Harvard Seismology: Event Location

 This project is currently funded by the Defense Threat Reduction Agency (DTRA) with the goal of achieving high precision regional and teleseismic event locations using 3D models of mantle structure.  Primary motivation for the research is successful monitoring of the proposed Comprehensive Test Ban Treaty (CTBT).  The CTBT has the stated goal of locating events to a precision greater than 1000 km^2.  This goal is frequently not obtained using conventional 1-D seismic velocity models and location techniques. Obtaining better quality teleseismic event locations is also important for studies which utilize the global distribution of seismicity, particularly in terms of depth.

Previously, we  have demonstrated that absolute locations determined using teleseismic arrival times can be improved through incorporation of three dimensional (3-D) models of the P-wave velocity structure of the mantle, even when these models are parameterized by a low-order expansion of the velocity heterogeneity in terms of spherical harmonics.  In addition, a global  study of P_n arrivals reported to the ISC indicates that azimuthal anisotropy is an important feature in upper mantle P-velocities.  This has important implications for the accuracy at which events can be located at regional distances.

Currently our research is focused in five main areas:

1. 1. Development of new techniques for improving current Earth models, such as multiple resolution algorithms and incorporation of regional seismic phases.
2. 2. Construction of higher resolution three-dimensional Earth models, with emphasis on P-wave structure.
3. 3. Evaulation of the new models using independent data.
4. 4. Development of additional station and/or path corrections, in conjunction with the 3D models, for use in location algorithms.
5. 5. Compilation of a high quality dataset of well located master events, including both explosions and earthquakes, for use in assessing improvements in location capability.

Michael Antolik, Göran Ekström, Adam Dziewonski, Yu J. Gu, Lapo Boschi, Jianfeng Pan,

Dr. Ivan Kitov, Institute for the Dynamics of the Geosphere, Moscow

• Analysis of improvement in teleseismic event locations obtainable by incorporating low-order mantle P-wave heterogeneity (Smith and Ekström, 1996)
• Variation of P_n velocities under contintents (Smith and Ekström, 1998)
• Construction of a high-resolution model of mantle P wave velocity (Boschi and Dziewonski, 1999).

• Tests of location accuracy obtained using current models of P-velocity These experiments are designed to test the location accuracy obtained using current mantle P-wave velocity models of varying resolution. The data used are teleseismic phases from reference or "ground truth" events (earthquakes and explosions) whose locations are known exactly or extremely well from local and regionazl network data.
• A new global P-velocity model parameterized in cubic splines This model is based on a new dataset of summary P wave travel times and surface wave phase measurements. The model is designed to recover long wavelength structure well, especially in the upper mantle to aid in teleseismic event location.
• Improving locations of oceanic earthquakes We are using a constrained inversion technique to relocate earthquakes on remote plate boundaries. This should help compile reference events for testing 3-D models in areas with sparse station coverage.

• G.P. Smith and G. Ekström, Improving teleseismic event locations using a three-dimensional Earth model Bull. Seism. Soc. Am., 86, 788-796, 1996
• G.P. Smith and G. Ekström. Interpretation of earthquake epicenter and CMT centroid locations, in terms of rupture length and direction, Phys. Earth and Plan. Inter., 102, 123-132, 1997
• G.P. Smith and G. Ekström, A global study of P_n anisotropy beneath continents, Journ. Geophys. Res., , in press, 1999)

Title Illustration. P-wave travel time residuals in 5^o x 5^o caps predicted by the mantle model S&P12/WM13 for an event in Kermadec. Red and blue triangles denote positive and negative residuals, respectively.