GEMOC ARC National Key Centre

O'Reilly, S. Y.¹, Griffin, W. L.^1,2, Gaul, O¹ and Poudjom Djomani, Y.¹

1. GEMOC National Key Centre, School of Earth Sciences, Macquarie University, Sydney, NSW, 2109, Australia
2. CSIRO Exploration and Mining, PO box 136, North Ryde, NSW, 2113, Australia

The lithospheric mantle and 4-D lithosphere mapping

The subcontinental lithospheric mantle (SCLM) carries a geochemical, thermal and chronological record of large-scale tectonic events that have shaped the Earth's crust. The SCLM is part of the continental plate, and moves with the plates over the less rigid asthenosphere. It has long been accepted that "old" (cratonic) lithosphere is relatively deep, depleted and cold; more recently it has been recognised that "young" lithosphere is relatively thin, fertile and hot.

Development of the 4-D Lithosphere Mapping methodology (O'Reilly and Griffin, 1996) has provided tools for constructing realistic geological sections of the SCLM. Xenoliths and garnet and chromite xenocrysts from mantle-derived volcanics (eg basalts, lamproites, kimberlites) provide samples of the lithospheric mantle at the time of eruption. Where sufficient xenoliths and/or xenocrysts of appropriate composition are available, we can determine the paleogeotherm, the depth to the crust-mantle boundary, the detailed distribution of rock types with depth, the spatial distribution of fluid-related (metasomatic and anatectic) processes and the depth to the lithosphere-asthenosphere (LAB) boundary within the tectosphere. Volcanic episodes of different ages in one region provide this information for different time-slices corresponding to ages of the volcanism, while geophysical data (seismic, gravity, magnetic, thermal) can be used to extend the geologically-derived profiles laterally or to interpret lithospheric domains with geophysical signatures that can be matched with geologically mapped sections.

The basis of Lithosphere Mapping is the direct evidence for the petrology of the lower crust and upper mantle provided by xenoliths and xenocrysts of deep-seated rock types entrained in basaltic, kimberlitic and lamproitic magmas. These samples are generally transported to the surface in 10-30 hours, too fast for alteration or significant re-equilibration to occur. They yield the compositions and locations of specific rock types in the underlying crust-mantle section, and large specimens can be used to determine the petrophysical characteristics (density, acoustic velocity, magnetic properties, electrical and thermal conductivity, heat production) of the rocks at given depths.

The key technique of Lithospheric Mapping is the construction of empirical (paleo)geotherms at specific localities by use of xenoliths and xenocrysts. This information is then used to place individual samples (for which temperature (T) can be calculated) in their original vertical sequence, and thus give the distribution with depth of rock types and mantle processes such as metasomatism. The thermal state of a lithospheric column also influences geophysical characteristics: T determines density and thus affects seismic velocities and gravity; magnetic responses are confined to rocks above the Curie isotherm. Mantle-derived material sampled by volcanic episodes of different ages allows interpretation of the evolution of the lithosphere in four dimensions (ie time as well as space).

4-D lithosphere mapping shows that the depth of the LAB can range from about 250 to 150 km in cratonic areas, while it is seldom >150 km in circumcratonic areas. Distinctive rock-type profiles (mantle stratigraphy) can be mapped, followed laterally, and correlated with surface geology. In Siberia within-craton domains with distinctive mantle stratigraphy coincide with crustal terranes mapped at the surface (Griffin et al. 1998b). Markedly different SCLM sections sampled by Ordovician kimberlites in the Sino-Korean craton suggest that the major TanLu fault penetrates the lithospheric mantle and separates two Archean terranes (Griffin et al. 1998e; Xu et al. 1998). These studies provide evidence that individual Archean terranes or microcontinents developed their own distinctive SCLM, which survived accretion of the terranes into cratons and plate-tectonic translation during subsequent aeons. However, later tectonothermal events, such as rifting or large-scale magmatism, are associated with major changes in SCLM thickness and stratigraphy, and especially in composition. Why is this so?

