Southern Rhinegraben Research Articles

Relocation of the 5 December 2004 Waldkirch, Germany, Earthquake with Regional 1D and 3D Seismic Velocity Models in the Presence of Upper-Mantle Anisotropy

Abstract The exact hypocenter location, that is, the depth of the 5 December 2004 Waldkirch earthquake in the southern Rhinegraben area, has been at the center of debate for some time. After various relocation efforts by several geophysical institutions, a depth range of 9–12 km was eventually reported for this event. Here we examine whether the Waldkirch event can be located with improved accuracy by a regional network with newly derived regional 1D and 3D seismic velocity models for the crust and upper mantle that also take into account the existing Pn upper‐mantle anisotropy beneath Germany. The preliminary arrival‐time analysis of the phase data hints at possible misidentifications of several phases ( Pg versus Pn ), in addition to that of errors due to anisotropy that lead to unrealistic shallow depth locations in the range of 0–5 km. After removal of these adverse effects on the relocation, the more realistic depth range of 10–15 km is obtained. The effects of an anisotropic upper mantle, the influence of the V P / V S ratio, and the influence of a varying Moho depth on the event relocation are then investigated. With the 1D model, the Waldkirch depth is located between 14 and 15 km, when P waves only or both P and S waves are used, whereas with the 3D P ‐wave velocity model, the depth again becomes 14 km. The effect of the Pn anisotropy is stronger for the epicenter than for the depth. The statistical location precision is determined by calculation of covariances and confidence ellipses as well as by Monte Carlo simulations with both randomly perturbed arrival times and initial hypocenters. In addition, the effects of uncertainties in the velocity model on the relocation are investigated. Based on this comprehensive analysis, the optimal Waldkirch event depth is determined to be 14 km, with a 95% confidence interval of ±2.5 km. Online Material: Tables of original phase data and statistical results with various relocations, and figures showing effects of phase corrections, 3D tomographic and anisotropic models, and results of Monte Carlo relocations.

100 years Jura décollement hypothesis: how it affects Steinmann’s (1892) “Schwarzwaldlinie”

Steinmann, then professor of geology at Freiburg (Germany), more than a 100 years ago wondered about the southern end of the extensional Rhinegraben and proposed that elements of the graben penetrated the contractional Jura. In particular, he recognized the "Schwarzwaldlinie” in the southern prolongation of the eastern border of the southern Rhinegraben, a line-up of topographic as well as structural irregularities. He conjectured that it was caused by normal faults of the Rhinegraben system. Subsequently—100 years ago—Buxtorf (1907) proposed the hypothesis, that the Jura was a thin-skinned nappe sheared off on Triassic evaporites. In the autochthonous basement underneath the wrinkled skin, the ``Schwarzwald line” is difficult to define. It probably consists of a gentle flexure punctuated by faults that approximately coincides with Steinmann’s original projection, although he sought to identify its constituent faults in the badly deformed allochthonous skin. Current data place the thin-skin elements of the Schwarzwald line in a more westerly, allochthonous position where most of them were reactivated into sinistrally transpressional structures.

2-D anisotropicPn-velocity tomography underneath Germany using regional traveltimes

