T34A-01 INVITED
Observations of Extremely Thin Continental Crust: Paradoxes and Implications
One of the most enigmatic problems in extensional tectonics is to understand the mechanics of thinning crust and lithosphere to an extreme amount. In magma starved settings, where extensional stresses can only be relieved by mechanical thinning, crust must be thinned from typically 30 km or more to only a few km thick before mantle exhumation takes place. Off the Galicia Bank, the zone of thinned crust consists of small half graben basins that appear to sit directly atop upper mantle. The existence of apparently only upper crustal rocks leads to imporant questions regarding what happens to mid and lower crust during such extreme extension. Along the Flemish Cap margin and conjugate to Galicia Bank, the thinning phase of breakup spans a very narrow region and is puzzling due to the lack of seismic evidence for signicant normal faulting in the uppermost crust. The region along the slope is nearly devoid of visible structure in seismic sections and the thinnest observed crust farther seaward shows only small throw normal faults and very little rotation. We summarize several competing hypotheses for the formation of these thin crust regions. In the absense of direct sampling, the available data may be unable to distinguish between these different models.
T34A-02
From Crustal Thinning to Continental Break-up in Magma-Poor Rifted Margins: How Important are Detachment Faults?
Detachment faults are widely regarded as playing an important role in crustal thinning and mantle exhumation at magma-poor rifted margins. However, how important are these structures and what is their role in crustal thinning and mantle exhumation? In our presentation, we review the pertinent observations made along the Iberia-Newfoundland, Western Pyrenees-Bay of Biscay and the Alpine margins. A reconstruction of the future distal margins 10 to 20 myr before continental break-up shows that: (1) extension was localized and the crust was thinned to less than 10 km despite subdued high-angle normal faulting, (2) detachment faults responsible for mantle unroofing are late and shallow crustal structures, (3) the mantle lithosphere was locally omitted and replaced by infiltrated asthenospheric mantle, and (4) shallow marine sediments were deposited over thinned crust suggesting a retardation of tectonic subsidence. These observations scrutinize our present knowledge of rift processes and ask for a paradigm shift in the way to interpret rifted margins. To further investigate the role of detachment faulting during crustal thinning and mantle exhumation, we developed a numerical model that: (1) uses initial conditions, strain distribution, rheology and deformation modes constrained from the study of Alpine margins, and (2) is able to replicate the well-documented tectonic evolution of the Iberia and Alpine margins, starting with pure-shear dominated crustal stretching and ending with mantle exhumation along detachment faults. The most interesting result of the experiment is the way the model evolves from the prescribed initial distributed stretching (e.g. pure-shear) to final localized exhumation (e.g. simple shear). During this transitional phase, referred to as thinning phase, extension is accommodated by a system of superimposed but decoupled concave-downward faults that simultaneously exhume middle crust to the seafloor and deeper mantle at the base of the crust. The two rolling hinges eventually merge to form one concave downward fault that unroofs the mantle to the seafloor. These results suggest that exhumation associated with detachment faulting is an important process. It can explain crustal thinning to less than 10 km; the lack of major fault bounded topography; and the emplacement of deep and infiltrated mantle rocks beneath thinned crust 10 to 20 myr before continental break-up. This has important consequences for the isostatic, thermal and rheological evolution of deep magma-poor margins that need to be constrained by data. At present, the thinning mode is neither understood nor constrained by geological or geophysical data. Because the proximal and distal margins are mainly preserving structures of the initial stretching and the final exhumation modes, we started to investigate the transition zone between the proximal and distal margins in the Alps and the zone ahead of a propagating ocean in the Western Pyrenees – Bay of Biscay. In our presentation, we present evidence for the existence of the thinning mode within these two domains and show how these results help to develop new conceptual ideas to re-interpret seismic data from present-day magma-poor rifted margins.
T34A-03
Mantle Exhumation at Magma-Poor Rifted Margins due to Melt Suppression During Continental Break-up and Seafloor Spreading Initiation.
