V33D-01 INVITED
What Determines the Thickness of the Lithosphere?
Cratonic continental lithosphere can be thicker than 200 km and survive for billions of years. This is commonly ascribed to greater buoyancy and strength of the cratonic lithospheric mantle compared to younger lithosphere. The boundary between the lithosphere and the asthenosphere is often defined as an isotherm even though it is a rheological boundary that determines where rocks become stiff enough to resist deformation by mantle flow. The viscosity of rocks depends on temperature, but also on composition. Refractory mantle has a significantly higher solidus than fertile mantle, implying that refractory mantle has a higher viscosity than fertile mantle at the same temperature. In fact, a significant part of the cratonic lithospheric mantle might be so highly viscous that it behaves elastically. Another important factor for the rheology of the mantle is the water content. The presence of water reduces the viscosity of olivine and therefore significantly weakens the mantle. Refractory mantle that has experienced large degrees of melt extraction should be depleted in water, which adds to the strength imposed by the high solidus and low temperatures. Numerical simulations show that cratonic keels indeed resist erosion by mantle convection for infinitely long times if realistic viscosity contrasts are applied. However, observations also show that some cratons are destroyed over geological time and loose their thick roots. This destruction of ancient roots appears to be preceded by complete refertilization of the lithospheric mantle through metasomatic melt infiltration, which confirms the notion that composition plays an important role for the ability of continental roots to resist erosion. Whereas the thickness of Archaean cratonic lithospheres might be determined by the initial thickness of refractory mantle created during their formation, younger lithospheres seem to stabilize at much lower, but uniform thicknesses, even though they probably form by different mechanisms. We propose that the inherent layering of the lithosphere that is caused by mantle phase transitions may play an important role for the level at which the lithosphere-asthenosphere boundary is established. The garnet-spinel phase transition is located at ca. 90 km depth at 1330 °C in fertile mantle and is associated with one percent change in density. This density contrast is large enough to trigger gravitational instabilities and aid delamination of the garnet-peridotite part of the lithosphere. At the same time, the lower density of spinel-peridotite acts as a barrier to ascending small scale convection. Inherited compositional layering re-enforces this barrier. In addition to the density contrast the garnet- spinel boundary might also influence the rheology due to changes in the modal proportion of olivine relative to pyroxenes. The amount of pyroxene is reduced in the spinel-peridotite to garnet-peridotite reaction. Pyroxenes can host about 10 times more water than olivine, and redistribution of water between the minerals due to the phase transition will increase the water content in olivine and thereby cause a reduction in viscosity of garnet-peridotite relative to spinel-peridotite. The development of gravitational instabilities depends on contrasts in densities and viscosities, which can possibly be both provided by the layering caused by the phase transition.
V33D-02
A Global Lithosphere-Asthenosphere Boundary?
The notion that a rigid lithosphere moves over a weaker asthenosphere is widely accepted. Indeed, a seismically fast lid above a slower asthenosphere is observed regionally and globally by surface-wave studies. Furthermore, recent P-to-S (Ps) and S-to-P (Sp) studies have imaged the boundary on a regional scale at depths that are in general agreement with surface-wave studies. The Ps and Sp conversions provide high-resolution imaging of the lithosphere-asthenosphere boundary, which, where modeled, suggests that the boundary is sharp, occurring over depths of 11 km or less. Yet a global map of this fundamental boundary has yet to be determined since its depth and other properties likely vary among tectonic environments. We use scattered wave energy recorded at permanent seismic stations from 1990-2004 to map sub-Moho interfaces that might be associated with the lithosphere-asthenosphere boundary. Stacks of high-quality Ps data binned by epicentral distance at some individual stations suggest the presence of one or two negative discontinuities (velocity decreases with depth) at depths ranging from 60 to 110 km, arriving after the Moho Ps phase and before the first reverberated Moho phase. Sp results also generally confirm the presence of these discontinuities. These single-station results do not cover a global swath, but they do suggest that a discontinuity at these depths is a persistent feature for which regional or global imaging may be possible. Therefore, we have processed the entire IRIS FARM dataset to look for common global and regional features. In data binned by epicentral distance, direct conversions from the Moho and the 410- and 660-km discontinuities are clearly visible. The second crustal reverberation (Ppss) is also visible, whereas the first crustal multiple (Ppps) is not apparent. In addition, energy that is not related to crustal multiples is seen at 60 to 110 km depth, and is consistent with the existence of a velocity drop at these depths. The interference of this negative-polarity phase with the first crustal multiple, which has positive polarity, may explain the unexpected absence of this crustal phase. Data that are collapsed into a single trace by stacking along predicted Ps travel-time curves show a similar pattern. Direct conversions from the Moho, 410, and 660 are all apparent. The crustal reverberations are muted, as expected given their moveout. As in the single-station results, a significant anomaly is seen corresponding to Ps conversions from 60 to 110 km depth. Sp stacks also support the existence of a discontinuity in this depth range. Synthetic modeling of Ps indicates that this phase is too broad to be explained by a sharp boundary (0-5 km) at a single depth. However, a broader boundary, or more likely, a boundary that varies in depth with respect to location may better explain the character of this phase.
