T31D-01
Evidence for three-dimensional melt migration at volcanic rifted margins
Magmatism at volcanic rifted margins may be expressed by the presence in seismic reflection profiles of seaward dipping reflector (SDR) sequences interpreted as lava flows. It may also be expressed in velocity models developed from wide-angle seismic data as a region at the base of the crust with velocities higher than those typically observed in the continental crust and lower than those in the mantle, normally in the range 7.2-7.6 km/s, attributed to mafic intrusion or underplating over regions typically 50-200 km wide. This voluminous magmatism is commonly attributed to the presence of anomalously high mantle temperatures at the time of rifting, though other factors such as convective upwelling and the presence of volatiles may also play a role in some cases. To date, models of volcanic margin development have been largely two- dimensional and have considered relatively simple rifting histories, and these models have been successful in explaining the large-scale characteristics observed in widely spaced seismic transects across such margins. We present two sets of observations that are poorly explained by such models and provide a basis for future consideration of the effects of more complex rifting histories and three-dimensional lithosphere geometry and mantle upwelling. The first set of observations comes from the Laxmi Ridge offshore western India. Here, although SDR sequences are poorly developed and the adjacent oceanic crust is only 5-6 km thick, a magmatic body > 100 km wide and up to 8 km thick is inferred at the base of the crust. The magmatic body is inferred to pre- date the formation of the oceanic crust by only a few million years, and during this time melt production decreased while extension increased. The second set of observations comes from the Eastern Black Sea Basin. Here, the eastern part of the basin is underlain by a volcanic rifted margin with characteristically high lower crustal velocities and 12-14-km-thick adjacent oceanic crust. Further west, however, there is little or no magmatic addition despite thinning of the crust by a factor of at least 4. A strike profile indicates that the change between these two extensional regimes occurs over a distance of < 30 km, which is too abrupt to be readily explained by along-strike changes in mantle temperature, volatile content or upwelling rate. In both locations, along-margin transport of melt may play a significant role.
T31D-02 INVITED
The Case for a Thermal Origin of Magmatism on the North Atlantic Continental Margin
The cause of the magmatism on 'volcanic' continental margins is still disputed, specifically as to whether it is due to increased mantle temperatures. New normal incidence and wide-angle seismic profiles across the Faroe and Hatton Bank volcanic margins in the NE Atlantic enable us to constrain the seismic velocities and volumes of both the extruded and intruded melt. Near the Faroe Islands, for every 1 km along strike, 360-- 400 cubic kilometers of basalt was extruded, while 540--600 cubic kilometers was intruded into the continent- ocean transition (COT). Lower-crustal intrusions are focused mainly into a narrow zone about 50 km wide on the COT, whereas extruded basalts flow more than 100 km from the rift. Deep-penetration seismic profiles show that melt is intruded into the lower crust as sills which cross-cut the continental fabric, rather than as 'underplate' of 100 per cent melt as has previously often been assumed. This means that measured lower-crustal velocities represent a mixture of continental crust and new igneous rock. Tomographic inversion of wide-angle traveltimes from 85 ocean bottom seismometers constrain average lower-crustal seismic velocities as 6.9-7.3 km/s under the COT, intermediate between the velocities of the continental crust and fully igneous oceanic crust on either side. By comparison with theoretical curves of igneous thickness versus seismic velocity (H- Vp), our observations are consistent with the dominant control on the melt production being elevated mantle temperatures, with no requirement for either significant active small-scale mantle convection under the rift or of the presence of fertile mantle at the time of continental breakup as suggested for the North Atlantic by other authors. The mantle temperature anomaly was c. 130-150°C above normal at the time of continental breakup, decreasing steadily by about 75°C over the first 10 Ma of seafloor spreading. Comparison with the conjugate Greenland margins reveal a similar history of elevated mantle temperatures at breakup time, but asymmetry in the conjugate structures suggests that initial continental stretching led to asymmetric crustal structures, as is frequently observed on non-volcanic margins, before the entire rift region was covered by igneous rocks. White, R. S. et al. (2008). Lower-crustal intrusion on the North Atlantic continental margin, Nature, v. 452, pp. 460-464.
