Earthquakes in Switzerland and surrounding regions during 2012

Abstract

This report of the Swiss Seismological Service summarizes the seismic activity in Switzerland and surrounding regions during 2012. During this period, 497 earthquakes and 88 quarry blasts were detected and located in the region under consideration. With a total of only 13 events with M_L ≥ 2.5, the seismic activity in the year 2012 was far below the average over the previous 37 years. Most noteworthy were the earthquake sequence of Filisur (GR) in January with two events of M_L 3.3 and 3.5, the M_L 4.2 and M_L 3.5 earthquakes at a depth of 32 km below Zug in February and the M_L 3.6 event near Vallorcine in October. The epicentral intensity of the M_L 4.2 event close to Zug was IV, with a maximum intensity of V reached in a few areas, probably due to site amplification effects.

Zusammenfassung

Dieser Bericht des Schweizerischen Erdbebendienstes stellt eine Zusammenfassung der im Vorjahr in der Schweiz und Umgebung aufgetretenen Erdbeben dar. Im Jahr 2012 wurden im erwähnten Gebiet 497 Erdbeben sowie 88 Sprengungen erfasst und lokalisiert. Mit nur 13 Beben der Magnitude M_L ≥ 2.5, lag die seismische Aktivität im Jahr 2012 weit unter dem Durchschnitt der vorhergehenden 37 Jahre. Die bedeutendsten Ereignisse waren die Erdbebensequenz von Filisur im Januar mit zwei Beben der Magnitude M_L 3.5 und 3.3, die Beben in 32 km Tiefe bei Zug (M_L 4.2 und M_L 3.5) im Februar sowie das Beben bei Vallorcine (M_L 3.6) im Oktober. Das M_L 4.2 Beben nahe Zug verursachte Erschütterungen der Intensität IV. Sehr vereinzelt wurden Erschütterungen der Intensitäten V verspürt, vermutlich verursacht durch lokale Amplifikationseffekte.

Resumé

Le présent rapport du Service Sismologique Suisse résume l’activité sismique en Suisse et dans les régions limitrophes au cours de l’année 2012. Durant cette période, 497 tremblements de terre et 88 tirs de carrière ont été détectés et localisés dans la région considérée. Avec seulement 13 événements de magnitude M_L ≥ 2.5, l’activité sismique de l’année 2012 est bien en-dessous de la moyenne des 37 années précédentes. Les événements les plus significatifs ont été la séquence de Filisur en janvier avec deux séismes de magnitude M_L 3.5 et 3.3, les séismes à une profondeur de 32 km près de Zoug (M_L 4.2 et M_L 3.5) en février et le séisme près de Vallorcine en Octobre (M_L 3.6). L’intensité de l’événement de M_L 4.2 près de Zoug a été de IV, avec une intensité maximale de V atteinte dans quelques zones sans doute en lien avec des effets d’amplification locaux.

1 Introduction

Past earthquake activity in and around Switzerland has been documented in an uninterrupted series of annual reports from 1879 until 1963 (Jahresberichte des Schweizerischen Erdbebendienstes). Three additional annual reports have been published for the years 1972–1974. These reports together with historical records of earthquakes dating back to the thirteenth century have been summarized by Pavoni (1977) and provided the basis for the first seismic hazard map of Switzerland (Sägesser and Mayer-Rosa 1978). With the advent of routine data processing by computer, the wealth of data acquired by the nationwide seismograph network has been regularly documented in bulletins with detailed lists of all recorded events (Monthly Bulletin of the Swiss Seismological Service). Since 1996, annual reports summarizing the seismic activity in Switzerland and surrounding regions have been published in the present form (Baer et al. 1997, 1999, 2001, 2003, 2005, 2007; Deichmann et al. 1998, 2000a, 2002, 2004, 2006, 2008, 2009, 2010, 2011, 2012). In the course of reassessing the seismic hazard in Switzerland, a uniform earthquake catalogue covering both the historical and instrumental periods was compiled in 2002 (Fäh et al. 2003). The official seismic hazard map of Switzerland based on this catalogue was released in 2004 (Giardini et al. 2004; Wiemer et al. 2009). In 2009, the Earthquake Catalogue of Switzerland was revised (ECOS-09) and is now available on-line (http://www.seismo.ethz.ch/prod/catalog/index). In addition, numerous studies covering different aspects of the recent seismicity of Switzerland have been published in the scientific literature (for overviews and additional references see, e.g. Deichmann 1990; Pavoni and Roth 1990; Rüttener 1995; Rüttener et al. 1996; Pavoni et al. 1997; Deichmann et al. 2000b; Kastrup et al. 2004, 2007; Husen et al. 2007; Marschall et al. 2013).

