[S01-1] Open session I

[S01-1-02] Automated seismic event location combining waveform stacking and relative location techniques

^○Francesco Grigoli¹, Simone Cesca², Frederic Massin¹, Anne Obermann¹, Wilfried Strauch³, John Clinton¹, Stefan Wiemer¹ (1.ETH-Zurich, Swiss Seismological Service (SED), Switzerland, 2.GFZ-Potsdam, German Research Centre for Geosciences, Germany, 3.Instituto Nicaraguense de Estudios Territoriales (INETER), Nicaragua)

Microseismic monitoring became a common operation in many applications (e. g. induced seismicity). The analysis of microseismic data is challenging, because of the large number of recorded events often characterized by low signal-to-noise ratio. In the last years, to improve the performance of the automated location procedures, various waveform based methods for microseismicity characterization have been proposed. These picking-free methods exploits the coherence of the waveforms recorded at different stations and appear to be more noise robust than the traditional methods (based on automatic phase picking). However, like any other absolute location method, the accuracy strongly depends on the knowledge of the velocity model. The use of simplified 1D velocity models in areas characterized by complex 3D velocity structures or by pronounced topography may strongly affect the locations accuracy. The poor knowledge of the velocity model is, in fact, the largest source of error in the seismic event location process. In this work we present a location method which combines features of relative location techniques with the waveform based location methods. This approach inherits all the advantages of the fullwaveform location methods without its main drawback (i.e. effect of poorly known velocity model). In fact, this method is less dependent on the knowledge of the velocity model, improving the location accuracy: 1) it accounts for phase delays due to local site effects, e.g. surface topography or variable sediment thickness 2) theoretical velocity model are only used to estimate travel time within the source volume, and not along the entire source-sensor path. In order to validate this approach we tested it with different synthetic and real datasets, including induced seismicity related to geothermal energy exploitation in St. Gallen (Switzerland) and seismic swarms of volcanic origin in Nicaragua.