Crustal structure of the Eastern Alps and their foreland: Seismic model beneath the CEL10/Alp04 profile and tectonic implications

Summary

The subject of this paper concerns the seismic modelling of the crustal structure in the transition zone from the Bohemian Massif, across the Molasse basin and the Eastern Alps to the Southern Alps, mainly on the territory of Austria. The CEL10/Alp04 profile crosses the triple point of the European plate, Adriatic microplate and the recently identified Pannonian fragment. The seismic data along the presented profile originate from two large experiments: CELEBRATION 2000 and ALP 2002. The wavefield observed in the Eastern Alps is more complex than in the Bohemian Massif. Strong first arrivals (Pg) are distinct up to 60–90 km offset and are characterized by large variations in apparent velocity and amplitude. The contact between the Molasse basin and the Eastern Alps represents a barrier for seismic waves. Mid-crustal reflections (Pc) are usually recorded at short distance intervals (20–50 km) and are also characterized by variations in apparent velocity and amplitude. The Moho reflections are usually strong and well correlated, while Pn arrivals are only fragmentarily recorded. Detailed 2-D forward modelling of all refracted, post-critical and reflected phases, identified in the correlation process, was undertaken using a ray-tracing technique. The P-wave velocity in the crystalline upper crust of the Bohemian Massif and Molasse basin is about 6.15 km s⁻¹, which is slightly higher than in the Alpine area (about 6.0 km s⁻¹). Below the northern accretionary wedge of the Eastern Alps low-velocity sediments penetrate towards the southwest (SW) down to about 10 km depth. In the middle crust of the Alpine part, a reflective zone was modelled by a lamellae structure with alternating high and low velocities and thicknesses of about 2–3 km. The lower crust in this part of the model is more homogeneous, with a velocity of about 6.9 km s⁻¹. In the Bohemian Massif, a high-velocity (HV) body (V_P ∼ 7 km s⁻¹) of a few kilometres thickness was delineated in the depth interval 18–23 km. The crustal thickness along the CEL10/Alp04 profile changes from about 42–44 km in the SW (Alpine part), to around 40 km in the central part of the profile (Molasse basin), and 38–40 km in the NE (Bohemian Massif). The velocity in the uppermost mantle determined from Pn wave traveltimes is about 8 km s⁻¹ along the whole profile. The interpretation of the seismic wavefield is supplemented by an existing 3-D P-wave velocity model of the area. Main features derived by 2-D modelling (low velocities beneath the accretionary wedge, high velocities in the lower crust of the Bohemian Massif, Moho topography) well correlate with the 3-D model. Furthermore, the 3-D model allows assessing the lateral extent of significant features alongside the CEL10/Alp04 profile. This area is affected by both collision and escape tectonics. The high-reflectivity zone in the middle crust is explained by intermediate to mafic intrusions, rather than by ductile extensional deformation as generally observed in the lower crust.

Controlled source seismology, Body waves, Continental tectonics: compressional, Crustal structure

Introduction

During the last decade, Central Europe has been covered by a series of large seismic refraction and wide-angle reflection experiments (POLONAISE'97, CELEBRATION 2000, ALP 2002 and SUDETES 2003) to obtain a better knowledge of the lithospheric structure of this area. (Guterch et al. 1999, 2000, 2001; Brückl et al. 2003; Grad et al. 2003b; Guterch et al. 2003a,b). This paper looks at the two overlapping profiles CEL10 and Alp04, from the CELEBRATION 2000 and ALP 2002 experiments. Together with other profiles in this area, the CEL10/Alp04 profile provides a comprehensive seismic coverage along a section through the southeastern Bohemian Massif, across the Molasse basin, and through the Eastern and Southern Alps. Another paper on CEL10/Alp04 concentrates on the Bohemian part of this profile and the transition to Trans-European Suture Zone (TESZ; Hrubcová 2008). The first results from the evaluation of particular lines within the ALP 2002 project were obtained using 2-D ray-tracing techniques (Bleibinhaus et al. 2006; Brückl et al. 2007). Similar 2-D interpretation was also done for other areas of the Alps (e.g. Lüschen et al. 2004; Bleibinhaus & Gebrande 2006), for neighbouring regions of the Bohemian Massif (Ru°žek et al. 2003; Hrubcová 2005; Majdański et al. 2006), for the TESZ (Janik et al. 2002; Grad et al. 2003a; Janik et al. 2005), as well as for the Carpathians and the Pannonian Basin (Grad et al. 2006; Środa et al. 2006).

During each of the above-mentioned experiments, seismic shots were recorded simultaneously by all receivers. In each experiment, the data volume contains about 20 per cent inline recordings, while the rest are crossline recordings. Data of the third CELEBRATION 2000 deployment and from the whole ALP 2002 experiment were used for the generation of a 3-D model by stacking and tomographic methods (Behm 2006; Behm et al. 2007). 3-D interpretation has also been done for bordering regions: for the northern Bohemian Massif (Majdański et al. 2007; Ru°žek et al. 2007); for the TESZ and East European craton (EEC) in central Poland (Czuba et al. 2002; Środa et al. 2002); for SE Poland at the transition from EEC to the Carpathians (Malinowski et al. 2008) and for the northeastern part of the Pannonian basin (Hajnal et al. 2004). In the future, such 3-D approaches will provide an opportunity for consistent 3-D interpretation of all new surveys, and data from older profiles for the whole area between the Baltic and the Adriatic Seas. In this paper, we concentrate on the 2-D data of the CEL10/Alp04 profile, but we also refer to results of the 3-D model by Behm et al. (2007).