2 Data acquisition and analysis

2.1 Seismic stations in operation during 2012

The Swiss Seismological Service (SED) operates two separate nationwide seismic networks, a high-gain predominantly broad-band seismometer network (Table 1) and a low-gain accelerograph network (Table 2). The former is designed to continuously monitor ongoing earthquake activity down to magnitudes well below the human perception threshold, whereas the latter is principally aimed at engineering concerns and thus focuses on recording ‘strong motions’ in urban areas. In addition, the Seismological Service operates a number of temporary stations for various projects (Table 3). SED stations with on-line data acquisition that were operational at the end of 2012 are shown in Fig. 1.

Table 1 High-gain seismograph stations of the Swiss national network operational at the end of 2012

Table 2 Strong-motion stations of the Swiss national network with on-line data acquisition operational at the end of 2012

Table 3 Local seismic networks with on-line data acquisition operational at the end of 2012

Fig. 1

Seismograph stations in Switzerland with on-line data acquisition operational at the end of 2012. The stations defined as high-gain (HG) are mostly equipped with broad-band or 5-second sensors, whereas the strong-motion stations (SM) are accelerometers (see also Tables 1, 2, 3)

The broadband station at Emosson dam, EMV, was removed in June 2012 due to construction in the area that has permanently destroyed the vault. A new broadband station in the vicinity is planned once a suitable site is found.

The ongoing densification of the Strong Motion accelerograph network with real-time continuous very-broadband accelerometers (Clinton et al. 2011; Cauzzi and Clinton 2013) continued apace in 2012. The first phase neared completion, with 12 new stations added (SAIG, SALTS, SARK, SBEG, SBRS, SBUH, SIEB, SINS, SMAO, SOLZ, STGK, SZUZ). At SBEG, an analogue accelerometric sensor installed in 1992 was removed. Note that in 2011, the triggered dial-up analogue strong motion hardware at SIOM and SVAM were replaced by real-time digital hardware, a fact not reflected in the 2011 yearly report.

The micro-network setup originally by the SED in 2005 in the region between the Lukmanier Pass and the Leventina Valley was dismantled in 2012. This network monitored an ongoing sequence of earthquakes in the immediate vicinity of the southern segment of the new Gotthard railway tunnel (Husen et al. 2011). The short period stations (and co-located acceleration sensors where installed) at CUNA, CURA, LUKA1, DOETR, RITOM and TONGO were removed. The broadband surface station NALPS at the southern shore of Lai di Nalps continues to run and has been included in the broadband seismic network.

In spring 2012, the ownership of the remaining operational Basel borehole stations that were installed in 2006 to monitor the geothermal energy project was passed from GeoPower Basel Ltd to the Kanton Basel Stadt. Station JOHAN was discontinued during the handover. MATTE and OTER2 are operated by SED under a contract with Kanton Basel Stadt, and complement the existing SED borehole station OTER1.

The earthquakes induced by the geothermal project in Basel have raised concerns about potential seismicity induced by other geothermal projects (even those that do not involve the enhancement of permeability through massive water injections). Therefore, responsible operators have established the practice to include local monitoring capabilities for such projects. The four-station short period array around the southeast tip of Lake Geneva, associated to a project near Noville to drill for natural gas, continued operation through 2012. An array to monitor a geothermal experiment planned for 2012 but delayed to 2013 near Sankt Gallen was installed in January and February 2012 in cooperation with the Sankt Galler Stadtwerke and the Swiss Federal Office of Energy—station SGT00 includes a borehole short period sensor at 205 m depth and a surface accelerometer located beside the drill site. Five additional broadband stations (SGT01, SGT02, SGT03, SGT04 and SGT05) are installed at the surface within a 10 km radius around the borehole landing point (Fig. 1).

