Thermal State of the Lithosphere Beneath Mongolia and Southern Baikal Area: Implications for Lithospheric Structure and Mantle Dynamics in Central Asia

Dmitri A. Ionov1, Suzanne Y. O'Reilly 1 and William L. Griffin1,2

1Key Centre for Geochemical Evolution and Metallogeny of Continents (GEMOC), School of Earth Sciences, Macquarie University, NSW 2109

2CSIRO Exploration and Mining, Box 136, North Ryde, NSW 2113

A suite of garnet-spinel lherzolites, garnet websterites and garnet granulites (27 samples) from the Shavaryn-Tsaram eruption centre in the Pliocene-Pleistocene Tariat volcanic field, northern Hangai Mountains, central Mongolia yields pressure (P) and temperature (T) information for the upper mantle and lower crust in the region. The P-T data obtained using the Ca-in-opx thermometer of Brey & Köhler (1990) and barometer of Nickel & Green (1985) define a geotherm that passes through 12 kbar at 850°C and 20 kbar at 1220°C. T and P estimates for composite xenoliths (garnet pyroxenite veins in garnet-spinel and spinel lherzolites) are consistent for each xenolith and agree with phase transition boundaries for lherzolitic and pyroxenitic systems. T values calculated for 30 spinel lherzolite xenoliths range from 850°C to 1070°C but ~75% of these fall into a narrow interval of 895±20°C. Projection of the cut-off temperature for spinel lherzolites (870°C) onto the geotherm (assuming that the lowest T lherzolite represents the uppermost mantle) defines pressure at the crust-mantle boundary of 12-13 kbar indicating a rather high crustal thickness of about 45 km. This value is significantly lower than estimates (50-60 km) obtained by Zorin et al. (1990) from interpretation of gravity data.

Garnet and garnet-spinel lherzolite xenoliths also occur in a Paleocene picritic tuff deposit and in a few Pliocene-Pleistocene basanitic volcanoes on the Vitim Highland some 200 km east of Lake Baikal (Siberia) and 1100 km north east of the Tariat region. Garnet pyroxenites and granulites are very rare or absent in these occurrences. The thermobarometric data obtained for the lherzolite xenoliths show rather narrow P-T arrays (small depth range) and therefore do not define a full geotherm locus. However, these data indicate significant differences between thermal regimes recorded by xenoliths from the Paleocene picritic tuff and the younger basanites. The P-T data for xenoliths from the younger basanites plot on the high-P-T segment of the Tariat geotherm suggesting a similar thermal state of the lithosphere in the corresponding depth range in these two regions in the late Cenozoic. However, the xenoliths from the older (Paleocene) picritic tuff yield T estimates that are about 100°C lower at the same P suggesting that the lithospheric mantle beneath Vitim has been heated since the Paleocene. The heating may be related to the alkali basaltic volcanism on the Vitim Highland that largely took place in the Miocene.

Large volcanic fields made up of Miocene and less common Pliocene-Pleistocene lavas occur in the Hamar-Daban Range south of Lake Baikal bordering on northern Mongolia and about halfway between the Vitim and Tariat areas. Garnet-bearing xenoliths appropriate for P-T estimates are not available in the area. Spinel peridotite xenoliths from young (1-6 Ma) basaltic rocks in northern Hamar-Daban represent a narrow T range of 980±30°C; most spinel lherzolites from the southern part of Hamar-Daban (Bartoy) also fall into this range but some extend to higher T values (up to 1150°C). If the cut-off T values for spinel lherzolites from Hamar-Daban (~950°C) are projected onto the Tariat geotherm they yield the depth to the crust-mantle boundary (~15 kbar, ie ~55 km) that is much higher that estimates from regional geophysical data (40-45 km). It is possible that the uppermost part of the mantle has not been sampled, but this is unlikely because many Hamar-Daban xenoliths contain accessory feldspar. Alternatively, the geotherm in this region may be different, ie temperatures at the crust-mantle boundary (CMB) could be higher in southern Baikal area (Hamar-Daban) than in central Mongolia. A possible explanation is that voluminous underplating of basaltic magma at the CMB took place beneath the southern Baikal area in the Miocene, but not in the Tariat area where volcanic activity began in the Pleistocene.

The geotherm inferred for central Mongolia (Tariat) converges with eastern Australia geotherm at high pressures (~20 kbar) but yields significantly lower T values (~100°C) near the CMB. The CMB depth inferred for central Mongolia (45 km) is also much deeper than the average eastern Australia CMB (O'Reilly & Griffin, 1995). These differences in the thermal state of the lithosphere are consistent with the apparently more intense magmatic activity in eastern Australia in the Mesozoic-Cenozoic accompanied by under- and overplating at the CMB. Our data also indicate that the late Cenozoic volcanic activity in the Baikal area has resulted in some lithospheric heating.

The uppermost mantle sampled by peridotite xenoliths in the Tariat, Hamar-Daban and Vitim areas has fertile compositions (enriched in basaltic components) providing a geochemical contrast with more depleted and more magnesian mantle beneath the Siberian Platform to the north. The combination of the non-refractory compositions and the moderately high geothermal gradient (relative to stable platforms) is consistent with anomalously low seismic velocities right below the Moho in the region.