Secular Variation in Lithosphere Composition

Boyd (1989, 1997) recognised a fundamental distinction between Archean cratonic mantle, represented by xenoliths in African and Siberian kimberlites, and Phanerozoic circumcratonic mantle, represented by xenoliths in intraplate basalts and by orogenic lherzolite massifs. Archean xenoliths are not only more depleted on average, but have higher Si/Mg (higher opx/olivine), and subcalcic harzburgites are well-represented in Archean xenolith and xenocryst suites, but essentially absent in younger ones. Analysis of >13,000 garnet xenocrysts from volcanic rocks worldwide shows a clear correlation of garnet composition with the tectonothermal age of the crust penetrated by the volcanic rocks (Griffin et al. 1998a,d). The xenolith and garnet data, taken together, indicate that the Archean/Proterozoic boundary represents a major change in the nature of lithosphere-forming processes. The garnet data further indicate that newly-formed SCLM has become progressively less depleted from Archean, through Proterozoic to Phanerozoic time.

In xenoliths, the Cr2O3 content of garnet correlates well with the Al2O3 content of the host rock (Griffin et al. 1998c). Xenolith suites also show good correlations between the content of Al2O3 and those of other major and minor elements; these correlations make it feasible to calculate the composition of a mantle section, given the median Cr2O3 content of garnet xenocrysts from that section (Griffin et al. 1998c; Table 1). The mean composition of SCLM beneath terrains of Archean, Proterozoic and Phanerozoic tectonothermal age, calculated in this way, show a clear secular evolution in all measures of depletion, such as Al, Ca, mg#, and Fe/Al (Table 2). Pristine Protoerozoic SCLM is moderately depleted, and intermediate in composition between Archean and Phanerozoic SCLM. Cenozoic SCLM, exemplified by Zabargad peridotites and by garnet peridotite xenoliths from young extensional areas of China, Siberia and Australia, is only mildly depleted relative to Primitive Mantle. SCLM beneath some Phanerozoic terrains, especially in Europe, is more depleted and may represent reworked Proterozoic SCLM (Table 2:"preferred").

Significance of SCLM evolution to geophysical interpretation

Average mineral compositions for each age group have been used to calculate average modes, densities and seismic velocities (Table 2). Archean SCLM is 2.5% less dense than the asthenosphere (approximated by PM); for the less-depleted Phanerozoic mantle the difference is <1%. Thermal expansion coefficients are identical within error for all compositions, so that these differences persist to high temperatures. At 25°C, the Vp and Vs of Archean SCLM are higher than that of Phanerozoic SCLM by ca 0.5% and 1.2%, respectively; the compositional differences thus account for ca 25% of the range observed by seismic tomography. Typical geotherms for cratonic and Phanerozoic areas were used to calculate the difference in Vp and Vs at 100 km depth; the Archean values are higher by 4-5%, corresponding to the ranges commonly seen by seismic tomography.

Lithosphere evolution and destruction

These physical property data are important constraints on the delamination and recycling of the SCLM. Archean lithosphere is highly refractory and buoyant compared to the asthenosphere (ÅPM; Table 2); this is progressively less true of younger lithosphere. Archean lithospheric mantle is significantly buoyant relative to the underlying asthenosphere. For Proterozoic lithosphere, density decreases moderately with depth at the typical geotherms of about 40-45 mW/m2, the lithosphere-asthenosphere boundary (LAB) is around 150 - 180km thick and down to about 80 km, the lithosphere column will have an integrated density approximately the same as the asthenosphere at the LAB. Sections thicker than 80 km are significantly buoyant and difficult to remove or delaminate by gravity alone. Phanerozoic lithosphere is thin and fertile, and for the lower geotherms (50-45mW/m2) the integrated density of the column is greater than the density of the asthenosphere for lithosphere up to about 100km thick and thus gravitationally unstable. This effect explains the thickness and apparent longevity of existing Archean lithosphere, but suggests that Phanerozoic lithosphere can be delaminated when cool, if >100km thick.

Tectonic or magmatic events that lead to the replacement of old SCLM by younger material cause changes in the density and geotherm of the lithospheric column, with major effects at the surface. In the Kaapvaal Craton (Brown and Griffin, 1997) thermal and chemical erosion produced in a thinner, hotter and chemically recharged (metasomatised) lithosphere, and led to significant uplift of the craton. In the eastern Sino-Korean craton, the removal of 3100 km of Archean lithosphere during the late Mesozoic was accompanied by uplift, basin formation and widespread magmatism. In this case, lithosphere replacement involved rifting, with contemporaneous upwelling of fertile asthenospheric material (Griffin et al., 1998e; Yuan 1996).