A long-term project to monitor seismicity down to magnitude M_L 1.0 across north-eastern Switzerland began in 2011. This project, under a contract with the National Cooperative for the Disposal of Radioactive Waste (Nagra), aims to increase the understanding of seismotectonic processes in the vicinity of proposed sites for deep nuclear waste repositories. In order to reach these goals, the existing network in northern Switzerland and southern Germany is being significantly densified with seven surface stations and three short-period borehole stations at depths of 100–200 m. The additional stations will be in operation for a minimum of 10 years. In 2012, surface stations were installed at BERGE (in Germany) and EMMET; while at STIEG, a 123 m deep short period borehole sensor and a surface accelerometer were installed (Fig. 1). Due to changes in the scope of the project, the station SAIRA in the Swiss Jura was dismantled.

To improve the reliability of locations for events at the periphery or outside of Switzerland, the SED continues to be engaged in an ongoing cross-frontier cooperative effort to exchange seismic data in real-time. The SED continuously records and archives signals from stations in Austria operated by the Zentralanstalt für Meteorologie und Geodynamik in Vienna (ZAMG); in Italy operated by the Istituto Nazionale di Geofisica e Vulcanologia in Rome (INGV), the Istituto di Geofisica, Universita di Genova, the Zivilschutz der Autonomen Provinz Bozen-Südtirol and the Istituto Nazionale di Oceanografia e di Geofisica Sperimentale (OGS) in Trieste; in Germany from the Landeserdbebendienst Baden-Württemberg in Freiburg (LED) and the Bundesanstalt für Geowissenschaften und Rohstoffe in Hannover (BGR); and in France from the Réseau Sismologique et Géodésique Français (RESIF). A total of 35 foreign stations were monitored at the SED in 2012, and the number continues to increase as new high-quality stations come on-line in the border region.

2.2 Hypocenter location, magnitude, focal mechanisms and monitoring software

Since the year 2005, hypocenter locations of most of the local earthquakes have been determined using the software package NonLinLoc (Lomax et al. 2000). The P-wave velocity model used was derived from a 3D tomographic inversion of local earthquake data with constraints from controlled source seismics (Husen et al. 2003), and the S-velocities are calculated from the P-velocities using a Vp/Vs ratio of 1.71.

Local magnitudes (M_L) are calculated from the maximum amplitude of the horizontal components of the digital broad-band seismograms filtered to simulate the response of a Wood-Anderson seismograph. The attenuation with epicentral distance is accounted for by an empirically determined relation (Kradolfer and Mayer-Rosa, 1988). The final magnitude corresponds to the median value of all individual station magnitudes. For the stronger events, the traditional determination of focal mechanisms from the azimuthal distribution of first-motion polarities (fault-plane solutions) is complemented by moment tensors based on full-waveform inversion. This procedure, based on a time domain inversion scheme developed by Dreger (2003), also provides a moment magnitude, M_W, the best fitting double couple, and an optimal depth estimate based on the given location. An additional procedure has been implemented that routinely and automatically provides estimates of M_W also for earthquakes of lower magnitudes. M_W values are computed using a spectral fitting technique following the method of Edwards et al. (2010). The far-field signal moments are obtained by simultaneously fitting the spectrum of a theoretical source model (Brune 1970, 1971) along with path variable attenuation and event stress-drop to observed Fourier velocity spectra. The seismic moment is derived from the far-field signal moments assuming a simple geometrical spreading model that accounts for body and surface wave propagation. Site amplification is derived for the whole network following the approach detailed in Edwards et al. (2013). Using this information all data are corrected to a reference hard-rock recording site (Poggi et al. 2011) before computing M_W.

A more detailed documentation of the data analysis can be found in previous annual reports (Deichmann et al. 2006; Baer et al. 2007).

In October 2012, the Seismic Network migrated to a new software suite for automatic earthquake detection and characterization as well as manual event review and catalogue management. This software, SeisComP3 (Hanka et al. 2010; http://www.seiscomp3.org), is an open source, community supported tool developed at the GeoForschungsZentrum Potsdam. During a development and testing period lasting several years, all critical features from the previous internally developed software have been included in SeisComp3. This includes the Baer-Kradolfer automated picker (Baer and Kradolfer 1987), the Swiss Local magnitude relationship (Kradolfer and Mayer-Rosa 1988), the NonLinLoc earthquake location algorithm (Lomax et al. 2000), and all relevant velocity models (e.g. Husen et al. 2003). The extensive development and testing ensured that there is no change in the detection threshold of earthquakes and the location accuracy of the reviewed catalogue remains the same. Additionally, there is no systematic bias introduced to the magnitude estimation. SeisComP3 provides a modern software framework with an extensible modular structure, and we expect that the network will use this software for the foreseeable future.