Tectonic Setting and Geological Cross-section

The Eastern Alps result from orogenic processes in the Cretaceous and early Tertiary, representing a continent-continent collision between the European and Adriatic-Apulian plates. The major geological units of the Eastern Alps and their surrounding tectonic provinces are shown in Fig. 1 (Oberhauser 1980; Franke & Żelaźniewicz 2000; Schmid et al. 2004). The Bohemian Massif in the north represents the European platform. To the south, European crust dips below the Molasse basin, the foreland of the Alpine orogen. The accretionary wedge of the Eastern Alps comprises the Flysch belt and Austroalpine nappes including the Northern Calcareous Alps (NCA). European crust has been exhumed in the Tauern Window. The Periadriatic Lineament (PAL) separates the Eastern Alps from the Southern Alps. The latter are bounded to the south by the External Dinarides, the Po plain and the Istria Peninsula. To the northeast (NE), the Eastern Alps continue into the Carpathians, and to the east they border to the Pannonian basin.

Figure 1.

Tectonic setting of the investigated area (top) and geologic cross-section along profile CEL10/Alp04 (bottom). Map and cross-section are compiled in accordance with Schmid et al. (2004), Oberhauser (1980) and Franke & Żelaźniewicz (2000). On the cross-section numbered triangles refer to shotpoints. SP26800 shows common position with another shot fired in this place: SP26801. Crossing points with Alp01, Alp02 and CEL09 are indicated. Geological units and major faults are designated by the following abbreviations: in the Bohemian Massif: BA, Barrandian; MD—Moldanubian, MR—Moravian; in the Eastern Alps: F—Flysch belt, NCA—Northern Calcareous Alps, GW—Greywacke Zone, MC—Mesozoic cover of Austroalpine basement nappes, ‘**’ designates the Austroalpine basement nappes DG—Drautal/Gurktal, OB—Ötztal/Bundschuh, KW—Koralpe/Wölz, TW—Tauern Window, G—Gneiss core of TW; Southern Alps: ST—post-volcanic and Mesozoic cover, SB—Palaeozoic basement; ED—External Dinarides; ID—Internal Dinarides; AF—Adriatic foreland; TU—Tisza Unit. Thrusts and lineaments: NAT—North Alpine Thrust, NCT—Northern Calcareous Alps Thrust, SEMP—Salzach/Enns/Maria-Zell/Puchberg-Line, PAL—Periadriatic Lineament; SAT—South Alpine Thrust, CT—Cicarija Thrust, MHZ—Mid-Hungarian Zone.

Major tectonic processes of the Eastern Alps include orogenic cycles in the Cretaceous (Eoalpine phase) and early Tertiary, representing a continent-continent collision between the European and Adriatic-Apulian plates. Crustal shortening of the Eastern Alps in the north-south direction followed. The maximum extent of shortening is assumed to be 100 km, which corresponds to 50 per cent of the original width. Since the Late Oligocene and Early Miocene, the ongoing north-south oriented compression of the Eastern Alps has been accompanied by vertical and lateral extrusion (in particular of the Tauern Window) and tectonic escape of large crustal wedges to the unconstrained margin represented by the Pannonian basin in the east. Major fault systems (e.g. PAL, SEMP; see Fig. 1) were reactivated or formed by this tectonic process (Ratschbacher et al. 1991).

A geological cross-section along the CEL10/Alp04 profile down to a depth of 5 km is shown in Fig. 1. This cross-section is based on geological maps and sections by Franke & Żelaźniewicz (2000), Oberhauser (1980), Schmid et al. (2004) and Wessely (2006). In the southwest (SW), the profile begins at the northwestern corner of the External Dinarides. After crossing the South-Alpine thrust fault, the profile enters the Southern Alps with their Mesozoic cover and further to the north with their Palaeozoic basement. Around the profile distance of 60 km, the profile intersects the PAL and reaches the Eastern Alps. This is also the location of the most southwestern shotpoint SP31140. Within the Eastern Alps, the profile first crosses the southern Mesozoic cover (MC in Fig. 1) and the crystalline basement of the Austroalpine nappes (DG—Drautal/Gurktal, OB—Ötztal/Bundschuh, KW-Koralpe/Wölz in Fig. 1), then it crosses the Greywacke Zone (GW), and the NCA. The NCA overthrust the Flysch belt in the north (NCT). The Eastern Alps and the Flysch belt form the accretionary wedge of the Eastern Alpine orogen, which overthrusts the Molasse at the North Alpine thrust (NAT). The extent of the Molasse zone is less than 20 km along CEL10/Alp04. Thereafter, the profile continues within Moldanubian and Moravian units of the Bohemian Massif.

Seismic Data

The field layout of seismic experiments within the territory of Austria and its surroundings is shown in Fig. 2. The shots and receivers used in this paper are highlighted. The northern part of profile Alp04 from the ALP 2002 experiment overlaps with the southern part of the CEL10 profile, recorded during the CELEBRATION 2000 experiment. The combined CEL10/Alp04 profile presented in this paper is 380-km long, with nine shotpoints at an average interval of 40 km. The profile extends from the Austrian/Czech border in the north to upper Italy in the south. Along the CEL10 part of the profile 89 single channel Texan-RefTek 125 receivers were deployed with a 3-km average spacing, which recorded eight shots beginning with SP20010 in the SW and ending with SP29140 in the NE. Along Alp04, 68 receivers with a 5.5-km average spacing recorded SP31140. These 68 receivers comprised 30 three-component recording units with 1-Hz seismometers, and 38 single channel Texan-RefTek 125 type with 4.5-Hz vertical geophones. The sampling rate was 0.01 s and the recording time window was 300 s for each shot. Table 1 summarizes information on the shots used in the present study. SP31140 is the crossing point of the CEL10/Alp04 profile with two other profiles, Alp01 and Alp02 (Brückl et al. 2007). The northernmost shot used here (SP29140) is located on profile CEL09 close to its intersection with CEL10. The standard shooting procedure in Austria was to employ 200–500 kg of explosives distributed in five to eight boreholes 30–50-m deep. Shots 26800 and 26801 were blasted at approximately the same location in an open pit mine using a production shooting scheme. The timing of the seismic shots was either done by automatic shooting at GPS-controlled time or by recording the ignition current. The field records were cut to a length of 100 s starting at 0 s of reduced time for a reduction velocity of 8 km s⁻¹.

