Volcanology, Geochemistry, and Petrology [V]

V32A  MW:3007   Wednesday
Recycling of Deep Continental Lithosphere I: Consequences for the Mantle and Crust
Presiding: R L Rudnick, University of Maryland; S Gao, Chinese University of Geosciences; A Lenardic, Rice University; R Zhu, Chinese Academy of Sciences

V32A-01 

Subduction of Continental Crust into Earth's Transition Zone: an Experimental Investigation with Implications for the Fate of Continental Subduction

* Wu, Y (zmjin@cug.edu.cn), The State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences, 388 Lumo Road, Wuhan, 430074, China * Wu, Y (zmjin@cug.edu.cn), Geophysical Laboratory, Carnegie Institution of Washington, 5251 Broad Branch Road, N.W., Washington, DC 20018, United States Fei, Y (y.fei@gl.ciw.edu), Geophysical Laboratory, Carnegie Institution of Washington, 5251 Broad Branch Road, N.W., Washington, DC 20018, United States Jin, Z (zmjin@cug.edu.cn), The State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences, 388 Lumo Road, Wuhan, 430074, China

The occurrence of f ultra-high pressure metamorphic (UHPM) index minerals within rocks from continental collision zones indicate that continental rocks may descend into the Earth's upper mantle, perhaps even to the mantle transition zone, and leave specific geochemical signatures during the continental collision. The conceiving evidence of continental crust being subducted to depth >350 km has recently been considered with the discovery of former stishovite pseudomorphy in UHPM rocks from the Altyn Tagh, western China and relevant experiments at high pressure (>10-13 GPa). The idea that volumetrically abundant lithologies of continental crust, along with UHP eclogite could be subducted into the transition zone has also been supported by geochemical isotopic and seismic tomographical studies. In order to understand the fate of subducted continental crust, we conducted experimental studies up to 24GPa and 2073 K on natural UHPM gneisses powders in two bulk compositions corresponding to average upper continental crust and terrigenous rocks with piston-cylinder and multi-anvil apparatus. Phase identifications were based on raman spectra and microprobe analyses. Coesite, clinopyroxene, orthoclase and ganrnet with minor phengite or k-mica with unknown structure and epidote/lawsonite are found to be stable in runs up to 9GPa. At pressure between 9-14GPa, the charges mainly contain stishovite, cpx/jadeite, K-hollandite and garnet. Abundant mymekites consisting of K-hollandite and majoritic garnet (Si=3.18pfu at 14GPa and 1673K ) indicate the breakdown of biotite in starting materials. Rutile with α-PbO2structure has also been identified as break down product of titanite. At pressures from 14 to 24GPa, the Si and Mg contents in garnet increase with increasing pressure, while Al and Fe and Ca contents decrease accordingly. Contents of Al and divalent cations in stishovite show slightly increase with increasing pressure. The composition of cpx becomes gradually less aluminous and higher in jadeite component with increasing pressure. The solubility of NaAlSi3O8 and CaAl2Si2O8 component in K-hollandite, extending to 42 mol% and 11 mol% at 24GPa and 2073K respectively, shows a strong correlation with increasing temperature. Our experimental results suggest that the density of continental crust would be equal to or denser than that of pyrolite when those rocks have been transported to depth >300 km (9-10GPa), which would favor continental subduction into the lower part of the upper mantle.

V32A-02 INVITED 

Continental Crust and the Convective Thinning Mechanism

* Houseman, G A (greg@earth.leeds.ac.uk), School of Earth and Environment University of Leeds, Woodhouse Lane, Leeds, LS2 9JT, United Kingdom

Continental lithosphere is potentially unstable to gravitational instability triggered by convergent orogeny. Such events occur quickly on a geological timescale and are limited in the areal extent of lithosphere affected. The volume of lithospheric mantle that may be removed by one such event is illustrated at one extreme by the Tibetan Plateau or the Basin and Range province, and at the other extreme by the convergence associated with the Alpine Fault in New Zealand and the San Andreas Fault in southern California. Miocene convergence between Africa and Europe has resulted in a complex evolution which has almost certainly involved instability of continental lithospheric mantle in basins such as the Pannonian and Alboran Sea. The replacement of continental lithosphere by hot asthenosphere may result in lithospheric extension occurring even in the middle of a region affected by a convergent stress field. In parts of the Mediterranean the extension has naturally progressed to sea- floor spreading. Seismic tomography provides evidence that the removal of lithospheric mantle from beneath the Mediterranean and from beneath Tibet has occurred rapidly enough that unequilibrated, seismically fast, material remains now in the transition zone between 410 and 670 km depth beneath those regions. The extent to which continental crust is involved in such convective thinning events is questionable. The large density contrast between typical continental crust and mantle clearly would prevent unaltered crust being involved in anything like a Rayleigh-Taylor instability of the mantle lithosphere. If the transition to eclogite facies occurs, however, then the lower crust could well participate and in fact might provide an important contribution to the gravitational energy that drives a convective thinning event. Estimates of the volumes of continental crust that might be involved in continental convective thinning events are relatively small. For example, the present crustal volume of the Tibetan Plateau is consistent with what is expected from mass conservation calculations of the continental collision. It seems unlikely that more than 10% of that volume is unaccounted for. Thus it also seems unlikely that removal of eclogitised continental crust by the mantle drip mechanism is significant relative to the volume of oceanic crust consumed in subduction zones.