3 Seismic activity during 2012

3.1 Overview

During 2012, the Swiss Seismological Service detected and located 497 earthquakes in the region shown in Fig. 2. Based on criteria such as the time of occurrence, the location, signal character or on direct communication, 88 additional seismic events were identified as quarry blasts.

Fig. 2

Epicenters and focal mechanisms of earthquakes recorded by the Swiss Seismological Service during 2012. Epicenters of events mentioned in the text are Beckenried (Bec), Breithorn (Bre), Filisur (Fil), Santa Maria (SMa), Vallorcine (Val), and Zug (Zug)

Magnitude values of the events recorded in 2012 range from M_L 0.0 to 4.2 (Fig. 3). The events with M_L ≥ 2.5 and the criteria used to assign the quality rating for the given locations as well as the corresponding estimated location accuracy are listed in Tables 4 and 5. Table 4 also includes the available M_w values derived from the spectral fitting method of Edwards et al. (2010) and from method of Dreger (2003) when available. Fault-plane solutions based on first-motion polarities are shown in Fig. 4 (see also Fig. 2) and their parameters are listed in Table 6.

Fig. 3

Earthquake activity during 2012: a magnitude of each event and cumulative number of events; b histogram of magnitudes

Table 4 Earthquakes with M_L ≥ 2.5

Table 5 Criteria and location uncertainty corresponding to the quality rating (Q) of the hypocentral parameters in the event list in Table 4

Fig. 4

Fault-plane solutions based on first-motion polarities (see Table 6). All stereographs are lower hemisphere, equal-area projections. Solid circles correspond to compressive first motion (up); empty circles correspond to dilatational first motion (down). Full waveform moment tensor solutions are shown as beach-ball symbols for the two largest events where high quality fits are obtained

Table 6 Focal mechanism parameters based on first-motion polarities of five earthquakes in 2012

Figure 5 shows the epicenters of the 893 earthquakes with M_L ≥ 2.5, which have been recorded in Switzerland and surrounding regions over the period 1975–2012. These events represent about 7 % of the total number of events detected during that time period in the same area. The chosen magnitude threshold of M_L 2.5 ensures that the data set is complete for the given period (Nanjo et al. 2010) and that the number of unidentified quarry blasts and of badly mislocated epicenters is negligible.

Fig. 5

Epicenters of earthquakes with magnitudes M_L ≥ 2.5, during the period 1975–2012. Grey circles denote earthquakes in the period 1975–2011, bold red circles indicate earthquakes in 2012. The grey dashed line and the solid grey box mark the location of the profile A–A′ shown in Fig. 7

3.2 Significant earthquakes of 2012

3.2.1 Filisur

On January 1st and 2nd, 2012, two moderate earthquakes with M_L 3.3 and 3.5 jolted the central parts of Kanton Graubünden. They were part of a sequence of nine events that occurred over 3 days with epicenters about 5 km east-northeast of Filisur (Marschall et al. 2013). The M_L 3.5 event was felt in the part of the Graubünden and maximum intensities of IV have been reported. With the closest station 14 km away (DAVOX), routinely computed focal depths are uncertain, and individual events scatter between 2 and 8 km, depending on the number of available arrivals. A comparison of the arrival-time differences between the reflection at the Moho (PmP) and the direct arrival (Pg) observed for the mainshock of this sequence at stations WILA and ZUR with those of one of the events of the Sertig sequence of 2003, located 8 km further to the east-northeast, shows that the source of the Filisur sequence was slightly shallower than the Sertig sequence (Fig. 6). Since the focal depth of the Sertig sequence (7 km) is well constrained by records of a temporary station installed in the immediate vicinity of the epicenter (Deichmann et al. 2004), we conclude that the hypocenters of the Filisur sequence are located at a depth of about 6 km. This was confirmed by 2D ray-trace modeling of the Pg and PmP travel times observed at several stations in northern Switzerland.

