S33B-2758
Implementing a C++ Version of the Joint Seismic-Geodetic Algorithm for Finite-Fault Detection and Slip Inversion for Earthquake Early Warning

Wednesday, 16 December 2015
Poster Hall (Moscone South)
Deborah E. Smith1, Claude Felizardo2, Sarah E Minson1, Maren Boese3, John O Langbein1, Christian Guillemot1 and Jessica R Murray4, (1)U.S. Geological Survey, Earthquake Science Center, Menlo Park, CA, United States, (2)California Institute of Technology, Pasadena, CA, United States, (3)ETH Zurich, Zurich, Switzerland, (4)US Geological Survey, Earthquake Science Center, Menlo Park, CA, United States
Abstract:
The earthquake early warning (EEW) systems in California and elsewhere can greatly benefit from algorithms that generate estimates of finite-fault parameters. These estimates could significantly improve real-time shaking calculations and yield important information for immediate disaster response. Minson et al. (2015) determined that combining FinDer’s seismic-based algorithm (Böse et al., 2012) with BEFORES’ geodetic-based algorithm (Minson et al., 2014) yields a more robust and informative joint solution than using either algorithm alone. FinDer examines the distribution of peak ground accelerations from seismic stations and determines the best finite-fault extent and strike from template matching. BEFORES employs a Bayesian framework to search for the best slip inversion over all possible fault geometries in terms of strike and dip. Using FinDer and BEFORES together generates estimates of finite-fault extent, strike, dip, preferred slip, and magnitude.

To yield the quickest, most flexible, and open-source version of the joint algorithm, we translated BEFORES and FinDer from Matlab into C++. We are now developing a C++ Application Protocol Interface for these two algorithms to be connected to the seismic and geodetic data flowing from the EEW system. The interface that is being developed will also enable communication between the two algorithms to generate the joint solution of finite-fault parameters. Once this interface is developed and implemented, the next step will be to run test seismic and geodetic data through the system via the Earthworm module, Tank Player. This will allow us to examine algorithm performance on simulated data and past real events.