Figure 2.

Field layout of seismic experiments 2000–2003 in the area of Austria. Large stars with numbers are shot locations along CEL10 and Alp04 profiles. Table 1 provides detailed information about all used shotpoints. Smaller stars are shot locations along other CELEBRATION 2000, ALP 2002 and SUDETES 2003 profiles. Receiver locations are shown by dots. The yellow bar shows the CEL10/Alp04 profile modelled in this paper; the grey diamond in the southwestern end of the profile shows ‘zero’ of profile with geographic coordinates ϕ = 46.1852°N, λ = 13.0919°E. Country symbols: A—Austria, CZ—Czech Republic, D—Germany, H—Hungary, HR—Croatia, I—Italy, SLO—Slovenia.

Table 1.

Shotpoint data for profile CEL10/Alp04 in Austria; shot 26 800 was in a quarry and other 26 801 from this location was recorded.

Generally, the P wavefield recorded on the CEL10/Alp04 profile has a rather good signal-to-noise (S/N) ratio, but there are considerable differences between NE (Bohemian Massif) and SW (Alps). In Fig. 3 for shots SP20010 and SP29140 (recorded in Austria), examples of the full wavefield are shown for a reduced time from 0 to 40 s (reduction velocity 8 km s⁻¹). All the data were used to derive the velocity model of the crust presented in Fig. 4; its accuracy and resolution were tested for different phases (Fig. 5). Those parts of the sections documenting features in the crust are shown in Figs 6–10 and are discussed in the next sections.

Figure 3.

Example of trace-normalized vertical-component seismic record sections of P and S waves for SP20010 (left) and SP29140 (right) along profile CEL10. A band pass filter (2–15 Hz) has been applied. Pg, Sg—waves refracted from the basement; Pc, Sc—waves reflected from mid-crustal discontinuities; P*, S*—waves refracted from HV body in the middle crust; PmP, SmS—reflected waves from the Moho; Pcrustal, Scrustal—overcritical crustal waves; Pn, Sn—refracted waves from the Moho; P^I—waves from the lower lithosphere. Pg# and Sg# show places of abrupt termination of Pg and Sg traveltime curves related to low-velocity sediments. Reduction velocity = 8.0 km s⁻¹.

Figure 4.

Compilation of seismic models along CEL10/Alp04 profile. At the top, topography along the profile is shown. Numbered triangles refer to shotpoints; shot with underlined number 26800 shows common position with shot SP26801. Crossing point with Alp01 and Alp02 profiles is at SP31140. Vertical exaggeration Y/X is 4:1 for seimsic models, and 25:1 for topography. Depth in models is relative to sea level. (a) 2-D velocity model for CEL10/Alp04 profile obtained by forward ray-tracing modelling using the SEIS83 package (Červený & Pšenčík 1983b). The thick solid lines are well-documented layer boundaries and dashed lines are continuations of boundaries needed in ray-tracing model; thin lines are P-wave iso-velocity contours in km s⁻¹. Areas of missing ray coverage are marked by grey overlay. (b) Vertical section through the 3-D seismic model (by Behm et al. 2007). The thick black line indicates the Moho where it is constrained by data, a dashed line indicates the Moho obtained by interpolation from nearby data values. Areas of missing coverage are white.

Figure 5.

An assessment of model uncertainties from traveltimes calculated using the SEIS83 ray-tracing technique for SP20010. Thick lines marked by Pg, PmP and Sg are traveltimes calculated for the final ray-tracing model of the structure shown in Fig. 4a. The arrival times of the Pg phase with the velocity perturbed by +0.2 km s⁻¹ and −0.2 km s⁻¹ shown by lighter lines are clearly too early and too late in relation to recorded first breaks. This shows that the velocity of the upper crust could be determined from the Pg wave with an uncertainty ±0.1 km s⁻¹. The arrival time of PmP phase reflected from the Moho in the final model (∼42 km depth) is shown together with arrival times with the depth perturbed by ±2 km marked by thinner lines. The sensitivity for estimation of the S-wave velocity in the upper crust was tested using Sg phase with V_P/V_S perturbed by ± 0.03. Early and late arrivals of the Sg phase are shown by lighter lines. See the text for further discussion. For this section, a bandpass filter (2–10 Hz) has been applied. Reduction velocity is 4.5 km s⁻¹. In the inset, enlarged 15 right side traces are shown with 1-s time bar.

Figure 6.

Amplitude-normalized seismic record sections for SP20020, SP20030 and SP26801 with theoretical traveltimes of P waves calculated for the crustal model derived using the SEIS83 ray-tracing technique: (a) synthetic seismograms, (b) record sections, (c) model and diagrams with chosen rays. Phases labelled as in Fig. 3. Numbers are P-wave velocities in km s⁻¹.

Figure 7.

Traveltimes of Pg wave for shotpoint SP31140 recorded in six directions along profiles Alp01, Alp02 and Alp04. Seismic record sections for SP31140 show slightly lower apparent velocities towards NE and ESE, and slightly higher towards S. In the traveltime plot, the vertical size of ellipses corresponds to time error of first arrival picking: ±0.05 s (offsets 20–30 km), ±0.10 s (offsets > 30 km) and ±0.15 s (for very low S/N ratio).

Figure 8.

Comparison of modelling results for SP31140. (a) Synthetic seismograms calculated using reflectivity method (Fuchs & Müller 1971) for 1-D, (b) synthetic seismograms of P waves calculated for the crustal model derived using the SEIS83 ray-tracing technique, (c) amplitude-normalized seismic record sections with theoretical traveltimes for Pg, PmP and Pn waves and (d) model with chosen rays. Phases labelled as in Fig. 3. Numbers are P-wave velocities in km s⁻¹.

Figure 9.

