A Unified Approach to Joint Regional/Teleseismic Calibration and Event Location with a 3D Earth Model
Abstract
This project is investigating the problem of locating seismic events from combined data sets of regional and teleseismic arrival times, based on the use of a unified 3D model of the Earth's velocity structure to predict travel times for both types of data. Inherent to this problem is the joint tomographic calibration of the Earth model with both regional and teleseismic ground-truth data. The goal of the unified approach is to remove the inconsistencies that result when travel-time predictions are performed with a mixture of separately calibrated regional and global models and empirical corrections, which can lead to degradation in event location accuracy. Additionally, the unified approach provides a rigorous framework for choosing optimal relative weighting of different types of data used in locating an event. We are addressing a number of practical difficulties that arise in the pursuit of a unified location/calibration capability. One is to develop fast and accurate raytracing techniques for modeling different types of seismic arrivals. A second is the computational challenge of performing joint tomographic calibration with very large numbers of data and model parameters. We are considering various strategies for reducing the problem size, such as averaging data from proximate events and stations, and for breaking the calibration problem into equivalent sub-problems by region or depth range. The calculation of model uncertainty, and its translation to travel-time prediction uncertainty provides optimal data weighting for locating events, but adds greatly to the computational challenge. A significant computational savings, in both calibration and location, is achieved by employing a linearized approximation to travel-time calculations, obtained by integrating the slowness function of a 3D model along ray paths computed in a 1D reference model. While this approximation is not adequate for regional travel times, we have performed numerical experiments that indicate
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 2010
- Accession Number
- ADA568700
Entities
People
- Delaine T. Reiter
- Nathan A. Simmons
- Stephen C. Myers
- William L. Rodi
Organizations
- Massachusetts Institute of Technology