Given the importance of shaking prediction for earthquake early warning (EEW) and, rapid situational awareness (including damage assessments and the evaluation of future seismic hazards such as aftershocks), it is crucial to quickly determine the spatial extent, in particular the length and orientation of fault ruptures during significant earthquakes. This controls both the amplitudes and distribution of ground motion (Böse et al., 2018). The FinDer algorithm (Böse et al., 2012, 2015, 2018, 2023, 2024) was developed to address this need. FinDer uses template matching to rapidly compute finite-source models that provide information on the spatial extent of high-frequency seismic ground motion (especially peak ground acceleration, PGA) in the proximity of the developing fault rupture. FinDer can work with real-time seismic data and capture the fault rupture evolution as it occurs.
The development of FinDer began with a focus on large earthquakes (Böse et al., 2012). However, the algorithm also showed promising performance on medium-sized events through a real-time prototype implementation in California (Böse et al., 2015). This success led to various updates to the original algorithm, such as the addition of the ability to detect and characterize earthquakes with magnitudes from 2.5 to 9.0 (FinDer version 2; Böse et al., 2018), and more recently, support for fault-specific template sets (FinDer version 3; Böse et al., 2023).
FinDer applies binary template matching to select a template with spatially distributed ground motion amplitudes modeled for a given source geometry (e.g., for a line source or a more complex 2D rupture) that best matches the currently observed ground motion pattern in the epicentral region. The FinDer magnitude, Mfd, is estimated from empirical rupture length-magnitude scaling relationships. The latest version of FinDer (version 3) supports two types of template sets:
Fault-specific templates do not replace generic templates, but are used in a multi-template environment to improve the performance of FinDer, especially for large-scale earthquakes in areas with mapped active faults. Ongoing developments in FinDer address the incorporation of low-frequency displacement amplitudes to estimate slip and alternate magnitudes (Böse et al., 2021a).
FinDer is or has been used in real timereal-time in various seismic networks around the world, e.g. along the US West Coast (ShakeAlert; e.g. Böse et al., 2023), Central America (Massin et al., 2022; Porras et al., 2021), Switzerland (Massin et al., 2021), New Zealand (Andrews et al., 2023), Italy, Chile, and Taiwan. FinDer has been tested offline in Japan and China. While originally developed for EEW purposes, FinDer has added value in enhancing ShakeMaps and rapid damage estimates after large earthquakes, as demonstrated for the 2008 M 7.8 Wenchuan (China; Li et al., 2020) and 2023 M 7.8 Pazarcik (Türkiye) earthquakes (Böse et al., 2024). As many countries may not have the resources to build and maintain dense seismic networks required to determine rupture geometry, Böse et al. (2021b) and Bossu et al. (2024) demonstrated the successful operation of FinDer using intensity felt reports sourced from the population.
The implementation of FinDer is carried out in close collaboration with the US Geological Survey (USGS), the California Institute of Technology (Caltech), and GNS Science (NZ). Since 2021 FinDer integrates has integrated with the ETHZ-SED SeisComP EEW (ESE) system. The code for the scfinder wrapper module is free and open source from SeisComP v4.6.0, but the executable is not included in the SeisComP distribution. scfinder is distributed under the GNU Affero General Public License (Free Software Foundation, version 3 or later) .The FinDer and scfinder libraries and executables are distributed separately on request.