This paper, complementing the recent 1-D Pn time-term analysis of Song et al. (2001a), presents tomographic inversions of regional Pn traveltimes to image lateral variations of seismic velocity and anisotropy within the uppermost mantle beneath Germany. As such our study fills a gap of previous regional 2-D anisotropic Pn studies for France and Italy by expanding anisotropic Pn tomography to a continental scale covering much of Europe. In a novel way, the resolving power of the used Pn tomography system is investigated through a series of checkerboard tests combined with tradeoff analyses between resolution and covariance of the models. Methodologically important and commonly interesting problems of seismic tomography are addressed with the data set, such as model parametrization for a given source-receiver configuration, effects of time terms on the model parameters, and ambiguity between lateral velocity perturbations and anisotropy. The reliability and significance of various inversion models (1-D, 2-D, isotropic, anisotropic) is investigated in detail using statistical F tests. The latter indicate that, while both 2-D isotropic and anisotropic models provide better fits of the Pn data than a standard homogeneous, isotropic upper mantle, the most significant improvement arises from a large-scale 1-D anisotropy ellipsoid with the fast Pn-velocity axis in 25°NE direction. This result correlates well with a number of trends related to upper mantle and plate dynamics, for example, major plate motion. Besides this anisotropic dominant traveltime pattern, we observe a number of lateral velocity anomalies. Lower than average velocities were detected in the southwestern Rhenish massif (but not in the area of the Eifel plume), similar to findings for the adjacent region in France. In addition, lower velocities are seen in parts of the Alps and in a region along the upper Main river, south of the Rhön. However, the most striking Pn anomalies are associated with prominent high velocity anomalies in the Ingolstadt region and an area centred between the Harz and the Vogelsberg, where perturbations of up to +5 per cent are found. As a possible mechanism for these features we consider underplating within ancient plume activity. In addition, results for the time-term station delays are consistent with findings of crustal structure in other studies with different methodologies and data, and indicate, for example, crustal updoming of the Moho in the southern Rhine graben around the Kaiserstuhl area, but crustal thickening towards the Alps.

Plio-Pleistocene folding in the southern Rhinegraben recorded by the evolution of the drainage network (Sundgau area; northwestern Switzerland and France)

The southern Rhinegraben represents a key area to unravel the Neogene paleohydrographic history of the northern Alpine foreland. At least, three successive main drainage systems are documented by the youngest sediments of the graben fill. They are respectively pre-, syn- and post-dating the folding of the Jura Mountains (about 10-5 Ma ago). Pliocene to recent uplift and shortening of the Pliocene fluvial gravel (“Sundgauschotter”), which accumulated on a nearly plain surface, and the progressive deflection and capture of rivers resulted from reactivation of pre-existing Paleogene faults. The alluvial terraces along the major rivers of the Sundgau area show upward warping near the uplift areas. Furthermore, river beds responded to uplift by lateral shift. During the Pliocene-Pleistocene, fluvial drainage diverted stepwise from the initial, westwards directed course (towards the Bressegraben) to the N (into the Rhinegraben): Firstly, uplift in the south at the boundary between the Rhinegraben and the Jura Mountains induced a northwards directed migration of the middle Pliocene, E-W flowing paleo-drainage system. Successively, a late Pliocene/early Pleistocene lowering of the base level in the Rhinegraben north of Mulhouse resulted in the capture by southwards directed backward erosion of this system into the Rhinegraben, as likewise documented by the reconstitution of the paleo-drainage patterns in the Sundgau area. In addition, a gradually capture moving from NE to SW has been identified for the Sundgau catchment area.

The southern Rhine graben: A new view of the initial phase

One of the puzzling features of the southern end of the Rhine graben is the Dinkelberg-Tabular Jura block on the eastern shoulder of the graben. It is dissected by a large number of faults, the most notable ones forming a field of narrow little grabens and half-grabens whose bordering faults converge at the level of the Middle Triassic evaporites, which points to decollement at that horizon. The little grabens were traditionally considered to be of Oligocene age, coeval with the main taphrogenesis of the Rhine graben. Two hypotheses were offered for their formation, one ascribing them to extension on the extrados of large basement folds, the other to gravity sliding on paleoslopes. Recent field work uncovered overwhelming evidence for an Eocene age of the little grabens, the time of the initial phase of Rhine graben formation. At that time there were neither large basement folds nor paleoslopes of any significance, and therefore the two hypotheses offered until now do not work. However, the map-view pattern of the field of faults offers a somewhat unusual way out of the dilemma. This pattern is most prominently displayed in the Dinkelberg area north of the Rhine. There a lane of narrow decollement grabens with a mean NNE strike is confined within the NW- striking Dinkelberg graben, which is much wider and rooted in the basement. It is also very shallow, with a subsidence on the order of 100 m. The lane of decollement grabens forms a dextral en-echelon pattern with respect to the Dinkelberg graben, suggesting stretching of the post-evaporite sequence above a basement essentially extended by strike slip. This model, though not as clearly expressed, is also compatible with the data in the rest of the Dinkelberg-Tabular Jura block. It also fits surprisingly well a theoretical model by Withjack and Scheiner (1982) that predicts a dominance of strike-slip in the marginal area of a system consisting of extension superimposed on doming.