V32A-03 INVITED 

On foundering lithosphere and volatile migration: Upside-down melting

* Elkins-Tanton, L (ltelkins@mit.edu), Massachusetts Institute of Technology, 77 Massachusetts Ave. 54-824, Cambridge, MA 02139, United States

On Earth magmatism occurs on continents in the absence of subduction, often producing volatile-rich magmas such as those in the Leucite Hills, the Sierra Nevada, and Peru's Altiplano. The primary hypothesis to explain this volcanism is foundering of the lower lithosphere into the mantle. Here loss of the lower lithosphere is hypothesized to occur in a ductile manner in response to a density contrast such as would be caused by intruding mantle melts that freeze as eclogites. This mechanism requires no specific structural weakness beyond a dense region in the lithosphere that is gravitationally unstable with respect to the underlying mantle and that possesses a rheology conducive to flow. Density contrasts of as little as 1% are fully sufficient to drive gravitational instabilities. A gravitational instability forms when a perturbation in a boundary grows through lateral flow, causing the perturbation to grow. The growing instability begins to sink into the underlying mantle material as a drip, exactly analogous to but reversed in the sense of growth from an ascending plume head. The unstable material will sink more rapidly than lateral flow in the lower lithosphere can continue to add material to it, resulting in an annulus of thinned lithosphere centered on the instability. Thus the lithosphere is thinned slightly in the region around the drip, but no dome forms in the lower lithosphere during ductile delamination. Traditionally magmatism associated with instabilities has been attributed to return flow of the asthenosphere into such a dome, but maintaining a dome in the lithosphere requires unusual rheological conditions not expected in such a setting. Any volatile content in the sinking material may act in petrologically significant ways. The sinking lower lithosphere may contain 0.1 to 0.2 mass% of water if only nominally anhydrous minerals are present, and up to several weight percent of water if phlogopite or amphibole are present. The sinking lithospheric material heats conductively in the asthenosphere. Depending upon its rate of descent and volatile content, the sinking material may (1) devolatilize (as a descending slab in a subduction zone does), (2) carry volatiles to depth, sinking in some cases faster than slabs and thus carrying volatiles to depth more efficiently, or (3) heat sufficiently quickly to cross its solidus and itself produce magma. Because melting in instabilities would occur as they sink, we have termed this novel melting mechanism "upside-down melting." Upside-down melting has the potential to create primitive hydrous basaltic magmas with high alkali contents and lithospheric trace element signatures, typical of small-volume continental magmas worldwide. These magmas are compositionally distinct from the relatively dry adiabatic melts that result at mid-ocean ridges, and which the dry adiabatic melting created by the movement of the sinking instability would more closely resemble; the models successfully predict their range of mantle source conditions.

V32A-04 INVITED 

Recycling of mafic lithologies during continent formation

* Lee, C (ctlee@rice.edu), Rice University, 6100 Main St., MS-126, Houston, TX 77005, United States Hoink, T (tobias.hoeink@rice.edu), Rice University, 6100 Main St., MS-126, Houston, TX 77005, United States Luffi, P (pluffi@rice.edu), Rice University, 6100 Main St., MS-126, Houston, TX 77005, United States Lenardic, A (ajns@rice.edu), Rice University, 6100 Main St., MS-126, Houston, TX 77005, United States Anderson, D (dla@gps.caltech.edu), Caltech, 1200 E. California Blvd, Pasadena, CA 91125, United States