Consequences

Correlations between mantle type and crustal age indicate that continental crust and its underlying mantle are formed together and remain coupled for geologically long times. Destruction of Archean SCLM is difficult, but where it occurs, by thermal and chemical erosion and/or rifting, thinning and displacement, it has major thermal and tectonic consequences. These processes are important in area selection using geophysical techniques. For example, in diamond exploration, Archean areas penetrated by old kimberlites and subsequently affected by lithosphere erosion will not show the geophysical signature of old, cold lithosphere even though the old, potentially diamond-bearing kimberlites or alluvial diamond concentrations derived from them may still be present at the surface, as in the Sino-Korean craton.

Some applications of 4-D lithosphere mapping to Australia

Geothermal state of lithospheric domains

Empirical paleogeotherms constructed from geothermobarometry of deep-seated xenoliths and garnet ± chromite concentrates from basalts, lamproites and kimberlites around Australia reveal regions of different paleogeothermal signatures (O'Reilly et al., 1997). Differences between the tectonically young eastern areas and the cratonic western part of the Australian continent correspond to those shown on a large scale by long-wavelength magnetic data, which integrate the total magnetic signature from the lithospheric column where temperatures are below the Curie Point. Surface heat-flow measurements may not reflect deeper geothermal gradients and model-dependent extrapolations to lower crust and mantle depths must be used cautiously. In eastern Australia, where xenolith data are available (coinciding with the basaltic provinces), there is a remarkably consistent geotherm, which is independent of the age of the basaltic volcanism. This inflected (advective) geotherm is higher than conventional ocean basin geotherms and reflects thermal perturbation associated with volcanic episodes. It records the thermal state at the time of the particular volcanic activity, and decays towards a conductive geotherm over a period of <100 m.y. Data from the eastern craton margin (in South Australia and western New South Wales) indicate significant changes in the thermal state through time, while Archaean and Proterozoic areas in Western Australia reflect typically low geotherms. Knowledge of a geotherm for a specific lithospheric column can be used to construct a realistic distribution of rock types with depth. This lithospheric column provides constraints for the geologically meaningful interpretation of geophysical data and for placing geochemical and mantle process information in a spatial context.

Lithosphere transect across southeastern Australia

A lithosphere transect from Jugiong in Phanerozoic eastern NSW across the Tasman Line to the Eyre Peninsular in Proterozoic South Australia was carried out using xenoliths and garnet concentrates. The paleogeotherm decreases westward from greater than 50 mW/m2 at Jugiong to around 40 mW/m2 in South Australia.

"Chemical sections" were constructed using minor and trace element characteristics of garnet for each vertical section making up the transect, revealing:

. maximum Cr2O3 content of garnet increases from 3% in Phanerozoic areas to >10% in the Proterozoic sections

. mean TiO2 content does not vary systematically across the transect but increases with depth in all sections, reflecting increasing influence of melt-related metasomatism with depth

. mean Y/Ga ratios are extremely high in the Jugiong section and decrease rapidly to the west, indicating an increase in the degree of depletion through partial melting in this direction.

Although garnet reflects important chemical variations in the lithosphere, it is olivine that is the most abundant mantle mineral. The Fe/Mg ratio of olivine is important in controlling the physical properties of lithospheric regions (density, Vs, Vs) and indicating the melting history (depletion events through basalt extraction) in the lithosphere. We have developed a technique involving inversion of O'Neill and Wood's (1979) garnet-olivine Fe-Mg exchange geothermometer to calculate the Fe/Mg of olivine coexisting with each garnet grain. The algorithm for his inversion was calibrated using garnets from xenoliths where the composition of olivine was known. Xenoliths from both Archaean (Kaapvaal Craton) and Phanerozoic (eastern China) mantle were used in this test to ensure applicability to a variety of tectonic areas. The tests indicate that the inversion can reproduce olivine Fo contents with an error of ±0.5.

Figure 1. Variation in Mg content of olivine across the southern Australian transect.