Plio-Pleistocene transpressional reactivation of Paleozoic and Paleogene structures in the Rhine-Bresse transform zone (northern Switzerland and eastern France)

Pliocene to recent uplift and shortening in the southern Rhinegraben is documented by deformation of Pliocene fluvial gravels, deposited on a nearly planar surface, as well as by progressive deflection and capture of rivers. This deformation is suggested to result from thick-skinned tectonic movements as evidenced by ob- servations on seismic records, which demonstrate a spatial coincidence between en-�chelon anticlines at the surface and faults located in the crystalline basement. These findings contradict the often invoked thin-skinned tectonism in the recent tectonic history of the Rhine- graben. In particular the transfer zone between the Rhinegraben and the Bressegraben is very suitable for reactivation under the present day stress field. Thick- skinned reactivation of faults in the basement is also expressed by focal plane mechanisms of recent earth- quakes showing strike-slip- rather than reverse faulting characteristics. This is of importance for the densely populated and industrialised southern Rhinegraben, pre- viously affected by large earthquakes in historical times (e.g. Basel 1356).

New insight in the tectonics of the southern Rhine Graben‐Jura region using local earthquake seismology

A new seismic data set was acquired using a dense temporary network of mobile seismic stations. It is combined with existing seismological data for the southern Rhine Graben. The data yielded some new focal mechanisms and allowed for a detailed local earthquake tomography. These methods give new insight in the structure of the crust and stress conditions in the area. By combining the tomography with the earthquake hypocenter locations and focal mechanisms, some of the structures in the P wave velocity model are interpreted as active faults. The eastern boundary fault zone of the graben is found to be activated in a left‐lateral strike‐slip sense. The southern continuation of these faults is formed by Variscan structures in the basement. A change in orientation of the graben axis between the main graben and its southern end have played a role during the development of the system. This difference in orientation caused the development of the Dannemarie and Allschwil subbasins during a late stage of the opening of the Rhine Graben. This study shows evidence for active faulting in the basement underneath the detachment of the Jura Mountains. These faults, which are interpreted as reactivated Permo‐Carboniferous troughs, coincide with the frontal thrusts of the Jura Mountains. It might be that these structures in the basement have influence on the detachment tectonics in the region.

A contribution to the middle Oligocene paleogeography of central Europe: New evidence from fission track ages of the southern Rhine-Graben

A contribution to the middle Oligocene paleogeography of central Europe: New evidence from fission track ages of the southern Rhine-Graben

Seismicity pattern and style of seismic faulting at the eastern borderfault of the southern Rhine Graben

Seismicity recorded on high-gain, digital networks in the southern Rhine Graben reveals that the area around the eastern borderfault system shows the highest level of seismic activity. Across this fault system remarkable differences of maximum focal depths exist: 12–13 km are characteristic within the graben, and 18–20 km beneath the western foothills of the southern Black Forest, located approximately between the eastern borderfault of the graben and the river Rhine (inner borderfault). Farther to the east, in the southern Black Forest proper, maximum depths around 20–22 km are found. At particular sites a few earthquakes with depths of up to 26 km, close to the Moho, were detected. The deep cut-off of seismic activity in the southern Black Forest and its western foothills is persistent for about 100 km distance between the Freiburg focal area in the north and the Swiss Jura in the south. Seismic reflection profiling indicates that the deep foci lie well within a reflective lower crust (southern Black Forst, Swiss Jura) or just above a highly reflective top of the lower crust (western foothills of the southern Black Forest, Dinkelberg block). Within the graben proper, seismic activity is restricted to the upper crust and ceases about 5 km above the top of the lower crust.Low seismic activity is found in areas of low P-wave velocities as in the upper crust of the central Black Forest, north of the Freiburg focal area, and at midcrustal levels in the southern Black Forest (at depths of about ±10 km).Well constrained focal mechanisms are presented for 47 events. Most of them occurred adjacent to the eastern borderfault. Nearly all mechanisms show an oblique component of slip. Strike-slip and normal-faulting mechanisms prevail. Only a few events are of thrust type. Dominance of a certain faulting style with depth was not found. Focal mechanisms of events which occur within 10 km of the borderfault show variations of faulting style, indicating local deviations from the average regional stress tensor and/or from the frictional parameters in the proximity of the particular rupture planes. Although the nodal planes correlate with the strike directions of surface trends of major faults in general, it is difficult to recognize their link to individual faults.

