ITR: Development of an Integrated Tool for Modeling Quasi-static Deformation in the Solid Earth

Rensselaer Polytechnic Institute, Troy NY

Investigators

Abstract

This project is providing a complete modeling environment for solving quasi-static problems in geophysics. The goal is to provide a tool that is easy to use, open-source, and able to solve problems of all scales, including large-scale problems that are not presently solvable. With this tool, users are able to address problems on time scales ranging from days (e.g., postseismic relaxation following a large earthquake, magma chamber expansion/contraction beneath an active volcano), to years (e.g., models of interseismic stress/strain accumulation to evaluate seismic hazard), to hundreds or thousands of years (e.g., multi-earthquake simulations of an active seismic zone), to millions of years (e.g., evolution of tectonic features). Ease of use is aided by incorporating the code into the Pyre framework being developed at Caltech. By using this framework, users automatically gain access to other tools already included in Pyre, including the ability to couple different codes. Use of the Pyre framework also greatly reduces the complexity of creating a model, exploring parameter space, and visualizing the results. Comprehensive documentation and a suite of test problems are also being provided as a starting point for new users, allowing them to rapidly develop their own models. The core of this project is a version of the finite element code TECTON, which has been specifically designed to solve tectonic problems. In addition to incorporating the code into the Pyre framework, the code is also being parallelized using the PETSc parallel libraries. The parallel implementation will allow the code to be used on a wide range of computer architectures, including serial (single processor) machines, shared memory (multiprocessor) machines, networks of PC's or workstations, or specially-designed workstation clusters. Thus, users will be able to solve problems of all scales, depending on the available computational resources