V33D-03 INVITED
Multi-scale dynamics near the Moho: The role of mass and energy exchange between the crust, sub-continental lithosphere and asthenosphere in crustal development
Although there are a wide range of models for the genesis and evolution of continental crust, a nearly universal feature of these models, either overt or implicit, is a statement of the mass balance problem: the mean continental crust is significantly more silicic relative to what is thought to be the mean influx of mantle melts. Revised estimates for mantle melt flux into the crust are even greater than previously posited, further exacerbating the mass balance problem (Dimalanta, et al, 2002; Jicha, et al., 2006). Geophysical observations, xenolith suites, and improved understanding of lower crustal phase assemblages and rheologies have all contributed to the view that crustal level material can potentially be involved in a delamination event and return to mantle depths, providing one process that may alleviate the mass balance problem and preferentially removes mafic components. This two-way coupling of material between the mantle and crust has specific implications for the thermal and chemical evolution of the underlying mantle. Here we assess melting processes occurring at multiple scales at or below the Moho discontinuity, along with equivalent processes occurring in the lower crust. We combine three types of physical modeling, finite volume, finite element, and lattice boltzmann techniques to understand the bulk dynamics, stress evolution and fine-scale segregation and reaction of melts with the surrounding residuum. We use fine scale models to develop "sub-grid" models to understand better how large-scale motions translate into melting and segregation at the fine-scale. We find that a thermally mature crust, due to a combination of increased thickness and elevated melt flux, promotes homogeneity of intermediate composition melts in the lower crust. Stochastic numerical simulation of melt intrusion in the lower crust indicates that provided that pressure is greater than ~9 kbar, and melting and extraction has proceeded to ~0.2 volume fraction or greater, the phase assemblage and rheology will favor density instabilities. The downwelling garnet pyroxenites generate return flow in the mantle that can produce excess melt and brings deeper reservoir material into the sub- continental lithosphere, providing a source for episodic production of mantle heterogeneity.
V33D-04
A Chemistry-Based Classification for Peridotite Xenoliths
The development of a petrological and geochemical database for mantle xenoliths is important for interpreting EarthScope geophysical results. Interpretation of compositional characteristics of xenoliths requires a sound basis for comparing geochemical results, even when no petrographic modes are available. Peridotite xenoliths are generally classified on the basis of mineralogy (Streckeisen, 1973) derived from point-counting methods. Modal estimates, particularly on heterogeneous samples, are conducted using various methodologies and are therefore subject to large statistical error. Also, many studies simply do not report the modes. Other classifications for peridotite xenoliths based on host matrix or tectonic setting (cratonic vs. non-cratonic) are poorly defined and provide little information on where samples from transitional settings fit within a classification scheme (e.g., xenoliths from circum-cratonic locations). We present here a classification for peridotite xenoliths based on bulk rock major element chemistry, which is one of the most common types of data reported in the literature. A chemical dataset of over 1150 peridotite xenoliths is compiled from two online geochemistry databases, the EarthChem Deep Lithosphere Dataset and from GEOROC (http://www.earthchem.org), and is downloaded with the rock names reported in the original publications. Ternary plots of combinations of the SiO2- CaO-Al2O3-MgO (SCAM) components display sharp boundaries that define the dunite, harzburgite, lherzolite, or wehrlite-pyroxenite fields and provide a graphical basis for classification. In addition, for the CaO-Al2O3-MgO (CAM) diagram, a boundary between harzburgite and lherzolite at approximately 19% CaO is defined by a plot of over 160 abyssal peridotite compositions calculated from observed modes using the methods of Asimow (1999) and Baker and Beckett (1999). We anticipate that our SCAM classification is a first step in the development of a uniform basis for classifying mantle xenoliths and will facilitate the use of databases to model physical characteristics such as density and anisotropy for integration with geophysical measurements. http://www.earthchem.org