Magmatic expression of lithospheric thinning across continental rifts

Abstract Studies of magmatism associated with continental rifting have traditionally focused only on volcanism within the downfaulted axial zone and along its immediate flanks. Teleseismic travel-time delay studies during the last decade have confirmed the results of earlier gravity surveys of rifted areas, showing that thinning at the base of the continental lithosphere occurs throughout a zone up to about 10 times wider than the physiographic expression of the rift. It is, therefore, logical to consider rifting-related magmatism on the same scale. Potential sources of mafic magmas in rift zones are the thinned subcontinental lithospheric mantle (SCLM), the convecting mantle beneath the continental plate and mixtures of the two. Detailed elemental and radiogenic isotope geochemical studies show that, during the initial extension of continental rifts, the associated mafic magmatism tends to be: (1) relatively sodic and from predominantly convecting mantle sources at the rift axis; (2) relatively potassic and from predominantly lithospheric mantle sources at the margins of the thinned-plate zone. This underlying geochemical pattern is obscured in many instances by such processes as crustal contamination and magma mixing within open-system reservoirs. The mafic ultrapotassic component that provides a distinctive input to SCLM-source magmas appears to be largely fusible at temperatures well below the dry solidus of SCLM; so that, in some cases, prolonged magmatism at a site causes removal of most or all of the potassic lithosphere-source melt (as mafic ultrapotassic magmas or as a contribution to mixed-source melts) without destruction of that lithosphere segment as a geophysically defined unit. Such a zone of refractory lithosphere permits subsequent, recognisable, convecting mantle source melts to penetrate it and reach the surface. These principles are illustrated by discussion of the Neogene-Quaternary magmatism of the Rio Grande, East African, Rhine and Baikal rifts, in the context of the most recent published models of their geophysical structures to depths > 200 km. Teleseismic and gravity studies identify lithospheric thinning beneath the Rio Grande, East African and Baikal rifts across zones 700–800 km wide. The failure of the southern Rhine graben to show a similar deep seismic structure may be a result of efficient buoyant migration of low-viscosity mafic alkalic melt out of the underlying mantle during the 7 Ma period since magmatism ceased, causing seismically defined asthenosphere to revert to lithosphere. A 700 km geochemical traverse across the Rio Grande rift at ~ 37°N, focusing on Oligocene-Miocene magmatism minimally affected by post-genesis processes, shows a clear symmetrical pattern of relatively sodic volcanics at the rift axis and mafic, ultrapotassic magmatism on its outer flanks. The geochemistry of these contrasting magma types is consistent with the view that they originated predominantly within the convecting mantle and SCLM, respectively. The same geochemical pattern is detectable in the volcanism within the equatorial segment of the East African rift system but it is complicated in two zones: east of the Gregory rift and southwest of Lake Kivu, by the effects of previous Cretaceous-Palaeogene magmatism. Limited, published, appropriate, geochemical data show that regional compositional variation in the volcanics associated with the Baikal rift appear to fit the pattern proposed here as a general model for rifting-related magmatism.

Bootstrap inversion for vertical and lateral variations of the S wave structure and the υ p/υ s-ratio from shallow earthquakes in the Rhinegraben seismic zone, Germany