Interpretation and modelling of record section for SP20020: (a) synthetic seismograms, (b) trace-normalized seismic record sections with theoretical traveltimes calculated for the 2-D ray-tracing model, (c) diagram with chosen rays. P*—wave from the HV body in the middle crust; other phases labelled as in Fig. 3. Note termination of Pg phase at a distance of about 240 km and a strong Pn wave. Numbers are P-wave velocities in km s⁻¹.

Figure 10.

Interpretation and modelling of record section for SP20060: (a) synthetic seismograms, (b) trace-normalized seismic record sections with theoretical traveltimes calculated for the 2-D ray-tracing model, (c) diagram with chosen rays. P*—wave from the HV body in the middle crust; other phases labelled as in Fig. 3. Numbers are P-wave velocities in km s⁻¹.

Wavefield

Waves from the sedimentary cover (Psed) are observed in first arrivals only in the close vicinity of a few shotpoints, up to offsets of about 10–20 km (e.g. SP20060 in Fig. 10). The main crustal phases (Pg, Pc) and Moho reflections (PmP) were correlated for all record sections, while the interpretation of refracted waves from the uppermost mantle (Pn) is sometimes hampered by low S/N ratio. The Pg phase can be observed up to 60–90 km offset in the southwestern part of the profile and up to 150 km in the northeastern part. Its apparent velocity ranges from 5.8 to 6.2 km s⁻¹. An interesting feature is the abrupt termination of the Pg wave at a profile distance of 240 km (marked by Pg# in the left section of Fig. 3 and in Figs 6 and 9). Mid-crustal waves (Pc, P*) are usually recorded along short distance intervals (20–50 km). The exception is the record section for the southernmost SP31140, where a strong mid-crustal reflection (Pc) is followed by a 2–3-s long group of reflected signals (in Fig. 8c in distance 100–160 km and reduced time 4.5–6.5 s). The Moho reflections (PmP) are usually strong and well correlated. Moho refractions (Pn) in the southern part of the profile were recorded only fragmentarily, with weak arrivals, for example, for SP20010 (Fig. 3) and SP20060 (Fig. 10). In the northern part of the profile, Pn arrivals are very clear, for example, for SP20020 (Fig. 9) and SP29140 (Fig. 3). Mid-crustal overcritical reflections (Pcrustal and Scrustal) are well developed in the northern part of the profile (see e.g. SP29140 in Fig. 3), but are not observed for shots in the Eastern Alps. As seen from Fig. 3, main P phases have their counterparts in the observed S wavefield. Particularly, the northernmost shot SP29140 shows distinct S phases. Mantle lithospheric waves (P^I) were observed only fragmentarily (Figs 8–10). The upper mantle reflector is well resolved only between profile distances 150 and 240 km, and its depth is about 70 km.

Derivation of Crustal Model Using a Ray-tracing Technique

Detailed 2-D forward modelling of all refracted, reflected and post-critical phases identified in the correlation process was undertaken using a ray-tracing technique. The sometimes-difficult identification and correlation of seismic phases were done manually on a computer screen using the ZPLOT software which allows scaling, filtering and reduction velocity (Zelt 1994; Środa 1999). Computations of traveltimes, ray paths and synthetic seismograms were performed using the ray-theory package SEIS83 (Červený & Pšenčík 1983a,b) enhanced by employing the interactive graphical interfaces MODEL (Komminaho 1997). The 2-D velocity model along the profile CEL10/Alp04 was successively altered by trial and error, and traveltimes were recalculated many times until agreement was obtained between observed and model-derived traveltimes. A misfit of the order of 0.1–0.2 s was typical.

Apart of P waves, also S-wave modelling was performed for some shotpoints along the CEL10/Alp04 by determining the V_P/V_S ratio for each layer. As S-wave traveltime data were not sufficient for independent modelling of V_S structure, we kept the geometry of the P-wave model (Fig. 4a) and fitted theoretical S-wave traveltimes to the observed ones by adapting the V_P/V_S ratio. In the uppermost 2 km, we assumed V_P/V_S = 1.80. In the crystalline basement with V_P = 6.0–6.2 km s⁻¹, the shear wave velocities are relatively higher, and the modelled V_P/V_S is about 1.68. These values were well determined from good quality Pg and Sg wave traveltimes. For the deeper part of the model, a standard value V_P/V_S value of 1.73 fits the observed S-wave traveltimes quite well. Particularly in the range of the HV body in the middle crust, a V_P/V_S ratio of 1.73 explains the observed P* and S* phases very well.

In addition to kinematic modelling, synthetic seismograms (including both P and S waves) were calculated to control velocity gradients within the layers and velocity contrasts across seismic boundaries. Finally, the synthetic seismograms show good qualitative agreement with the relative amplitudes of observed refracted and reflected waves.

The model derived for the structure along the profile CEL10/Alp04 is shown in Fig. 4a. The model shows large variations in the crustal structure, while the Moho topography changes in a relatively narrow depth interval of 38–44 km. P-wave velocities in the CEL10/Alp04 model agree with the models at cross-points with other profiles: Alp01 and Alp02 profiles in the south, beneath SP31140 (Brückl et al. 2007) and with CEL09 profile in the north, beneath SP29140 (Hrubcová 2005).