V33D-05 INVITED
Earth's deep water cycle
Major minerals in the upper mantle and particularly in the transition zone can hold considerable quantities of hydrogen from water. Exactly how much there is and how this hydrogen affects the mineral's physical properties is actively being researched. Observed seismic waves show clear evidence for regions of anomalous seismic velocity, which may or may not be a result of spatial variations in hydrogen content. How much hydrogen is needed to explain the observations and how to discriminate the effects of hydrogen from those of heat is also a subject of ongoing research. We will discuss some of this recent and ongoing research on detecting water in the mantle and then focus on a seismic anomaly beneath the eastern margin of the US. This anomaly is a region of weakly reduced shear-wave velocity that stretches laterally in a direction parallel to the eastern margin and stretches vertically from the top of the lower mantle to lithospheric depths. In the lower mantle, this low S-velocity anomaly sits adjacent to and above an east-dipping high S-velocity region, which likely represents subducted lithosphere from the Farallon Plate. The least unlikely explanation for the low-velocity anomaly is that it is relatively hydrous, with up to 1 wt % of water above average. If this hydrous region is slightly buoyant, it could well up and hydrate the strong lithosphere at the currently passive margin between the North-American continent and Atlantic Ocean. If the lithosphere is sufficiently weakened by hydration it might give way to typical forces from ridge push and sediment loads to develop a new subduction zone. We will discuss implications of such a scenario for the evolution of plate tectonics.
V33D-06
Water contents in mantle xenoliths from the Colorado Plateau and vicinity: Implications for the rheology and hydration-induced thinning of continental lithosphere
Low angle subduction of the Farallon Plate beneath western North America during the early Cenozoic has been suggested to have introduced water into the overriding North America lithosphere. However, most of the evidence comes from trace element systematics and the presence of hydrous phases in lavas and some peridotite and eclogite xenoliths. Here, we present results from a systematic study on water contents in nominally anhydrous minerals (NAMs, including olivine, clinopyroxene and orthopyroxene) of peridotite xenoliths collected from the Colorado Plateau and vicinity. After water loss during magma ascent was corrected, olivines from 4 out of 5 sample localities are found to contain ~40 to 70 ppm H2O compared to ~10 to 40 ppm H2O in olivine from the fertile asthenosphere. Assuming this amount of water (~70 ppm H2O in olivine) was introduced into the base of the lithospheric mantle by hydration, application of an updated flow law for dislocation creep of wet olivine to lithospheric mantle conditions beneath the Colorado Plateau predicts that for given background shear stress, hydration alone could have resulted in a more than 1 order of magnitude viscosity drop at the base of the lithosphere. The reduction in viscosity would have eased any basal erosion of the lithosphere driven by the convective stresses associated with low angle subduction and Basin and Range extension. We show that hydration could have easily permitted more than 15 km of lithospheric thinning. Viscosity reduction and lithospheric thinning of even larger extents (up to ~100 km lithospheric thinning) are predicted to be possible when thicker lithosphere (such as Archean cratons) and larger water contents (up to water-saturated condition) are considered. If our interpretations are correct, the implications of our study go beyond western North America and hint at a plausible way of recycling continental mantle, including cratonic mantle, back into the convecting mantle.