Using S wave arrival times of ~ 400 earthquakes in the southern Rhinegraben, 1-D and 3-D inversions for the shear (S)-wave crustal structure of that region are performed. These models are then combined with previously obtained 1-D and 3-D P wave models to compute vertical and lateral variations of the υ p / υ s (Poisson's)-ratio. Such an independent P and S wave inversion approach circumvents many of the limitations of the classical Wadati-diagram method usually used for the determination of crustal υ p/υ s-ratios. For the statistical evaluation of the 1-D models we have used a new Monte-Carlo bootstrap resampling technique which has been found useful in many statistical estimation problems where the underlying statistics of the data is poorly known. The 1-D bootstrap υ p / υ s crustal models exhibit significant vertical variations. Whereas the upper sedimentary and crystalline layer of the crust is characterized by a significant increase of υ p / υ s over the reference value, υ p / υ s in the upper and middle crust is lowered. This correlates well with several findings from Wadati-diagrams for υ p / υ s in other seismically active zones and might be explained through the gradual closing of saturated microcracks in that depth range. However, the υ p/υ s-reduction can also be interpreted in terms of a particular petrological composition such as orthoclase feldspar without the indispensable need for fluids. A lowering of υ p / gu s is also observed for the lower crust, which might reflect the strong lamination found here from previous deep seismic sounding experiments. For the upper mantle a strong increase of υ p / υ s is observed which is primarily triggered by a decrease of υ s , and is best explained by an eclogite petrological assemblage. For the 3-D S wave models significant reductions of more than 100% were obtained for the unexplained travel-time residuals. Variations of υ p / υ s in the upper and middle crust are mostly induced by changes in υ p rather than υ s . For the upper crust slight decreases of υ p / υ s are observed for the Rhinegraben, but increases are detected for the adjacent Vosges and Black Forest areas. In the middle crust the refraction-seismic deduced reduction of υ p / υ s under parts of the Black Forest can neither be corroborated nor rejected. However, because of some experimental ambiguity on the effects of petrological composition and interstitial pore fluids, this does not rule out the presence of the latter. In the lower crust a local reduction of υ p / υ s is observed under the Kaiserstuhl and graben-Black Forest shoulder and has been explained in terms of a metamorphic, dehydrating phase change of the hydrous amphibolite facies to the anhydrous granulite facies. Depending on the pore-pressure of the dehydrated water and the fractional amount of interstitial CO 2, granulite facies metamorphism occurs in the 650–800°C temperature range. Temperatures at the upper end of this interval can be generated by the mechanism of accretionary magmatic underplating as it has been proposed for some extensional rift zones where asthenospheric upwelling, as in the Rhinegraben, occurs.

Quantification of surface faulting potential in a low to moderate active region: an example from the southern Rhinegraben area : Loew, S; Jenni, J P; Blanc, B Int Assoc Engng Geol BullN40, Oct 1989, P111–117

Quantification of surface faulting potential in a low to moderate active region: an example from the southern Rhinegraben area : Loew, S; Jenni, J P; Blanc, B Int Assoc Engng Geol BullN40, Oct 1989, P111–117

Finite element modelling of a continental rift structure (Rhinegraben) with a large-deformations algorithm

The formation of continental rifts with small extension values can be simulated over several tens of millions of years by means of a finite element program able to handle large deformations. As an example an attempt has been made to model the formation of the northern and southern Rhinegraben over the past 40 m.y. Deviatoric extensional stresses of 25–40 MPa were required and, to model the Moho uplift, mass anomalies in the lower lithosphere had to be invoked with density contrasts of −0.15 and −0.30 g/ cm 3 for the northern and southern Rhinegraben respectively. Trial models without a mass anomaly in the lower lithosphere resulted in crustal thickening beneath the graben and not crustal thinning. In the northern Rhinegraben the rifting is more passive and uniform and model subsidence values of 4000 m were obtained together with model shoulder and Moho uplifts of 800 m and 1500 m respectively. In the southern Rhinegraben the rifting is more active and episodic and, in fact, between 13 and 10 m.y. ago the model graben floor was uplifted in order to simulate the observed graben floor uplift at around 10 m.y. ago. Model subsidence values of 2000 m were obtained while the model shoulder and Moho uplifts were 1900 m and 7000 m respectively in good agreement with the observed values.

