Rupture geometry from high-precision relative hypocentre locations of microearthquake clusters

Abstract

S U M M A R Y In 1987, two microearthquake sequences, with a duration of about a week each and consisting of 37 and 46 events respectively, occurred in the upper crust below the Jura Mountains of northern Switzerland. The seismograms within each sequence exhibit a high degree of similarity, indicating tight clustering of hypocentres and similar focal mechanisms. Using a cross-correlation technique applied to the seismograms in the time domain, together with a new least-squares adjustment procedure, it was possible to determine relative hypocentre locations of the earthquakes in each cluster with a precision of a few tens of metres. The results show that hypocentres of events with the same focal mechanism lie on a plane which coincides exactly with one of. the nodal planes of the fault-plane solution, and that consequently swarm-like sequences of similar earthquakes are due to repeated slip on the same fault. In one case, slip occurred as right-lateral motion on a steeply dipping plane striking in WNW-ESE direction, whereas the other case corresponds to left-lateral slip on two almost vertical planes striking roughly N-S. In 1988, an additional earthquake triplet occurred on a WNW-ESE oriented normal fault nearby. All three mechanisms are consistent with a general NNW-SSE oriented direction of maximum crustal shortening and a corresponding WSW-ENE oriented extension. Repeated slip on the same fault is indicative of a large degree of heterogeneity and of short-term temporal variability of both frictional resistance and stress distribution on the fault. An explanation for the occurrence of such swarm-like seismic activity in terms of barriers or asperities would require that shear stress on the unbroken patches increase from one event to the next. Since, however, the overall shear stress would be expected to decrease as a consequence of the stress released in each event, a more plausible mechanism involves pore-pressure fluctuations, caused by fluids under suprahydrostatic pressures migrating upward through pre-existing zones of weakness in the crust.

Extensional neotectonics around the bend of the Western/Central Alps: an overview

“…It is based on cross-correlation (time or frequency) of similar seismograms (similar body waves due to similar ray-paths). This method gives the relative location of the event of a given crisis with respect to the master event, with an overall precision in the order of tens of metres (e.g., Fréchet 1985;Deichmann and Garcia-Fernandez 1992;Maurer and Deichmann 1995). In most Alpine cases, the alignment obtained using relative relocations is coherent with one of the nodal planes of the main shock's focal mechanism.…”

“…These types of relocations in the Western Alps have been compiled on the map Fig. 5 a (red symbols) after Deichmann and Garcia-Fernandez (1992), Augliera et al (1995), Maurer and Deichmann (1995), Kastrup et al (2004), and Delacou et al (2005b).…”

Spatial characterization of Agron, southern Spain, 1988–1989 seismic series

“…As demonstrated in a pletora of studies, the importance of recognizing the occurrence of these cluster of events principally relies on the fact that the similarities among waveforms allow for a highly consistent determination of phase onsets, which are in turn used for precise relative hypocenter determination (e.g. Ito, 1985;Fremont and Malone, 1987;Deichmann and Garcia-Fernandez, 1992;Got et al, 1994;Augliera et al, 1995;Cattaneo et al, 1997;Shearer, 1997;Shearer, 1998;Phillips, 2000), other than for measuring temporal changes in seismic velocity (e.g. Poupinet et al, 1984;Haase et al, 1995).…”

“…However, several source of errors affect the correlation estimate: our signal are noisy and band-limited, so that in general ∆ ik = ∆ ji + ∆ jk . We therefore follow Deichmann and Garcia-Fernandez (1992), and once the multiplet is composed of at least three events, we can express the generic time difference ∆ ij (j > 1, i > j) as a difference between ∆ i1 and ∆ j 1 , thus making the system of equations over-determined with respect to these latter quantities.…”

“…In both the approaches, the time-or frequency-domain cross-correlation functions are interpolated thus allowing for a precision in the estimate of the relative delays between pair of signals in the order of a tenth or even hundredth of the original digitization rate. More precisely, for the frequency-domain approach, the phase of the cross-spectrum is least-squares fitted by a linear relationship whose intercept with the y-axis gives the time difference, while in the time-domain the correlation function is interpolated through a polynomial (Deichmann and Garcia-Fernandez, 1992) or a sinc ( Augliera et al, 1995) function. We choose a method very similar to the latter one, and interpolate the correlation function using a cubic SP-line interpolator (Press et al, 1992) to estimate the maximum of the cross correlation function with a precision of 1 ms (one-tenth of the original sampling rate; see Fig.…”