The results of applying this inversion to the southern Australian transect show high Mg# olivine (Fo 92-93) in the shallow portions of the Proterozoic westerly sections (Fig. 1)with olivine of Fo 90-91 in the Phanerozoic east. There also is an overall trend to lower Mg content values with increasing depth in each section.

References

Boyd, F.R., 1989, Composition and distinction between oceanic and cratonic lithosphere. Earth Planet. Sci. Lett., 96, 15-26.

Boyd, F.R., 1997, Origin of peridotite xenoliths: major element considerations. In G. Ranalli et al. (eds.), High pressure and high temperature research on lithosphere and mantle materials. Univ. of Siena.

Brown, R. W., Gallagher. K., Griffin, W.L., Ryan, C.G., de Wit, M.C.J., Belton, D.X., and Harmon, R. 1998. Abstracts, 7th International Kimberlite Conference, Cape Town April, 1998.

Griffin, W.L., Fisher, N.I., Friedman, J., Ryan, C.G. and O'Reilly, S.Y. 1998a. Cr-pyrope garnets in the lithospheric mantle: I. Compositional systematics and relations to tectonic setting. Jour. Petrol. (in press)

Griffin, W.L., Kaminsky, F.V., Ryan, C.G., O'Reilly, S.Y., Natapov, L.M. and Ilupin, I.P., 1998b, The Siberian Lithosphere Traverse: Mantle terranes and the assembly of the Siberian craton. Tectonophysics (subm).

Griffin, W.L., O'Reilly, S.Y. and Ryan, C.G., 1998c, The composition and origin of subcontinental lithospheric mantle. In Y. Fei (ed.) Mantle Petrology: Field observations and high-pressure experimentation (in press).

Griffin, W.L., O'Reilly, S.Y., Ryan, C.G., Gaul, O. and Ionov, D. 1997d. Secular variation in the composition of subcontinental lithospheric mantle. In J. Braun et al. (eds), Structure and evolution of the Australian continent, Geodynamics Vol. 26, Amer. Geophys. Union, 1-26.

Griffin, W.L., Zhang, A., O'Reilly, S.Y. and Ryan, C.G. 1998e, Phanerozoic evolution of the lithosphere beneath the Sino-Korean Craton. In: M. Flower et al. (eds) Mantle Dynamics and Plate Interactions in East Asia Amer. Geophys. Union, 107-126.

O'Neill, H.StC. and Wood, B.J., 1979, An experimental study of Fe-Mg partitioning between garnet and olivine and its calibration as a geothermometer: Contributions to Mineralogy and Petrology, 70, p. 59-70.

O'Reilly, S.Y. and Griffin, W.L. 1993,. Heat pulses of the Earth: the volcano/ lithosphere/ asthenosphere connection through time. Plenary Address, IAVCEI General Assembly, September 1993: Ancient Volcanism & Modern Analogues, p82.

O'Reilly, S.Y. and Griffin, W.L. 1996, 4-D Lithospheric Mapping: a review of the methodology with examples. Tectonophysics 262, 3-18.

O'Reilly, S.Y., Griffin, W.L. and Gaul, O., 1997. Paleogeothermal gradients in Australia: key to 4-D lithosphere mapping. AGSO Journal of Australian Geology and Geophysics, Centenary Volume 17, 63-72.Yuan, X.,1996, Velocity structure of the Qinling lithosphere and mushroom cloud model. Science in China, Series D, 39, 235-244.

Xu, X., O'Reilly, S. Y., Griffin, W. L. and Zhou, X., 1998. The Nature of the Cenozoic Lithosphere at Nushan, Eastern China. In Flower, M., Chung, S.L., Lo, C.H. and Lee, T. Y. (eds) Mantle Dynamics and Plate Interactions in East Asia , Geodynamics Series Vol. 27, Amer. Geophys. Union, Washington D.C. pp. 167-196.

Yuan, X.,1996, Velocity structure of the Qinling lithosphere and mushroom cloud model. Science in China, Series D, 39, 235-244.
 
 

Table 1. Comparison of mean mantle compositions calculated from garnets, with average compositions of xenolith suites
 

Table 2. Calculated mean compositions for Archean, Proterozoic and Phanerozoic SCLM