Next-generation Boundary Element for Earthquake Science

Thomas Thompson, Harvard University

Photo of Thomas Thompson

Boundary element methods (BEMs) are immensely useful tools for modeling fault behavior. However, existing BEM tools are incapable of including three-dimensional fault geometry, surface topography and material variations. Three-dimensional BEMs have been held back by the difficulty of calculating the Green's function integrals necessary for the method. I have developed a general-purpose solution to this challenge. I use a range of numerical tools, including adaptive quadrature, numerical limit-taking procedures and interpolation tables, to create an efficient algorithm for calculating almost any boundary element integral, regardless of geometry or the particular Green's function.

Many earthquakes occur in regions with mountains or nearby sedimentary basins. Both topographic variations and material property variations will have significant influence on both observed coseismic and interseismic displacements. I study these influences using both idealized models and the 2008 Mw 7.9 Wenchuan earthquake. For the Wenchuan earthquake, I use real geometries for the fault, surface topography and basin/bedrock interfaces. The influence of such added geometric effects is surprising and non-intuitive.

Abstract Author(s): T. Ben Thompson, Brendan J. Meade