Rifted continental margins exhibit large variations in magmatic activity. Non-volcanic margins may display a broad ocean-continent transition, up to 150 km wide, of exhumed mantle separating oceanic crust from thinned continental crust, whilst voluminous volcanism accompanies break-up at volcanic margins. Previous studies have shown the importance of asthenospheric temperature (White and McKenzie, 1989), lithosphere thinning rate (Bown and White, 1995; Pérez-Gussinyé et al., 2006) and initial continental geotherm (Reston and Morgan, 2004) on melt production at rifted margins, assuming that continental break-up occurs by pure-shear stretching of the lithosphere. Pure-shear models of continental lithosphere thinning generally predict melt generation before continental break-up, unless anomalously cool asthenosphere temperatures (Minshull et al., 2001) or depleted mantle source (Pérez-Gussinyé et al., 2006) are invoked. As a consequence pure-shear models have difficulty explaining mantle exhumation at non-volcanic margins. We model the onset and development of melt production during rifting of continental margins and seafloor spreading initiation using a model of depth- dependent lithospheric thinning and extension. For a mantle potential temperature of 1333°C and a half- spreading rate of 10mm/yr we predict that approximately 100km of lower crust and mantle is exhumed prior to melt production. Melt production reaches a steady state as the model reaches thermal equilibrium and seafloor spreading proceeds. The predicted width of exhumed mantle increases if the modelled spreading rate is decreased, if melt is retained in the mantle (e.g. as gabbroic intrusions), or if the subcontinental mantle is initially cool or depleted. In this model the thinning of continental lithosphere leading to break-up and sea-floor spreading initiation occurs in response to an upwelling and divergent flow-field. The melt parameterisations of Katz et al. (2003) are used to calculate melt fractions in the model space. The ability of the model to predict mantle exhumation prior to the onset of volcanism and normal seafloor spreading suggests that an upwelling and divergent flow-field may provide a better kinematic representation of continental break-up at non-volcanic margins (e.g. Iberia) than a pure shear model.
T34A-04 INVITED
Lena Trough (Arctic Ocean): Active mantle exhumation on a continental rifted margin
Lena Trough is the northern continuation of the Mid-Atlantic Ridge through Fram Strait and into the Arctic Ocean. The rifting of Lena Trough began in the Miocene, and significantly, is the final and the most recent event in the separation of the North American from the Eurasian continent. Lena Trough was mapped in 1999, 2001 and 2004 by PFS Polarstern (Alfred Wegener Institute for Polar and Marine Research, Bremerhaven, Germany), revealing sea floor structures that are inconsistent with any normally conceived mid-ocean ridge spreading, and instead indicative of late continental rifting. Lena Trough is shown to be a deep, fault-bounded basin with depths of 3800-4200m, and irregular, steep valley sides that are oblique to the spreading direction. Basement horst structures that outcrop as sigmoidal ridges with steeply dipping sides project out of the valley floor. These basement ridges are roughly parallel along flow lines to the valley walls on either side. Ridge-orthogonal topography is simply absent (ie no segments trending parallel nor fracture zones perpendicular to Gakkel Ridge). Most faults trend approximately SSE-NNW, an obliquity with respect to Gakkel Ridge (SW-NE) of about 55°. The basement ridges are composed nearly entirely of fertile mantle peridotite, as are the valley walls. Only at the northern and southern extremities of Lena Trough do basalts appear at all. The peridotites compositions are consistent with either continental or oceanic (asthenospheric) mantle. They show evidence of low-degree mantle melting, followed by high-level stagnation in a thick lithosphere. This evidence (veining, impregnation) is more evident where little or no basaltic cover is present, while peridotites dredged in the vicinity of basalts tend to be more residual. This may indicate some degree of magmatic focusing in the absence of a basaltic crust per se. Lena Trough contains rare, highly alkaline basalts that are unlike any compositions dredged from mid-ocean ridges. While nearly all alkaline E-MORB have less then 49 wt. % SiO2 and less than 15% Al2O3, the Lena Basalts have nearly 52% SiO2, and 18% Al2O3. This suggests that the melts formed at the quartz eclogite peritectic with residual garnet rather than at the peridotitic peritectic usual for MORB. Their trace element and isotopic characteristics moreover reflect a ubiquitous enriched component found in the Western Gakkel Ridge and the Mohn’s Ridge. We suggest that the Lena Trough has undergone almost no partial melting, and that the rare basalts found there are nearly uniquely the result of melting of early-melting heterogeneities (veins) in the upwelling asthenosphere.
T34A-05
Mantle detachment faults and the break-up of cold continental lithosphere
We use a novel numerical approach, which fully couples the energy, momentum and continuum equations, to investigate the physics of extension and break-up of cold continental lithosphere to form new ocean basins. Unlike hot continental systems, where flat-lying detachment faults are nucleated in the strong part of the upper crust, cold continental systems have flat-lying detachment faults nucleating in the strong upper mantle at a relatively early stage. These detachment faults subsequently control the development of a mantle core complex and associated crustal structures. The observed structures are analogous to those developed in mid-crustal core complexes during extension of relatively thick and hot continental crust. In the cold environment, however, a strong elastic core is developed within the mantle, shifting the stress-bearing part of the system to below the Moho. Our modelling results reproduce key tectonic elements of a natural system (the Iberia Margin) by stretching a randomly perturbed, unpatterned lithosphere. Results also explain the "upper plate paradox" by doming of continental mantle separated from the crust by two diffuse detachment zones dipping towards the two future continental margins. Doming is facilitated by channel flow of the lower crust.