T31D-03
Vertical Expressions of Horizontal Flows: Stratigraphic Manifestation of Transient Asthenospheric Flow
Convection within the Earth's mantle appears to be strongly time-dependent on geologic time scales. However, we have few direct observations which would help constrain the temporal and spatial variation of convection on timescales of 1--10 Myrs and length scales of 102--104 km. Recently, it has been demonstrated that transient uplift events punctuate the otherwise uniform thermal subsidence of sedimentary basins which fringe the Icelandic plume. In several of these basins, three-dimensional seismic reflection surveys calibrated by well-log information have been used to reconstruct a 55 Myr old transient event. The minimum amount of uplift is 500 m, which grew and decayed within 2 Myrs. There is good evidence that this event is diachronous and that its amplitude decreases with distance away from the reconstructed center of the putative Icelandic plume. These observations cannot be explained by glacio-eustatic sea-level changes or by crustal shortening. We describe a simple fluid dynamical model which accounts for these transient and diachronous observations. In this model, we assume that the Icelandic plume was already in existence and that it had an axisymmetrical geometry in which hot asthenosphere flows away from a central conduit within a horizontal layer. A transient temperature anomaly introduced at the plume center flows outward as an expanding annulus which generates transient uplift at the Earth's surface. Our stratigraphic observations can be accounted for using a plume flux of 1.3 × 108 km3 Myr-1. It is instructive to compare this value with a history of plume flux determined from the V-shaped ridges south of Iceland. This history suggests that plume flux waxed and waned through time. We suspect that the stratigraphic expression of transient convective behavior is common and that a careful examination of three-dimensional seismic reflection datasets will be fruitful.
T31D-04
Crustal Anatexis by Upwelling Mantle Melts in the N.Atlantic Igneous Province: the Isle of Rum, NW Scotland.
Crustal anatexis is a common process in the rift-to-drift evolution during continental breakup and the formation of Volcanic Rifted Margins (VRM) systems. 'Early felsic-later mafic' volcanic rock associations on the Continent Ocean Boundary (COB) of the N.Atlantic Ocean have been sampled by ODP drilling on the SE Greenland margin and the the Vøring Plateau (Norwegian Sea). Such associations also occur further inland in the British Paleocene Igneous Province, such as on the Isle of Rum (e.g., Troll et al., Contrib. Min. Petrol., 2004, 147, p.722). Sr and Nd isotope and trace element geochemistry show that the Rum rhyodacites are the products of melting of Lewisian amphibolite gneiss. There are no indications of a melt contribution from Lewisian granulite gneiss. The amphibolite gneiss parent rock had experienced an ancient Cs and Rb loss, possibly during a Caledonian event, which caused 87Sr/86Sr heterogeneity in the crustal source of silicic melts. The dacites and early gabbros of Rum are mixtures of crustal melts and primary mantle melts. Rare Earth Element modelling shows that late stage picritic melts on Rum are close analogues for the parent melts of the Rum Layered Suite, and for the mantle melts that caused crustal anatexis of the Lewisian gneiss. These primary mantle melts have close affinities to MORB whose trace element content varies from slightly depleted to slightly enriched. The 'early felsic-later mafic' volcanic associations from Rum, and from the now drowned seaward dipping wedges on the shelf of SE Greenland and on the Vøring Plateau show geochemical differences that result from variations in the regional crustal composition and the depth at which crustal anatexis took place.
T31D-05
Two-stage magmatism during the evolution of the transitional Ethiopian rift
The Ethiopian rift marks the transition between continental rifting and incipient seafloor spreading. The Ethiopia Afar Geoscientific Lithospheric Experiment (EAGLE) included a 400 km-long cross-rift profile with 97 broadband passive seismometers with the aim to investigate the change from mechanical to magmatic extension by defining the lithospheric structure and extent of magmatism beneath the rift. Complimentary studies of P-wave receiver functions, shear-wave splitting and teleseismic earthquake arrival times show that the lithospheric structure is inherently different beneath the north-western rift flank, rift valley and south- eastern rift flank, with contrasting crustal thickness and composition, upper mantle velocity and lithospheric anisotropy. Two stages of magmatic addition are interpreted: 1) a 6--18 km-thick underplate lens at the base of the crust, which probably formed synchronous with an Oligocene flood basalt event (and therefore pre-dates the adjacent rifting by ~20 Myr); and 2) a 20--30 km-wide zone of intense dyking and partial melt, which most likely pervades the entire crust beneath the rift valley and marks the locus of current rift extension. Furthermore, Precambrian collision-related lithospheric fabric is proposed to be the main source of the strong anisotropy that is observed along the entire cross-rift profile, which may be augmented by magmatism beneath the rift. An active, followed by a passive magma-assisted rifting model that is controlled by a combination of far-field plate stresses, the pre-existing lithospheric framework and magmatism is invoked to explain the rift evolution.