An assessment of the resolution and uncertainty of 2-D ray trace modelling

The resolution achieved by refraction and wide-angle reflection techniques was investigated when previous experiments such as POLONAISE'97 and CELEBRATION 2000 were interpreted. The corresponding profiles were characterized by similar methodology, source and receiver density and comparable data quality (Janik et al. 2002; Grad et al. 2003a, 2006). For the CEL10/Alp04 profile, an assessment of model uncertainties for calculated traveltimes using the SEIS83 ray-tracing technique is illustrated in Fig. 5. Traveltimes of Pg, PmP and Sg phases, calculated for the 2-D model, are marked by thick lines. The arrival times of the Pg phase with the velocity perturbed by +0.2 and −0.2 km s⁻¹ shown by thinner lines are noteably too early and too late in relation to recorded first breaks. The arrival times of PmP phases reflected from the Moho in the final model (∼42 km depth) are shown together with traveltimes calculated for the Moho being perturbed by ±2 km depth (marked by lighter lines). The sensitivity for estimation of the S-wave velocity in the upper crust was tested using Sg phases with V_P/V_S perturbed by ±0.03. Early and late arrivals of the Sg phase are shown by thinner lines. This indicates that for the upper crust (consolidated basement), where the coverage by Pg waves is at its highest, uncertainty of velocity in the basement could be ±0.1 km s⁻¹. We expect a similar uncertainty in the uppermost mantle (for areas where Pn waves are well recorded). Although waves refracted from the lower crust are very rarely observed in the first arrivals, in many cases the situation improves because of well-recorded overcritical crustal waves (Pcrustal) that penetrate the lower crust (the uncertainty of velocity determination here is about ±0.2 km s⁻¹). The depths of mid-crustal boundaries are usually determined with a depth uncertainty of ±2–3 km, and the Moho boundary with an even lower uncertainty of ±1–2 km, in the case of good ray coverage. The misfit in Moho topography between the new model 2-D (Fig. 4a) and the 3-D model (Fig. 4b) along CEL10/Alp04 profile is smaller than ±2 km.

The accuracy of S-wave velocity, or rather V_P/V_S ratio, could be estimated as ±0.02 for the upper crystalline crust (determined from Pg and Sg waves), and ±0.03 or more for deeper parts of the model where only envelopes of waves were correlated. In the case of low-quality data for shotpoints (record sections) with big noise, some areas could be of lower accuracy and resolution.

In the process of modelling, however, the limitations of the ray theory must be kept in mind. In addition, 2-D modelling does not take into account out-of-plane refracted and reflected arrivals, which must have occurred particularly in such structurally complex area as the Alps.

VERTICAL SECTION THROUGH 3-D SEISMIC MODEL ALONG CEL10/Alp04

Behm et al. (2007) used a 3-D approach to resolve the P-wave velocity structure of the crust, Moho topography and sub-Moho velocity in the Eastern Alps and their surroundings based on the ALP 2002 and CELEBRATION 2000 data. Their data set comprises approximately 79 000 traces of which about 80 per cent are crossline data (Fig. 2). The CEL10/Alp04 profile is located at the centre of their investigated area and is well covered by rays with different azimuths. The method is based on stacking and inversion techniques designed for 3-D wide-angle reflection and refraction data. Stacking was required since the S/N ratio was sometimes low in the Eastern Alps (Behm 2006; Behm et al. 2007). The applied methods make no difference between inline and crossline data and therefore ensure that the whole data set is used. Stacking of diving waves through the crust (Psed, Pg, Pcrustal) yields local 1-D traveltime curves, which are inverted for 1-D velocity models. These local 1-D models are combined to give a 3-D velocity model. Stacking of Pn waves provides velocities of the uppermost mantle and the Moho topography expressed as delay times. These delay times are combined with two-way traveltimes obtained from stacking of PmP waves. All models based on stacking are supplemented by traditional methods based on traveltime picks from the (single-fold) record sections wherever the S/N ratio permits accurate picking.

The crust is represented by a smooth 3-D velocity distribution and the Moho is modelled as a first-order discontinuity. Compared to standard procedures (e.g. 3-D tomography based solely on traveltime picks from record sections), the application of stacking results in models with large coverage but low resolution. In the following, we describe the most important features of the 3-D model along a vertical section following the CEL10/Alp04 profile (Fig. 4b).

Velocities in the upper crust (down to 10 km depth) of the Southern Alps and around the PAL (profile distance 0–80 km) range between 5.5 and 6.1 km s⁻¹. The upper crust of crystalline units in the Eastern Alps (∼90–150 km) shows higher velocities (5.7–6.2 km s⁻¹). Below the NCA and Flysch and Molasse basins (∼190–280 km), relatively low velocities (5.7–6.0 km s⁻¹) are found at a depth between 5 and 10 km, whereas the upper crust of the Bohemian Massif has moderate to high velocities again (5.6–6.2 km s⁻¹). The middle crust (10–25 km depth) shows velocities between 6.1 and 6.6 km s⁻¹, and no significant lateral velocity contrasts are found. However, the model also covers the lower crust in the northern part of the profile where relatively high velocities (6.7–6.9 km s⁻¹) are determined in the 25–35 km depth range. The adjacent area in the vicinity of the Vienna Basin (to the east of the northern end of profile CEL01/Alp04) shows even higher velocities in these depths (7.0–7.2 km s⁻¹).

The Moho has a depth of 35 km in the northernmost part of the profile. To the south, it remains relatively flat below the Bohemian Massif and starts to slope down below the NCA and reaches a depth of 43 km. Between profile distance 160 and 40 km, the Moho cannot be resolved along CEL10/Alp04, but is interpolated from nearby data values. Only at the southernmost end, a relatively deep Moho (45 km) is observed.

Discussion of Significant Structures Derived from P Waves

In the following section, we show seismic data, synthetic seismograms and ray diagrams indicating noteworthy structures of the seismic model of the CEL10/Alp04 profile (Fig. 4a). We further relate these structures to corresponding features of the 3-D model (Fig. 4b) to support the 2-D interpretation and to gain information on the lateral extent alongside the profile.