Steady state geothermal model of the crust and the problem of the boundary conditions: application to a rift system, the southern Rhinegraben

Seismic, gravity and magnetotelluric data are used to model the two-dimensional heat transfer beneath an E-W vertical cross section ranging from the Black Forest, Rhinegraben to the southern Vosges. The temperature field at the present time has been calculated taking into account the effect of temperature on thermal conductivity and the radiogenic heat sources. Heat flow density conditions at depth, computed by solving iteratively the heat equation and comparing the heat flow measured at the surface with the calculated one, have been used. As the resulting heat flow condition at depth is very sensitive to the interpolation method used to estimate the heat flow values at the surface, a confidence interval of the thermal field is then proposed by varying the possible mantle heat flow conditions using a Monte Carlo method. Arguments based on petrological considerations suggest that a temperature condition could be used at the Moho. The resulting model involves regional surface heat flow densities ranging from a normal crustal value of 70 mW m −2 to a highest value of 105 mW m −2, in agreement with actual measurements. The results are compatible with those obtained by the German KTB Program on the Black Forest (Gehlen et al., 1986) and on the Swiss traverse (Rybach, 1979). They suggest that the regional high heat flow (105 mW m −2) measured in the southern Vosges is caused by the high radiogenic elements content (especially thorium) of the Crêtes sub-alkaline granite and by an asymmetry in the lithospheric heating. High heat flow density values observed along the Rhinegraben give evidence of heat refraction phenomena around high conductivity rocks.

The Rhinegraben: Extension, subsidence and shoulder uplift

We apply a numerical model of the thermal evolution of a continental lithosphere undergoing extensional deformation to the Rhinegraben. Using reasonable parameters, the results of computations are in agreement with the present configuration of the rift and of its shoulders, as well as with their evolution since the Oligocene. Contrasting evolutions are reconstructed in the northern and southern Rhinegraben: whereas the Northern history is accounted for by a simple model of homogeneous stretching of the lithosphere, the history of the southern Rhinegraben cannot be explained without introducing two main periods of heterogeneous stretching. The results are locally supported by structural analyses of extensional patterns. Factors of extension average 1.2 in the axial portion of the Rhinegraben and 1.15 for the whole rift section, suggesting that the total amount of E-W extension reaches 6 km.

Deep crustal reflections from a Vibroseis survey in northern Switzerland

In 1982 a Vibroseis ® survey comprising 180 km of reflection profiles was run in northern Switzerland in order to investigate the suitability of the crystalline basement for the deposition of highly radioactive waste. A configuration was chosen with 144 channels, 25 m of geophone spacing, 20 s sweeps ranging from 11 to 61 Hz and stacking of 4 or 8 sweeps of 3 simultaneous vibrators at twice the geophone spacing. The listening time was generally 4 s and at 4 sites it was extended to 11s for the detection of deeper crustal reflectors. This survey unravelled the complicated fault and thrust system beneath the Swiss folded Jura mountains. The stack from 4 s to 11 s reveals clearly a strong sloping reflector between 3.0 and 3.5 s which is strong evidence for a pronounced differentiation in the upper crust. A series of reflections is observed between 5.8 and 7.2 s the top of which can be correlated with the Conrad discontinuity. A strong “layered” signal between 9.0 and 9.5 s is interpreted as reflections from the M-discontinuity. The main features are compatible with results from nearby refraction surveys in the southern Rhinegraben rift system which show a distinct velocity increase of about 0.5 km/s in the lower crust at a depth ranging from 15 to 20 km, followed by an inversion zone or a laminated structure before reaching the Moho at about 27 km depth. The correlation of the field recordings with the first 10 s of the up-sweep only, shows some loss of resolution in the uppermost 3 s because of the lower frequency content of the signal. However, the lower parts of the sections are nearly identical. The fact that the deeper reflectors in the sections can consistently be traced laterally is a strong argument for using this processing technique. Thus high-coverage Vibroseis surveys utilizing up-sweep can be processed for deep crustal reflections even if the recording time is restricted to the standard 4 s, provided the surface static corrections are carried out with high precision.

A review of the 1356 Basel earthquake: Basic data

Abstract The effects of the Basel earthquake on October 18,1356,22 hours local time, have been studied by collecting all available information in Switzerland, France and Germany. As far as possible the original sources such as chronicles, repair-bills and other reports were evaluated. By taking into account the type of construction in the area, MSK-intensities could be assigned in many cases. The macroseismic field shows, as a striking feature, a distinct extension of the isoseismals towards West-Southwest. In the epicentral zone, the maximum intensity is equal or greater than IX (MSK). The most probable location for the epi-center is about 10 km south of Basel, on the eastern main fault system of the southern Rhinegraben. The estimated focal depth is 10–20 km.