V33D-07
Formation and Evolution of the Continental Lithospheric Mantle: Perspectives From Radiogenic Isotopes of Silicate and Sulfide Inclusions in Macrodiamonds
Silicate and sulfide inclusions that occur in diamonds comprise the oldest (>3 Ga), deepest (>140 km) samples of mantle-derived minerals available for study. Their relevance to the evolution of the continental lithosphere is clear because terrestrial macrodiamonds are confined to regions of the Earth with continental lithospheric mantle keels. The goals of analytical work on inclusions in diamond are to obtain paragenesis constraints, radiogenic ages, and initial isotopic compositions. The purpose is to place diamond formation episodes into the broader framework of the geological processes that create and modify the continental lithosphere and to relate the source of the C and N in diamond-forming fluids to understanding the Earth's C and N cycles in the Archean. Although sulfide and silicate inclusions rarely occur in the same diamond, they both can be grouped according to their geochemical similarity with the chief rock types that comprise the mantle keel: peridotite and eclogite. Silicate inclusions are classified as harzburgitic (depleted; olivine > Fo91, garnet Cr2O3 > 3 wt% and CaO from 0 to 5 wt%), lherzolitic (fertile), or eclogitic (basaltic; garnet Cr2O3 < 2 wt% and CaO from 3 to 15 wt%, clinopyroxene with higher Na2O, Al2O3, and FeO); they are amenable for trace element study by SIMS and for Sm-Nd and Rb-Sr analysis by conventional P-TIMS after grouping by mineralogical similarity. Sulfide inclusions (chiefly FeS with lesser Ni, Cu, and Co) are classified as peridotitic (Ni > 14 wt%; Os > 2 ppm) versus eclogitic (Ni < 10 wt%; Os < 200 ppb); single sulfides are amenable for S isotopic study by SIMS or TIMS, and Re-Os analysis by N-TIMS. Work on inclusions in diamonds depends on the distribution of mined, diamond-bearing kimberlites, and the generosity of mining companies because of the extreme rarity of inclusions in suites of mostly gem-quality diamonds. Most isotopic work has been on the Kaapvaal-Zimbabwe craton with lesser work on the Slave, Siberian, and Australian cratons. Sm-Nd ages on silicate suites and Re-Os ages on sulfide suites confirm diamond formation from the Mesoarchean though the Neoproterozoic. Most important are the systematics across cratons in the context of crustal geology that lead to generalities about craton evolution. Inclusion suites date mantle keels as Mesoarchean and clearly point to subduction as the major process to form the earliest continental nuclei and to amalgamate the cratons in their present form. This is evident from the elevated initial Os isotopic compositions in 3.5 Ga Slave (Panda) and 2.9 Ga Kaapvaal (Kimberley) sulfides, the low Sm/Nd and elevated initial Sr isotopic compositions of 3.4 Ga Kaapvaal (Kimberley) harzburgitic garnets, the preponderance of 2.9 Ga eclogitic sulfides in every western Kaapvaal craton locality, and the occurrence of surficial, volcanogenic S in Kaapvaal (Orapa) sulfides. The continental lithosphere was accessible to melts and fluids from the asthenosphere throughout the Proterozoic as evident from silicate and sulfide inclusion suites of 0.9 to 2.0 Ga age in every locality studied in the Kaapvaal craton. The correspondence of silicate inclusion type with current seismic velocity structure of the Kaapvaal mantle keel shows that its structure is at least Bushveld age (2 Ga) and due to compositional differences. Seismic velocity structures of continental mantle keels may be more a function of their geologic history than current temperature distribution.
V33D-08
Thermal Evolution of Cratonic Roots
Thermal models for the stabilization of the cratonic lithosphere are constrained by present heat flow and heat production data from Archean Provinces. Archean provinces are presently characterized by low heat flow, with an average of 41 mWm-2 less than the global continental average (56 mWm-2). The range of regionally averaged heat flow values in Archean Provinces (18-54 mWm-2) is narrower than in Proterozoic and Paleozoic terranes. However, at the end of the Archean, when crustal heat production was double the present, surface heat flow varied over a range (~45-90 mWm-2) at least as wide as that presently observed in Paleozoic Provinces. The high crustal heat production during the Archean is not sufficient to account for elevated lower crustal temperatures without some additional heat input or without the crust being thicker or the vertical distribution of radio-elements being different from today's. Lithospheric heat production is a key variable in determining thermal conditions that permit stabilization of the crust and the preservation of a thick cratonic root. Stability of the crust and the cratonic root requires strong differentiation in the distribution of the radiogenic elements. Even for values of the surface heat flow higher than average in cratons, the crust can be stabilized before 2.5Ga if it is very differentiated and the radioelements are confined to shallow layers. Prior to differentiation, the lowermost crust could be near the solidus for present surface heat flow 40-45 mWm-2. Present heat production in the mantle root is constrained by the estimates of the mantle heat flow. Further constraints can be obtained by modeling the past thermal regime of the root when heat production was higher. If heat generation is high and/or if the root is thick, the lower lithosphere has cooled more rapidly than the convecting mantle, which implies that the temperature gradient was inverted at the base of the lithosphere and a weak mechanical layer in the middle of the root, precluding the survival of the root. A low temperature gradient at the base of the root leads to the development of convective instabilities and possible removal of the lowermost part. The large time-scale of diffusive heat transport implies that the lithospheric mantle can remain thermally decoupled from the crust for as long as 1 Gyr. Temperatures in the lithospheric mantle may rise above their initial values inherited from the process of root formation due to in-situ radiogenic heat production.