It is well-known that the continental crust is too silicic to have been derived directly as a melt from the mantle, so in order to balance the crust's composition with respect to basalt a missing mafic reservoir is needed. Thus, a multi-stage scenario is required to generate continental crust. First, basalt is extracted from the mantle. The basalt is then differentiated into felsic and mafic components, the latter of which is disposed of via various processes, loosely termed as delamination. Here, we discuss two of several possible crust formation scenarios. One mechanism is that of island accretion followed by further refinement by continental arc magmatism. Phanerozoic examples include the entire Cordilleran margin, extending from North to South America. The parental magmas are basaltic arc magmas generated in the mantle wedge. These arc basalts differentiate (by both fractional crystallization and re-melting of previously underplated basalt) into felsic and mafic lithologies, the former generating the crust and the latter residing in the deep crust in the form of "arc eclogites". The second mechanism is continent formation by underthrusting and stacking of oceanic lithospheres, a process likely to be confined to the Archean. In this scenario, as the underthrusted lithosphere heats up, serpentinized lithospheric mantle dehydrates, allowing the basaltic oceanic crust to undergo hydrous melting and generate felsic magmas. Such magmas rise up to form the crust, but the mafic residue remains in the underthrusted oceanic crust as eclogite. In both scenarios, the mafic residue or cumulate must be removed because there is little evidence that great quantities of these lithologies exist today within continents. In the arc accretion scenario, the mafic lower crust delaminates only after accretion. In the lithosphere stacking model, we propose that the underthrusted oceanic crust is later evacuated along the thrust faults without disturbing the peridotitic sections of the underthrusted lithospheres. The fate of these lithologies is uncertain. We speculate that underthrusted oceanic crust, once evacuated, may sink into the transition zone or deeper. As for the "arc-eclogites", some may heat up and melt soon after delamination because they are already warm. Such "arc-eclogites" would manifest themselves as fertile melting anomalies in the upper mantle.

V32A-05 

What are the differences between preserved vs. delaminated lower crust? Evidences from the Kohistan arc

* Jagoutz, O E (Oliver.Jagoutz@geo.unibe.ch), Institute for Geology, University of Bern, Baltzerstrasse 1+3, Bern, 3012, Switzerland Müntener, O (Othmar.Muntener@unil.ch), Institute of Mineralogy, University of Lausanne, Anthropole, Lausanne, 1015, Switzerland Burg, J (Burg@erdwethz.ch), Department of Earth Science, ETH Zurich, Clausiusstr, Zurich, 8092, Switzerland Schmidt, M (max.schmidt@erdw.ethz.ch), Department of Earth Science, ETH Zurich, Clausiusstr, Zurich, 8092, Switzerland Ulmer, P (peter.ulmer@erdw.ethz.ch), Department of Earth Science, ETH Zurich, Clausiusstr, Zurich, 8092, Switzerland

Delamination and foundering of the deep continental lithosphere into the mantle is an important crust forming mechanisms and might have implication for the formation of mantle heterogeneities. For example it is generally assumed that delamination of lower crust forms an enriched component within the mantle. However, the preserved lower continental crust is not complement to the upper crust. Therefore, if delamination is a significant crust forming process for the upper continental crust the delaminating crust must differ in mineralogy and composition from the preserved lower crust. A natural but unfortunate consequence of delamination is that the foundered crust is generally lost and therefore direct observations are scare. Consequently our knowledge of the composition and mineralogy of the preserved vs delaminated lower crust remains rather illusive. However, within the Kohistan arc (NE Pakistan) two mantle-crust sections (the Jijal and the Chilas section) are preserved which are markedly different in mineralogy. We describe lithologies of these sections and argue that fractionation mechanisms that produced them document two liquid lines of descent with largely differing initial water contents. We propose that the upper, non-sedimentary CC is dominantly formed by hydrous high-pressure fractionation of complementary garnet-pyroxene-amphibole-rich cumulates preserved in the Jijal section. In contrast, the lower crust of the Chilas section is formed by "less-hydrous" parental melts. Density considerations indicate that whereas the "Chilas lower crust" will be preserved over geological timescale the garnet-rich "Jijal lower crust" can delaminate and sink back into the mantle. The Chilas lower crust has a bulk composition remarkably similar to the bulk lower crust estimates. Contrasting the high density Jijal lower crust is strongly depleted in incompatible elements but has positive anomalies of Sr, Pb, and Eu and high Ba/Nb. If such a hydrous composition is transferred back into the depleted upper mantle its isotopic evolution can explain certain characteristics of intra plate magmatism.

