T52B-01
Observations on Rupture Behaviour of Fluid Induced Events at the Basel EGS Based on Empirical Green's Function Analysis

Friday, 18 December 2015: 10:20
302 (Moscone South)
Jonas Folesky1, Joern Kummerow1 and Serge Alexander Shapiro2, (1)Freie Universitaet Berlin, Berlin, Germany, (2)Freie Universität Berlin, Geophysics, Berlin, Germany
Abstract:
The Emprirical Green's Function (EGF) method uses pairs of events of high wave form similarity and adjacent hypocenters to decompose the influences of source time function, ray path, instrument site, and instrument response. The seismogram of the smaller event is considered as the Green's Function which then can be deconvolved from the other seismogram. The result provides a reconstructed relative source time function (RSTF) of the larger event of that event pair. The comparison of the RSTFs at different stations of the observation systems produces information on the rupture process of the larger event based on the observation of the directivity effect and on changing RSTFs complexities.
The Basel EGS dataset of 2006-2007 consists of about 2800 localized events of magnitudes between 0.0<ML<3.5 with event pairs of adequate magnitude difference for EGF analysis. The data has sufficient quality to analyse events with magnitudes down to ML=0.5 for an apparent directivity effect although the approximate rupture duration for those events is of only a few milliseconds. The dataset shows a number of multiplets and repeating earthquakes known from earlier studies. The larger events seem to appear close to the rim of the microseismic cloud. We are interested in their rupture behaviour. Using the EGF method we compute rupture orientations for about 190 event pairs and relate them to the event location, the known fault system, and stress regime. For the majority of events we observe a similar rupture direction which seems to correlate with the over all shape of the microseismic cloud. The large events, however, point back to the injection source.
Additionally the rupture direction fitting yields estimates for projections of the rupture velocity on the horizontal plane. They seem to vary between the multiplets in the reservoir from 0.3 to 0.7 times the S-wave velocity.
To our knowledge source characterization by EGF analysis has not yet been introduced to microseismic reservoirs with comparable data quality to the one from the Basel experiment. Our results show that EGF analysis can provide valuable additional insights on the distribution of rupture properties within the reservoir. The results suggest a preferred rupture direction of large induced events that occur at the reservoir rim back towards the injection source.