Fig. 6

Comparison of the PmP arrivals at stations WILA and ZUR of an event of the Sertig sequence (2003/07/18 11:01, M_L 3.9) and of the Filisur sequence (2012/01/02 01:42, M_L 3.5). The signals are integrated to displacement with a Wood-Anderson filter and filtered with a 1–20 Hz, 2nd order Butterworth band-pass filter

The fault-plane solution computed for the M_L 3.5 event (Table 6; Fig. 4) is based on take-off angles derived from the 2D ray-tracing assuming a linear velocity gradient between the base of the near-surface sediments and the Moho. The result is a normal-faulting focal mechanism with a NE–SW oriented T-axis, typical for this region (e.g. Marschall et al. 2013; Kastrup et al. 2004).

Judging from the high degree of similarity between the waveforms from the individual events in the sequence, the hypocenters must be closely co-located and their focal mechanisms nearly identical. To discriminate between the two nodal planes, a good constraint of the relative focal depths is essential. Given the large epicentral distance of the closest stations, this requires the inclusion of the PmP arrivals observed at stations in northern Switzerland in the cross-correlation. As a consequence, relative locations could be computed only for the three strongest events (Marschall et al. 2013). Mean travel-time residuals of the relative locations are between 2 and 3 ms, and the resulting location errors (one standard deviation) are on the order of 10 m horizontally and 23 m vertically. Within the location errors, the nodal plane dipping to the SW best matches the vertical distribution of the three hypocenters (Marschall et al. 2013).

3.2.2 Zug

The M_L 4.2 earthquake that occurred on February 11th at 11:45 pm local time (10:45 pm UTC) between Zuger- and Ägerisee, was the strongest earthquake to occur in Switzerland since the M_L 4.3 earthquake of Fribourg in 1999 and the M_L 4.9 Vallorcine earthquake in 2005, which was located just across the border with France, between Martigny and Chamonix. The hypocentral depth of 32 km is well constrained by P- and S-onsets at seven stations at epicentral distances between 20 and 40 km. According to the recently published crustal model of the Alps of Wagner et al. (2012), the Moho lies at a depth of about 34 km at this location (Fig. 7). Thus, the M_L 4.2 earthquake locates close to a previously known group of lower-crustal events in the area between the lakes of Zürich and Zug discussed in previous reports (e.g. Deichmann 1992; Deichmann et al. 2000a). Figure 8 shows a record section along a WSW trending profile of an M_L 3.5 aftershock with an almost identical hypocenter location at 32 km depth. In this direction the apparent velocity of the Pn arrivals is 8.15 km/s, while for the Pg arrivals beyond 80 km it is approximately 6.2 km/s. The unusually small Pn–Pg cross-over distance of about 70 km is evidence for the close proximity of the hypocenter to the Moho.

Fig. 7

Depth cross-section from the Black Forest to Lake Como (see Fig. 5 for location of profile) showing seismicity from 1984 to 2012 (circles) as recorded by the SED. The M_L 4.2 mainshock of February 11, 2012 is marked as a red dot. Only earthquakes within a swath of ±30 km of the profile (see box in Fig. 5) and with well constrained focal depths (i.e. azimuthal GAP ≤180°, number of observations ≥11, distance to closest station ≤30 km) are shown. These selection criteria are the same as those used for earlier compilations of focal depths (e.g. Deichmann et al. 2000a). Size of the circles is scaled by magnitude, as indicated. Crust–mantle boundary (Moho) from Wagner et al. (2012) along profile is shown as black line

Fig. 8

Record section of the seismograms of the M_L 3.5 Zug event of 2012/02/24 00:32 as a function of epicentral distance along a WSW trending profile. The signals are filtered with a 0.5–10 Hz Butterworth band-pass filter. In this direction, the apparent velocity of the Pn arrivals is 8.15 km/s and for the Pg arrivals beyond 80 km it is approximately 6.2 km/s. The unusually small Pn–Pg cross-over distance of about 70 km is evidence for the close proximity of the hypocenter to the Moho

The well-constrained fault-plane solution corresponds to a strike-slip mechanism with a minor component of normal-faulting (Table 6; Fig. 4). In agreement with other known focal mechanisms in this region, its P-axis trends towards NNW–SSE (Kastrup et al. 2004). In the 2 weeks following the February 11th mainshock, 3 aftershocks were recorded with similar locations and depths. A M_L 1.1 event occurred at 28 km depth on February 14th, followed by a M_L 1.6 event at 29 km depth on February 18th. A larger aftershock of M_L 3.5 occurred at 32 km depth on February 24th. The well-constrained fault-plane solution of this event is similar but not identical to that derived for the mainshock (Table 6; Fig. 4).