Low-velocity zone dipping under NCA and Greywacke Zone

Waves from the sedimentary cover or other near surface strata with P-wave velocities ranging from 5.0 to 5.5 km s⁻¹ are observed in the close vicinity of a few shotpoints (e.g. SP29140). The corresponding layer, about 2-km thick, can be followed along most of the profile. Clear arrivals of refracted and reflected P waves from the crystalline crust are typically observed along the whole profile. Velocities in the crystalline upper crust of the Bohemian Massif and Molasse basin are about 6.15 km s⁻¹, slightly higher than in the Alpine area (about 6.0 km s⁻¹). An exception is the record sections in the central part of the profile, where Pg phases travelling from SW to NE (SP20020, SP20030, SP26801), terminate at a profile distance of 240 km. Their termination is abrupt and in the record sections marked by Pg# (Figs 3 and 6). The corresponding shot-receiver distances are 90 km for SP20020, 60 km for SP20030 and 40 km for SP26801. We modelled this abrupt termination by a low-velocity zone (5.3–5.6 km s⁻¹) dipping under NCA and Greywacke Zone down to a depth of about 8 km. It is difficult to constrain velocity and thickness of this ‘low-velocity body’, which is surrounded by basement with higher velocities. However, its top is well constrained by reflected waves, and particularly from termination points of refracted waves in the crystalline basement Pg#. Synthetic seismograms (Fig. 6a) and ray diagrams (Fig. 6c) of shots SP20020, SP20030, SP26801 prove the consistency of our model with seismic data. Further to the SW, we introduce a nearly horizontal reflector at about a 10 km depth to explain crustal reflections observed in the distance interval of 120–160 km. A similar shape of the south dipping low-velocity zone was interpreted along the profile CEL10/Alp04 by Hrubcová (2008) and along profile Alp01 (about 150 km to the west) by Brückl et al. (2007). Corresponding low velocities are also obtained from the 3-D model where they extend at least 150 km to the west.

Basement and middle crust around PAL

Shotpoint 31140 in the southwestern part of the profile was a common point for three profiles: Alp01, Alp02 and Alp04. It gives the opportunity to compare Pg wave arrivals in six different directions, with an azimuth step of about 60° (Fig. 7). To enhance Pg first arrivals, the record sections are shown with very big amplification, even if later arrivals are overamplified. The difference between the traveltimes is not large, but traveltimes of Pg waves show slightly lower apparent velocities towards NE and ESE, and slightly higher towards south. The maximum arrival time difference at distances 50–80 km is only about 0.3–0.4 s (less than 3–4 per cent). This observation indicates that anisotropy in the upper crust is not significant. In the neighbouring areas of the Tauern Window, the Bohemian Massif in the Czech Republic and in the marginal zone of the EEC in SE Poland, the anisotropy of the upper crust can reach 5–10 per cent (Ru°žek et al. 2003; Lüschen et al. 2004; Vavryčuk et al. 2004; Bleibinhaus & Gebrande 2006; Środa 2006).

In the southwestern part of the CEL10/Alp04 profile, a strong mid-crustal reflection (Pc) marks the beginning of a 2–3-s long zone of ‘ringing’ reflections (SP31140 in Fig. 8c). Within this zone, amplitudes are higher than for Pg waves and a few individual ‘reflections’ could be correlated. Between the end of reflections and PmP arrivals, no strong arrivals are observed, which could be an indication of a more homogeneous lower crust in this part of the profile. The reflective zone observed in the record section was modelling using 2-D ray-tracing technique (Červený & Pšenčík 1983a,b; Fig. 8b) as a laminated middle crust with alternating high (6.4 and 6.6 km s⁻¹) and low velocities (6.3 and 6.4 km s⁻¹), and a thickness of about 2–3 km for each lamella. However, in the ray-tracing synthetic seismograms (Fig. 8b), we do not observe some reflected phases with their characteristic coda. Therefore, we used 1-D reflectivity method (Fuchs & Müller 1971; Fig. 8a) to simulate variability of the middle crust and transition to the Moho discontinuity. The 1-D velocity model used for the reflectivity method represents the average of the 2-D ray-tracing model between the distances of 50 and 100 km. Both synthetic sections in Fig. 8 reproduce the ringing reflections as observed in the record, and reflectivity method shows complexity of the wavefield not visible in the ray-tracing synthetics. The lower crust in this part of the model seems to be more homogeneous, and a velocity of about 6.9 km s⁻¹ is derived from PmP waves.

Further indications for increased crustal reflectivity in this part has also been reported from receiver function analyses on 30 passive monitoring stations deployed along the Austrian part of the CEL10/Alp04 profile during the ALP 2002 experiment (Rumpfhuber & ALP 2002 Working Group 2004).

Middle and lower crust of the European plate

In the central and NE parts of the profile (distance interval ∼170–320 km), an HV body of a few km thickness is delineated (V_P ∼ 7 km s⁻¹) in the depth interval 18–23 km. It is constrained using reflected (Pc) and refracted (P*) waves from several shotpoints (e.g. SP20020 and SP20060; Figs 9 and 10). At the northern end of the profile, an HV lower crust is found, with velocities of 7.2–7.4 km s⁻¹ in the depth interval 26–40 km. The estimation of these high velocities is based on refracted waves (secondary arrivals) in the offset range 200–250 km (e.g. SP 26801). The HV lower crust inferred from 2-D modelling at the northern end of the profile is also clearly visible in the 3-D model where it extends to the southeast (Fig. 11a).

Figure 11.

(a) Comparison of average P-wave velocities between 28- and 32 km depth from 2-D model (bar along CEL10/Alp04), and a depth slice at 30 km through the 3-D velocity model by Behm et al. (2007). (b) Moho surface and superimposed depth contours from the 3-D model by Behm et al. (2007). Bold yellow lines separate European, Adriatic and Pannonian plates. Bars indicated with (1) and (2) show the Moho depths obtained from 2-D ray tracing: (1) is associated with the European Moho and (2) is the protuberance (see the text) and correlates with the Pannonian Moho.