V32A-06 

Continental Relamination Drives Compositional and Physical-Property Changes in the Lower Crust

* Hacker, B R (hacker@geol.ucsb.edu), Earth Science, University of California, Santa Barbara, CA 93117, United States Kelemen, P B (peterk@ldeo.columbia.edu), Columbia University, POB 1000, Palisades, NY 10964, United States Behn, M D (mbehn@whoi.edu), Woods Hole, Oceanographic Institution, Woods Hole, MA 02543, United States

A long-standing paradigm for the genesis and evolution of Earth's continental crust holds that the crust is andesitic and reached this composition in the ‘subduction factory' by delamination or foundering of a dense, mafic or ultramafic component into the mantle from the base of initially basaltic arc crust. However, the range of suggested compositions for the lower crust and our incomplete understanding of subduction-zone processes render this paradigm non-unique. Recent discoveries from (ultra)high-pressure xenoliths and terranes, combined with re-evaluation of methods for inferring lower crustal compositions from seismic velocity data, show that "relamination" of buoyant, subducting continental crust may be an efficient means of altering the composition of the lower crust. Ultrahigh-pressure terranes show that large areas (>60,000 km2) of continental crust are subducted to depths >100 km where they undergo heating to temperatures of 600-1000{°}C for periods of up to 20 Myr. Xenoliths from the Pamir show that subduction erosion can drag continental rocks to depths >90 km and temperatures of {~}1200{°}C. In both settings, devolatilization and melting transform cold, hydrous, low-density crust into hot, less hydrous residues. Felsic and intermediate rocks attain densities similar to the middle–lower continental crust; buoyancy may drive such rocks to rise through the mantle to pond at the Moho or higher crust levels. The calculated seismic wavespeeds of such material are indistinguishable from the bulk lower crust. Both ultrahigh-pressure continental subduction and subduction erosion operate at rates of 1-1.5 km3/yr, such that over the lifetime of Earth either could have led to large-scale ‘continental relamination', refining the composition and physical properties of the continental lower crust. http://www.geol.ucsb.edu/faculty/hacker/

V32A-07 

Recycling deep cratonic lithosphere and generation of intraplate magmatism in the North China Craton

* Gao, S (sgao@263.net), State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences, Lumo Road 388, Wuhan, 430074, China * Gao, S (sgao@263.net), State Key Laboratory of Continental Dynamics, Department of Geology, Northwest University, Taibai Road 229, Xi'an, 710069, China Rudnick, R L (rudnick@geol.umd.edu), Geochemistry Laboratory, Department of Geology, University of Maryland, College Park, MD 20742, United States Xu, W (xuwl261@sina.com), School of Earth Sciences, Jilin University, Jianshen Road 2199, Changchun, 130061, China Yuan, H (hlyuan@263.net), State Key Laboratory of Continental Dynamics, Department of Geology, Northwest University, Taibai Road 229, Xi'an, 710069, China Liu, Y (yshliu@cug.edu.cn), State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences, Lumo Road 388, Wuhan, 430074, China Walker, R J (rjwalker@geol.umd.edu), Geochemistry Laboratory, Department of Geology, University of Maryland, College Park, MD 20742, United States Puchtel, I (ipuchtel@mail.umd.edu), Geochemistry Laboratory, Department of Geology, University of Maryland, College Park, MD 20742, United States Liu, X (xiaomingliu@263.net), State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences, Lumo Road 388, Wuhan, 430074, China Liu, X (xiaomingliu@263.net), State Key Laboratory of Continental Dynamics, Department of Geology, Northwest University, Taibai Road 229, Xi'an, 710069, China Huang, H (huanghua0913@163.com), State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences, Lumo Road 388, Wuhan, 430074, China Wang, X (wang-xiaorui@263.net), State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences, Lumo Road 388, Wuhan, 430074, China

Early Cretaceous alkaline picrites and high-magnesium basalts from the North China craton provide evidence for recycling of continental lithosphere by density foundering. Both the picrites and basalts contain xenocrystic olivines with high Fo92-93 and low CaO (<0.10%), consistent with their derivation from, or interaction with Archean mantle lithosphere. Most importantly, both the picritic and basaltic lavas contain unusual, reversely zoned clinopyroxene phenocrysts whose cores have low MgO, high Na2O (up to 2.4 wt.%, or 17.3 mol% Jd), and frequently contain ilmenite exsolution lamellae, consistent with their crystallization from an eclogite- derived melt. In contrast, the clinopyroxene mantles have low Na2O (<0.92 wt.%, or <6.5 mol% Jd) and are lamellae-free, suggesting crystallization from a mantle-derived melt (picrite or basalt). Both the cores and mantles have high Al2O3 contents (up to 6.9 wt.%). These observations suggest that both the cores and mantles crystallized from an aluminous melt at mantle depths, but that the cores formed at a significantly greater depth (> 2.5 GPa) than the surrounding mantles (> 1.5 GPa). High Ni/Cr and low 100Mn/Fe of both the picritic and basaltic lavas and correspondingly high Ni and low 100Mn/Fe of magmatic olivines at a given Fo number indicate derivation from a source containing limited or no olivine. High Sr/Y, LaN/YbN and Th/U and low Lu/Hf, together with radiogenic initial 87Sr/86Sr and 187Os/188Os ratios and negative epsilon Nd values implicate contributions from melts derived from foundered eclogitic lower continental crust. Modeling results suggest that the basalt source region contained a moderate proportion (20-30%) of eclogite-derived component whereas the source of the picritic lavas contained a substantially greater proportion (70-80%) of a different eclogite-derived component. Collectively, these results suggest that both the basaltic and picritic lavas originated by partial melting of Archean lithospheric mantle that experienced varying degrees of reaction with eclogite-derived melts. Together with previous studies, these findings provide new evidence that thinning of the North China craton was caused by the removal of the lower lithosphere (mantle and lower crust). Recycling and melting of eclogitic lower crust may contribute more to mantle heterogeneity than has previously been recognized.