The M_L 4.2 mainshock was felt throughout Switzerland, and was widely felt in the eastern and central Alpine foreland, spreading to the eastern extensions of the Jura to the Basel area and to the Lake of Constance (Fig. 9a). Significantly fewer reports were received from the southern and western parts of Switzerland and the Alpine area as a whole, which is even more evident in the distribution of intensities of the M_L 3.5 aftershock (Fig. 9b). The epicentral intensity on the European Macroseismic Scale (EMS) of the mainshock was IV, with a relatively “flat” intensity field typical for deep events. The maximum intensity of V has been reported in a few areas, probably due to site amplification effects (Fig. 9a).

Fig. 9

Macroseismic intensities (EMS-98) of a the M_L 4.2 Zug mainshock of February 11 and b the strongest aftershock with M_L 3.5 of February 24

Figure 10 shows the ground motion recordings of the M_L 4.2 mainshock at two stations of the Swiss National Strong Motion Network (SSMNet) installed in Lucerne about 21 km from the epicenter. Station SLUB (Bramberg) is a rock site on a hill above the city center. Station SLUW is located in the deep basin of Lucerne, near the main train station. The recordings at SLUB show an impulsive, unilateral signal pulse of the S-phase, with a peak displacement of 0.07 mm. The recordings at station SLUW show long-lasting reverberations due to local site effects and the peak displacement is amplified by a factor of 2 compared to the rock site.

Fig. 10

Displacement seismograms of the M_L 4.2 earthquake of Zug recorded at two accelerographs in the city of Lucerne. Epicentral distance to both stations is about 21 km, distance between stations is about 1.6 km. To stabilize the double integration of the original accelerograms, the signals where filtered with a 0.5 Hz, 4th order, zero-phase high-pass Butterworth filter and detrended with a second order polynomial. Note the impulsive unilateral signal pulse of the S-phase on the two horizontal components at station SLUB (right), situated on hard rock, compared to the long duration reverberations of the S-phase at station SLUW (left), situated on unconsolidated sediments near the lake shore. Note also the amplification by a factor of 2 of the peak displacement (given below each trace in mm) at station SLUW compared to station SLUB

3.2.3 Beckenried

On September 4th three earthquakes with M_L 2.3, 2.8, and 2.9 occurred within 4 h near Beckenried on the southern shore of the Vierwaldstättersee. The events are located close to the epicenter of the M_L 3.0 earthquake of 2000 (Baer et al. 2001). The high similarity between the signals of the three events of 2012 (Figs. 11, 12) suggests that the hypocenters must be co-located and their focal mechanisms nearly identical. The epicenter location of the 2000 event is located 2–3 km north-northeast of the others, which is in agreement with the difference in S–P times observed at station BNALP (located to the south of the epicenters) shown in Fig. 11.

Fig. 11

Vertical component seismograms recorded at station BNALP of the M_L 3.0 Beckenried event of the year 2000 compared to the three Beckenried events of 2012 at epicentral distances of about 8 km. Note the larger S–P travel-time difference for the event of 2000, which is compatible with the calculated location difference between the hypocenter of this event and that of the three events of 2012. Note also the significantly larger P/S amplitude ratio of the first event of 2012 compared to that of the subsequent two events, which is indicative of a slightly different focal mechanism

Fig. 12

Seismograms (ground velocity; 0.5 Hz, 4th order, zero-phase Butterworth high-pass filter) of the three Beckenried events of 2012 recorded at station SGT04 (epicentral distance 81 km). The vertical components (left panel) are aligned on the Pg arrival; the E–W components (right panel) are aligned on the Sg arrival. Note the almost identical travel-time differences between the reflections at the Moho (PmP, SmS) and the direct waves (Pg, Sg) for the three events, which are indicative of an essentially identical focal depth. This is a rare example of a sequence of clear and impulsive Pg, PmP, Sg and SmS phases recorded at the same location