Fragmentation of the Moho

PmP reflections from SP31140 with depth points around a profile distance of 100 km are observed as a strong wave (SP31140; Fig. 8c). Further to the NE, the Moho reflections, for example, from SP20020 (Fig. 9b) and SP20060 (Fig. 10b), are clear and form about 2-s long wave groups. Their midpoints are situated in the distance range of about 150–250 km. In this part of the profile and further to NE, the Moho is also well constrained by Pn phases. The crustal thickness along CEL10/Alp04 changes from about 44–42 km in the SW (Alpine part) to about 40 km in the central part of the profile (Molasse basin) and to 38–40 km in the NE (Bohemian Massif). The velocity in the uppermost mantle determined from Pn wave traveltimes is about 8 km s⁻¹ along the whole profile length.

The southward directed dip of the Moho is interrupted by a bulge at a profile distance of 100 km which was modelled using reflected waves from SP31140 (Fig. 8). A projection of the profile onto the Moho map (Fig. 11b; Behm et al. 2007) shows that CEL10/Alp04 crosses three Moho fragments that are the European, Pannonian and Adriatic Moho. Thus, the bulge in the 2-D model is interpreted as the westernmost tip of the Pannonian Moho fragment. Along the European Moho, where both the 2-D and the 3-D models provide Moho depths, the difference (3-D subtracted by 2-D) is −1.1 ± 0.8 km.

Tectonic Interpretation

The area of the CEL10/Alp04 profile is characterized by a complex tectonic setting. The profile begins in the Po plain near the northwestern corner of the External Dinarides and continues northeastwards over the Southern and Eastern Alps to the Flysch belt and the Molasse basin. It passes the central range of the Eastern Alps about 20 km east of the Tauern Window. Finally, it reaches the southeastern border of the Bohemian Massif near the Moldanubian-Moravian thrust. According to the Moho map constructed by Behm et al. (2007), the profile crosses the triple point of the European plate, Adriatic microplate and Pannonian fragment (Fig. 11b). The tectonic interpretation of the seismic model emphasizes structures and processes relevant for the Alpine orogenesis. Because of the orientation and location of the profile, crustal structures mainly correspond to a classic collision orogen (e.g. Moores & Twiss 1995; Hatcher & Williams 1986), but are also affected by extrusion and escape tectonics (e.g. Ratschbacher et al. 1991). Down to a depth of 5 km, the tectonic model was constrained by the geological profile (Fig. 1), while the interpretation of deeper structures relies on first-order discontinuities and significant velocity anomalies of both the 2-D and 3-D seismic models. The intersection of CEL10/Alp04 with the profiles Alp01 and Alp02 at the PAL (Brückl et al. 2007) and with profile CEL09 (Hrubcová. 2005) in the Bohemian Massif provides further constraints.

What follows is a detailed description of the tectonic model (Fig. 12), starting with the upper crust. We discriminate between new information from CEL10/Alp04, already existing information from the 3-D seismic model and intersecting profiles, and rather speculative interpretive features. A low-velocity zone dipping under NCA and Greywacke Zone is resolved by the new seismic data (Fig. 6). A comparison with the geological profile shows that the lower boundary of the low-velocity zone coincides with the crystalline basement of the Bohemian Massif, dipping below Molasse and overthrusted by the Alpine accretionary wedge. The horizontal reflector around a profile distance of 150 km at 10 km depth further delineates the decollement of Austroalpine nappes. We constrain the extent of the upper European crust to the south by a long and continuous velocity structure dipping from 7 km depth below the Bohemian Massif at a profile distance of 300 km to 15 km below the Eastern Alps. Back-thrusting of the upper Adriatic crust has taken place along the SAT. We may hypothetically extend the thrust down to the southern termination of upper European crust, thus following the scheme of a bivergent orogen (Hatcher & Williams 1986). The high-reflectivity (HR) zone in the middle crust around PAL, an area of crustal thickening, has been correlated with the indenting Adriatic crust. The genesis of this reflective zone will be discussed later in more detail. The elongated HV body at about 20 km depth within the European crust may either be an ophiolite, a felsic-mafic, or a mafic intrusion. The Moldanubian-Moravian thrust is very close to and subparallel to the CEL10/Alp04 profile in this area and ophiolite emplacement or magmatic activity is documented close to this suture. An HV middle crust was also found within the Bohemian Massif along profiles Alp01 (Brückl et al. 2007) and CEL09 (Hrubcová. 2005).

Figure 12.

Tectonic interpretation of the CEL10/Alp04 profile based on the seismic model (Fig. 4a) with geological cross-section from Fig. 1 (down to 5 km depth). Solid lines show seismic boundary elements or significant velocity structures mapped by reflected and/or refracted waves; chessboard areas are HR or HV zones. Bold dashed lines separate European and Adriatic plates. PAN marks Pannonian fragment. S-TR—Sub-Tauern ramp.

The European Moho has been resolved by Pn and PmP phases from profile distance 150 km to about 320 km. It dips slightly to the south from 39 km at a profile distance of 300 km to 43 km at a profile distance of 150 km. The overthrusting Eastern Alpine nappe stack causes this overall crustal thickening. However, the European crust, represented by the crystalline basement, thins out from 40 km in the north (profile distance 300 km) to 32 km in the south (profile distance 170 km). A similar observation was made along Alp01 (Brückl et al. 2007) and we again attribute it to an extensional phase of the European platform prior to the Alpine orogenesis (Roeder 1977).

Post-collisional escape and crustal extension generated the Pannonian fragment and lead to a Moho uplift (Behm et al. 2007; Brückl et al. 2007). Along CEL10/Alp04 at profile distance 100 km, the westernmost part of the Pannonian Moho is sampled by PmP reflections from the southwestern side. The Adriatic Moho is not revealed by seismic data along profile CEL10/Alp04. However, this area is well resolved by profiles Alp01 and Alp02 (Brückl et al. 2007). We introduce the Moho topography and the Sub-Tauern ramp (S-TR; TRANSALP Working Group 2002; Lüschen et al. 2006) from this tectonic interpretation into our model. In our interpretation of CEL10/Alp04, the S-TR does not continue up to the surface, because the SW-NE striking profile leaves the domain of the Tauern Window at a profile distance of about 100 km.

