about 30°. The walls of this structure appear to be sharp, in that, a fortuitous set of SKS phases traveling along this boundary display multipathing. To model such data, we need a sharp transition of less than 50 km, implying a chemical boundary. To model the above features, we employ a modification of the WKMJ method that is easily used in conjunction with the Cagniard-de Hoop method to generate 2D synthetics. As in tomography, we initiate the first stage by computing the paths reflecting off each layer of a 1D reference model (PREM), and noting the position of all ray segments. The velocity of each segment (layer crossing) is then adjusted to fit the tomographic model and a set of new paths generated to form a set of travel times (ti) and ray parameters (pi). These values can be used to generate a P-T curve and synthetic or used directly in a Cagniard-de Hoop code to handle the more difficult propagational problems, diffractions, etc. Synthetics for this procedure are compared with other methods with generally favorable results.


Go back to Schedule
Go back to Participants
Go back to Gilbert Fest home page