The focal depths of the 2012 M_L 2.8 and 2.9 events are constrained by P- and S-arrivals at stations BNALP and MUO at 8 and 13 km epicentral distance as well as by several Pn arrivals (stations KIZ, GUT, UBR and BRANT). At 9–10 km depth, these events are most likely located in the basement. This differentiates them from most of the other well-located events in central Switzerland, which occurred in the overlying sedimentary units (Deichmann et al. 2000b). Assuming a constant seismic velocity ratio (Vp/Vs) for the entire crystalline crust, the observed arrival-time differences between Pg, PmP and Sg, SmS at station SGT04 (Fig. 12) gives Vp/Vs = 1.73 ± 0.03.

The focal mechanism of the M_L 2.9 2012 event (Table 6; Fig. 4) is very similar to the one of 2000: strike-slip with more or less N–S and E–W trending nodal planes. The focal mechanisms of the M_L 2.9 and the M_L 2.8 events (not shown) are actually identical, as expected from the almost identical signals. The fault-plane solution of the M_L 2.3 event (not shown) is not well constrained, but it shows that the focal mechanism is not identical to that of the other events. This is also consistent with the differences in S/P amplitude ratios visible for station BNALP (less nodal) in Fig. 11. The orientation of the P- and T-axes is typical for earthquakes in central Switzerland (e.g. Kastrup et al. 2004).

3.2.4 Vallorcine

The M_L 3.6 earthquake that occurred on October 25th close to the village of Vallorcine (France) is part of the ongoing earthquake sequence that followed the M_L 4.9 earthquake of September 2005 (Fréchet et al. 2010). The M_L 3.6 event of October 2012 was the strongest earthquake in the region since the September 2005 mainshock. The 2012 event was felt in the southwestern part of the Valais and maximum intensities of IV have been reported. The routine epicenter location of the event is within the main cluster of activity, labeled C1 in the study of Fréchet et al. (2010).

The routinely computed focal depth of 11 km derived from using all available P-wave arrivals (71) is not consistent with focal depths computed for past seismicity of the Vallorcine sequence, including the M_L 4.9 2005 earthquake, despite the fact that this focal depth estimate is well constrained by the large number of Pg and Pn arrivals. Using only Pg- and Sg-wave arrivals up to 80 km distance yields a focal depth of 4 km, which is consistent with focal depths for past seismicity of the Vallorcine sequence. Due to the lack of observations at close distances and of Pn arrivals at large distances this focal depth estimate is not well constrained (uncertainty of ±3.0 km). A focal depth of 4 km yields large negative travel time residuals (>1 s) for Pn arrivals at larger distances but smaller travel time residuals for nearby stations OGSI (RESIF) and SALAN. This suggests that a focal depth of 4 km is more consistent with observations at nearby stations but that a focal depth of 11 km is more consistent with distant Pn arrivals. Hypocenter locations were computed using a 3-D P-wave velocity model of Husen et al. (2003), which includes a priori information on Moho depths beneath Switzerland and surrounding regions. This model, however, shows a relatively broad velocity gradient across the Moho instead of a rather sharp velocity gradient, which would be more realistic. Consequently, Pn arrivals are traced as diving waves in the model of Husen et al. (2003), which may lead to longer ray paths when compared to those computed in a model with a sharp velocity gradient across the Moho. Longer ray paths yield larger theoretical travel times and, hence, negative travel time residuals as observed.

Waveforms recorded at station AIGLE for the strongest aftershock (M_L 3.0) of the M_L 4.9 2005 Vallorcine earthquake and the M_L 3.6 2012 earthquake show nearly identical S–P times (Fig. 13). This suggests that the M_L 3.6 2012 earthquake occurred close to strongest aftershock of the M_L 4.9 2005 earthquake, consistent with the obtained epicenter location. PmP arrival times of the strongest aftershock and the M_L 3.6 2012 earthquake at station TORNY are very similar (Fig. 13). This suggests that both events occurred at a similar focal depth. Given the uncertainty in computing correct Pn ray paths, as discussed above, and a focal depth of 4 km for the aftershock constrained by a densely spaced deployment in September 2005 (Fréchet et al. 2010), a focal depth of 4 km for the M_L 3.6 2012 Vallorcine earthquake seems more plausible.