V32A-08 

Crustal and Upper Mantle Structure in the Eastern Northern China Craton from Seismic Array Studies and Implications for the Mesozoic-Cenozoic Lithospheric Reactivation

* Chen, L (lchen@mail.iggcas.ac.cn), Seismological Laboratory (SKL-LE), Institute of Geology and Geophysics, Chinese Academy of Sciences, No.19 Beituchengxilu, Deshengmenwai Qijiahuozi, Chaoyang District, Beijing, 100029, China Zhu, R (rxzhu@mail.iggcas.ac.cn), Seismological Laboratory (SKL-LE), Institute of Geology and Geophysics, Chinese Academy of Sciences, No.19 Beituchengxilu, Deshengmenwai Qijiahuozi, Chaoyang District, Beijing, 100029, China Ai, Y (ysai@mail.iggcas.ac.cn), Seismological Laboratory (SKL-LE), Institute of Geology and Geophysics, Chinese Academy of Sciences, No.19 Beituchengxilu, Deshengmenwai Qijiahuozi, Chaoyang District, Beijing, 100029, China Chen, Y (johnyc@pku.edu.cn), Department of Geophysics, School of Earth and Space Sciences, Peking University, No.5 Yiheyuan Street, Haidian District, Beijing, 100871, China

A detailed knowledge of the crustal and upper mantle structure beneath the North China Craton (NCC) is important for understanding the processes and mechanisms of the Mesozoic-Cenozoic lithospheric reactivation in the region. We report results from several dense temporary seismic arrays deployed in the region since 2000. Our imaging results show substantial lateral heterogeneities in crust and upper mantle structures among different tectonic settings in the eastern NCC and adjacent areas, indicating a complex crustal and upper mantle deformation process in the region. The crust and lithosphere of the region are generally thin and displays significant structural variations both laterally and vertically. In particular, the dramatically thinned lower crust, the 60~80-km thin lithosphere, and the general NW-SE anisotropy direction of shear waves in the east Bohai Bay Basin are all considered as specific features associated with the dominant NW-SE tectonic extension in the Mesozoic and Cenozoic. The well developed crust-mantle transition zone in addition to the relatively thick crust and lithosphere and the NE-SW fast shear wave polarizations under the west Taihang Mountain Range is presumably attributed to the extension related magmatic underplating at the base of the crust since the late Mesozoic. In contrast, the Yan Mountain Belt at the northeastern margin of the NCC is characterized by a sharp Moho, a slow but normal-in-thickness lower crust and a 90~100 km thin lithosphere. These features suggest that different lithospheric processes (e.g., multiple phases of contraction and probably crust involved delamination) might have played a significant role at the northeastern margin of the NCC besides the widespread tectonic extension that was more dominant within the cratonic interior. Most importantly, a sharp step of the lithospheric thickness from ~90 km to >130 km was imaged within a short distance of <100 km at the triple junction of the basin and mountains where an abrupt change of upper mantle seismic anisotropy was found. Such observations may indicate the preservation of the tectonic imprint of earlier contractional deformations in the present-day lithosphere and provide evidence for diverse mantle deformations in response to the Mesozoic-Cenozoic lithospheric thinning in the eastern NCC. Unlike the thinned crust and lithosphere that may reflect a warm shallow upper mantle beneath the region, the mantle transition zone was imaged to be on average 10~15 km thicker than the global average, implying a cold mantle environment presumably associated with the flat¡§Clying Pacific slab at the bottom of the upper mantle under the region. This discrepancy indicate that the shallow and deep upper mantle of the eastern NCC probably have not reached a thermal equilibrium, and the Cenozoic lithospheric process and magmatism in the region might have originated within the shallow upper mantle, possibly no deeper than the middle of the transition zone.