T31D-06
The Magma Plumbing System of Dabbahu and Gabho volcanoes (Afar rift, Ethiopia) from InSAR, GPS and Seismicity data
In September 2005, a 60-km-long dike, up to 8 meters thick, was intruded into the Dabbahu rift segment, a
nascent seafloor spreading center on the Nubia-Arabia plate boundary in the Afar Depression of Northern
Ethiopia. Localized subsidence of 2-3 meters at Dabbahu and Gabho, measured by InSAR, indicated that
some of the intrusion was fed from shallow magma chambers beneath Dabbahu and Gabho volcanoes, two
centers of focused silicic volcanism at the northern end of the rift segment. An array of 9 seismometers
recorded seismicity from October 2005 to April 2006 -- three were located in the area between Dabbahu and
Gabho, where an explosive, rhyolite eruption took place on 26 September 2005. Ten continuously-recording
GPS receivers were installed in January 2006, including one on the flanks of Dabbahu and one on Gabho. In
addition, Envisat was programmed to acquire SAR data on every overpass since September 2005, enabling
us to build time series of recent deformation.
The data show that:
(i) Gabho began to uplift aseismically in November/December 2005. Uplift was most rapid initially, with 25 cm
in the first six months, and continued until summer 2007. Since then it has been stable.
(ii) The southern flank of Dabbahu began subsiding immediately after the main dyke intruded, continuing until
~March 2006, and reaching a maximum of ~10 cm. This occurred above a band of seismicity that
dips to the north beneath Dabbahu.
(iii) The center of Dabbahu began to uplift in ~March 2006, and has continued steadily for at least 2
years. The total uplift (by July 2008) was ~50 cm. Seismicity in the first six months was concentrated at 3
km depth beneath the uplifting area.
(iv) Gabho and Dabbahu did not subside during the dyke injections that have occurred in the southern half of
the rift segment since 2005 (nine by July 2008).
Despite the remarkably similar behavior to the Krafla system in Iceland, which underwent a rifting episode
from 1975 to 1984, these observations require a more complex magma plumbing system. In contrast to the
single inferred shallow chamber beneath Krafla, multiple magmatic sources are required in the Dabbahu rift.
http://www.see.leeds.ac.uk/afar
T31D-07 INVITED
Lithospheric processes that enhance melting at rifts
Continental rifts are commonly sites for mantle melting, whether in the form of ridge melting to create new oceanic crust, or as the locus of flood basalt activity, or in the long initial period of rifting before lavas evolve fully into MORBs. The high topography in the lithosphere-asthenosphere boundary under a rift creates mantle upwelling and adiabatic melting even in the absence of a plume. This geometry itself, however, is conducive to lithospheric instability on the sides of the rifts. Unstable lithosphere may founder into the mantle, producing more complex aesthenospheric convective patterns and additional opportunities to produce melt. Lithospheric instabilities can produce additional adiabatic melting in convection produced as they sink, and they may also devolatilize as they sink, introducing the possibility of flux melting to the rift environment. We call this process upside-down melting, since devolatilization and melting proceed as the foundering lithosphere sinks, rather than while rising, as in the more familiar adiabatic decompression melting. Both adiabatic melting and flux melting would take place along the edges of the rift and may even move magmatism outside the rift, as has been seen in Ethiopia. In volcanism postdating the flood basalts on and adjacent to the Ethiopian Plateau there is evidence for both lithospheric thinning and volatile enrichment in the magmas, potentially consistent with the upside-down melting model. Here we present a physical model for the conjunction of adiabatic decompression melting to produce new oceanic crust in the rift, while lithospheric gravitational instabilities drive both adiabatic and flux melting at its margins.
T31D-08
Magmatism in rifting and basin formation
Whether heating and magmatism cause rifting or rifting processes cause magmatic activity is highly debated. The stretching factor in rift zones can be estimated as the relation between the initial and the final crustal thickness provided that the magmatic addition to the crust is insignificant. Recent research demonstrates substantial magmatic intrusion into the crust in the form of sill like structures in the lowest crust in the presently active Kenya and Baikal rift zones and the DonBas palaeo-rift zone in Ukraine. This result may be surprising as the Kenya Rift is associated with large amounts of volcanic products, whereas the Baikal Rift shows very little volcanism. Identification of large amounts of magmatic intrusion into the crust has strong implications for estimation of stretching factor, which in the case of Baikal Rift Zone is around 1.7 but direct estimation gives a value of 1.3-1.4 if the magmatic addition is not taken into account. This may indicate that much more stretching has taken place on rift systems than hitherto believed. Wide sedimentary basins may form around aborted rifts due to loading of the lithosphere by sedimentary and volcanic in-fill of the rift. This type of subsidence will create wide basins without faulting. The Norwegian- Danish basin in the North Sea area also has subsided gradually during the Triassic without faulting, but only few rift structures have been identified below the Triassic sequences. We have identified several mafic intrusions in the form of large batholiths, typically more than 100 km long, 20-40 km wide and 20 km thick. The associated heating would have lifted the surface by about 2 km, which may have been eroded before cooling. The subsequent contraction due to solidification and cooling would create subsidence in a geometry similar to basins that developed by loading. These new aspects of magmatism will be discussed with regard to rifting and basin formation.