Fig. 13

Displacement seismograms of the Vallorcine events of 2005/09/08 (M_L 3.0) and 2012/10/25 (M_L 3.6) recorded at stations AIGLE (EW-component, 35 km epicentral distance) and TORNY (Z-component, 82 km epicentral distance). To stabilize the integration and to accentuate differences at high frequency, the signals are filtered with a 1 Hz, 4th order, zero-phase Butterworth high-pass filter and an 8 Hz, 2nd order Butterworth low-pass filter. Because the Pg onsets at TORNY are rather emergent, both sets of seismograms are aligned on the Pg arrival at station AIGLE. Note the high signal similarity as well as the almost identical Sg–Pg arrival-time differences at AIGLE and PmP–Pg arrival-time differences at TORNY for the two events. These similarities are evidence of similar focal mechanisms and almost identical epicenters and focal depths

The focal mechanism derived from the analysis of first-motion polarities of the M_L 3.6 event (Table 6; Fig. 4) is almost identical with the mechanism of the 2005 mainshock published by Fréchet et al. (2010). The high similarity of waveforms to the 2005 aftershock and the similarity in the focal mechanisms suggest that, like in the 2005 mainshock, the plane striking N63°E was active in the 2012 event.

3.2.5 Breithorn

The M_L 2.9 event of November 28th was located 10 km south of Zermatt (labeled Bre in Fig. 2). The focal depth of 2 km is constrained by P and S-arrivals recorded at epicentral distances of 10 km (station SATI (INGV)) and 20 km (station MMK) and therefore uncertain. Due to the inhomogeneous distribution of first-motion observations, a focal-plane solution could not be reliably determined. The mainshock was followed by 5 aftershocks with M_L 0.7–1.8 in November and December.

3.2.6 Santa Maria

During 2012, a sequence of 15 earthquakes occurred in a region east of Santa Maria (Val Müstair). Based on the routine location and the high similarity of waveforms recorded at station MOSI (Bozen), seven of these events with M_L 1.3–1.9 appear to be co-located (cluster labeled SMa in Fig. 2). The sequence of co-located events started on October 20th and the last event occurred on December 7th. The largest of the events on November 8th was followed by a secondary event separated by about 3 s.

3.2.7 Seismicity associated with the former Deep Heat Mining project in Basel

The seismic activity induced by the geothermal project in Basel in 2006 and 2007 (e.g. Deichmann and Ernst 2009; Deichmann and Giardini 2009) continued to decrease over the years 2008–2011, with only one event (M_L 0.6) being recorded by the national network in 2010. In 2012 activity seems to have picked up again: five events, with magnitudes between M_L = 0.9–1.2, were strong enough to be recorded by the national broadband network. Master event relocation confirms that the two events of February 26th 2010 and May 20th 2012 as well as the three events which occurred between December 12th and 15th 2012 are located in the same region as the hypocenters of the events which occurred towards the end of 2007 (i.e. above the casing shoe, at the southern periphery of the stimulated rock volume), whereas the event of October 3rd 2012 is located about 400 m below the casing shoe and about 400 m to the east of the main hypocentral cloud. Due to the removal of three of the borehole sensors, relative location uncertainties for the five events of 2012 are about twice as large as for the earlier events. However, with computed standard deviations of 76 m (E–W), 147 m (N–S) and 67 m (Z), the outlier location of the event of October 3rd 2012 is well constrained. The seismic activity in the stimulated rock volume is expected to stay above the natural background level for several years (Bachmann et al. 2011).

4 Discussion

In 2012, as in previous years, a large portion of the seismic activity was concentrated in the Valais and the immediately adjacent regions. Routinely calculated focal depths for all but 13 events recorded in 2012 are less than 16 km. All of the deeper hypocenters occurred below the Molasse Basin and the Jura of northern Switzerland and southern Germany, mainly between 7°E and 10°E (e.g. Deichmann 1992; Deichmann et al. 2000a). Overall, the seismic activity in and around Switzerland, in terms of the number of events, was low in 2012. The total number of 13 events with M_L ≥ 2.5 is substantially below the yearly average of about 24 events over the previous 37 years in this magnitude range. The M_L 4.2 lower-crustal earthquake near Zug, on the other hand, had the largest magnitude since the 2005 Vallorcine earthquake.