T34A-06
Isostatic Response of Strongly Extended Continental Lithosphere to Refertilization
During stretching of the continental lithosphere, the density of the mantle column changes. In contrast to the simplification often used in geodynamic modeling, mantle density does not linearly decrease with increasing temperature, but is strongly dependent on pressure and bulk rock chemistry, and non-linear due to mineral phase transitions. We computed density as a function of temperature, pressure and composition for extending lithosphere from internally consistent thermodynamic data. If the temperatures are sufficiently high and the crust is thin a significant part of the lithosphere consists of light plagioclase peridotite, which causes the lithospheric mantle to become buoyant at strongly stretched continental margins. This effect is most pronounced for fertile bulk compositions. The amount of plagioclase peridotite in the mantle lithosphere, and therefore the magnitude of the buoyancy (uplift), is mainly determined by the bulk rock aluminum content and the sodium/aluminum ratio. Re-fertilization of the mantle by infiltrating fluids or melts should therefore enhance the formation of plagioclase, decrease the density and increase upward motion. Detailed field work on fossile and current passive margins has confirmed that lithosphere extension is inevitably accompanied by extensive infiltration of the stretched lithospheric mantle by magmatic fluids. This leads to re-fertilization of the mantle, crystallization of new plagioclase and formation of plagioclase at the expense of spinel. The metamorphic reaction is enhanced by the flux provided by the infiltrating melts, shifting ‘metamorphic closure temperatures' to lower values. We therefore propose that mantle phase transitions might play an important role for the exhumation of strongly stretched continental mantle at passive margins.
T34A-07 INVITED
Modes of normal faulting and mantle exhumation at slow-spreading mid-ocean ridges. To what extent could they apply to ocean continent transitions ?
We use data from the Southwest Indian and Mid-Atlantic ridges to illustrate current concepts on the modes of exhumation of mantle-derived ultramafics in slow-spread oceans. We then compare these concepts with observations from slow-rifted, magma-poor, ocean continent transitions seismically imaged and drilled along the Iberia-Newfoundland margins. Preferred models for slow-spread oceans couple large offset normal faulting (detachment faults), with magmatic processes that prevent the formation of steady-state melt reservoirs at crustal levels. In extreme cases, volcanism is suppressed altogether. Proposed magmatic processes include reduced mantle melting, melt migration along-axis to focused volcanic centers, melt crystallization in discrete crustal intrusions, the freezing of melt into the lower lithospheric mantle, and inhibited melt transport in dykes through a thick axial lithosphere. Melt suppression due to conductive cooling of the mantle at very slow spreading-rates is probably not the dominant mechanism, except at ultra-slow or highly oblique ridges that overlie anomalously cold mantle. Detachment faults are inferred, based mostly on seismicity, to have dips > 40deg, and are found to be be active for < 3-4 myrs. Such relatively short lifetimes for individual faults do not necessarily prevent longer-lasting tectonic asymmetry between the two diverging plates. Dome-shaped, corrugated surfaces appear to form along the longer-lasting and larger offset faults, with large flexural rotation of the footwall, and under specific conditions of melt supply to the crust. Higher, or lower melt supply to the crust appears to result in shorter duration of slip on individual detachment faults (up to 1-2 myrs), and possibly in less pronounced footwall rotation. This suggests that faulting modes, and the distribution of weak zones in the axial lithosphere, are closely related with melt emplacement at crustal levels.
T34A-08
Plate motion of Iberia relative to Europe in the Cretaceous: problems with the fit at M0 time
Conjugate sets of sea floor magnetic chron M0 (120 Ma, Early Aptian) that have been mapped between Iberia and Newfoundland provide a well-constrained plate reconstruction of Iberia relative to North America. Other North Atlantic-bounding plate positions are less well constrained in the Early Cretaceous because these reconstructions require estimates of total extension vectors in basins bordering the North Atlantic. Published estimates of extension can, however, be used to build a satisfactory plate fit of Europe to North America in the Early Aptian. The resulting plate reconstruction shows a large, eastward-widening gap between eastern Iberia and Europe in the space now occupied by the Pyrenees. With M0 being at the beginning of the Cretaceous Superchron, the next youngest available plate reconstruction constrained by sea floor magnetic data is chron 34 (83 Ma, Campanian). The position of Iberia relative to Europe in this reconstruction suggests that the gap that existed between Iberia and Europe in the Aptian (at least in the M0 reconstruction) was closed by Campanian time. This implies several hundred kilometers of convergence between Iberia and Europe in the mid Cretaceous. Although feasible, such a large convergence amount should be recorded as a foldbelt and/or subduction system. Available data from the area suggest, however, that the mid Cretaceous was a time of tectonic quiescence, with the Pyrenean orogeny only initiating in the Eocene. One possible implication of these observations is that the fit of M0 is incorrect and that the magnetic lineation mapped as M0 is in fact not a sea floor spreading lineation. M0 occurs at the boundary between normal oceanic crust and the wide zones of exhumed mantle material found on both the Iberian and Newfoundland margins. We suggest that the M0 lineation is caused by a boundary effect between these different domains. An additional implication is that the boundary itself is not an isochron and is diachronous, with the transition from exhumation to sea floor spreading getting younger to the north along both margins.