We mapped a laminated zone with HR in the middle crust and interpreted it as the indenting wedge of the Adriatic plate. The comparison of the recordings of SP31140 along profiles Alp01, Alp02 and Alp04 (Fig. 7) shows that a reflective zone is only observed along the Alp04 profile to NE. Along the profiles Alp01 and Alp02, only single floating reflectors were interpreted (Brückl et al. 2007). The HR zone is located in a very specific tectonic setting between the Tauern Window in the west and the emerging Pannonian fragment in the east, a circumstance which may be in related to its origin.

One cause of HR could be lamination due to ductile extensional deformation as generally observed in the lower crust by steep-angle reflection seismic profiles (Meissner & Kusznir 1987; Meissner et al. 2006). However, in our case the HR zone is restricted to the middle crust and was not observed along profile Alp02 further to the east, where the crust had been strongly influenced by extension due to tectonic escape (Brückl et al. 2007). Another explanation could be intermediate to mafic intrusions. These intrusions could have been triggered by a temperature pulse caused by the roughly contemporaneous vertical extrusion of the Tauern Window by more than 25 km (Fügenschuh et al. 1977) in the west, or the Moho uplift of the Pannonian fragment relative to the European and Adriatic plates in the east.

The suture between Adria and Europe has been drawn in Fig. 12 partly on the basis of the geologic profile and our new seismic data, partly hypothetically or by the integration of evidence from intersecting profiles, especially Alp01. The Molasse and Flysch belts separate Adria and Europe at the North Alpine front. We cannot decide from our data if overhrusted Flysch wedges out under the NCA, or if ophiolites (Penninicum) exist on top of the decollement at about 10 km depth and around profile distance 150 km. These units are exposed in the TW which has its eastern termination about 50–80 km west of profile CEL10/Alp04. The suture at mid-crustal level represents a crocodile structure with the HR zone (HR in Fig. 12) interpreted as the indenting Adriatic wedge. In the lower crust, we are very near to the intersecting Alp01 profile and south of the eastern termination of the TW. Here, we introduced the S-TR representing the suture zone in the lower crust. The S-TR is a major structure over which upfolding and extrusion of the TW took place (e.g. TRANSALP Working Group 2002; Lüschen et al. 2004). However, neither this structure nor the Adriatic Moho is constrained by the CEL10/Alp01 data. Therefore, we put a question mark to zone of our tectonic model in Fig. 12.

Conclusions

In general, the interpretation of the seismic data along profile CEL10/Alp04 agrees well with structures revealed by a 3-D model (Behm et al. 2007). The Moho depths derived at profile CEL10/Alp04 using 2-D ray-tracing technique are a good match to the Moho depths of the 3-D model (–1.1 ± 0.8 km) and are in accordance with the conception of a Pannonian fragment. It benefits from the information gained by the profiles Alp01 and Alp02 at their crossing points with Alp04 near PAL. Adriatic Moho could not be modelled by CEL/Alp04 data. Therefore, our new model cannot contribute to a clarification, if in the area of the TW European lithosphere was subducted below Adriatic lithosphere, or if the reverse is true (e.g. Kissling et al. 2006). The decollement of the Austroalpine nappe stack over Flysch, Molasse and European basement has been revealed down to a maximum depth of 10 km. Thinning of European basement crust from the Bohemian Massif to the south below Austroalpine nappes is in accordance with similar findings along Alp01 (Brückl et al. 2007). The suture between the European plate and the Adriatic microplate was interpreted on the basis of a 170-km long, coherent upper crustal velocity structure in the European crust and an HR zone in the Adriatic crust. An HV body in the middle European crust and an HV lower crust under the Bohemian massif correlate with an HV zone derived from the 3-D seismic model. The reflector in the middle crust may be related to Moldanubian-Moravian overthrusting during the Variscian orogenesis. The HV lower crust extends from CEL10/Alp04 to the Vienna basin and the northwestern Pannonian basin. The most likely an explanation of its generation has to be given within the context of the roll back of the Carpathian arc and the generation of the Vienna and Pannonian basins. However, this is beyond the scope of this study.

Acknowledgments

The CELEBRATION 2000 and ALP 2002 experiments were made possible by the scientific and financial contribution from 15 countries: Austria, Belarus, Canada, Croatia, Czech Republic, Denmark, Finland, Germany, Hungary, Poland, Russia, Slovakia, Slovenia, Turkey, USA. It is only through the participation of almost 1000 individuals that the preparation, field work, data acquisition and processing in both experiments could be successfully completed. We further extend our gratitude to Johanna Brückl for compiling a geological map and cross-sections. The authors are grateful to two anonymous reviewers for helpful and exhaustive comments.

Author notes

*

CELEBRATION 2000 and ALP 2002 Working Groups comprise: S. Acevedo, K. Aric, I. Asudeh, M. Behm, A.A. Belinsky, F. Bleibinhaus, T. Bodoky, M. Brož, E. Brückl, W. Chwatal, R. Clowes, W. Czuba, T. Fancsik, M. Fort, E. Gaczyński, H. Gebrande, A. Gosar, M. Grad, H. Grassl, A. Guterch, Z. Hajnal, S. Harder, E. Hegedűs, S. Hock, V. Höck, P. Hrubcová, T. Janik, G. Jentzsch, P. Joergensen, A. Kabas, G. Kaip, G.R. Keller, F. Kohlbeck, K. Komminaho, S.L. Kostiuchenko, A. Kovacs, D. Kracke, W. Loderer, K.C. Miller, A.F. Morozov, J. Oreskovic, K. Posgay, E.-M. Rumpfhuber, Ch. Schmid, R. Schmöller, O. Selvi, C. Snelson, A. Špičák, P. Środa, F. Šumanovac, E. Takács, H. Thybo, T. Tiira, Č. Tomek, Ch. Ullrich, J. Vozár, F. Weber, M. Wilde-Piórko, J. Yliniemi.