LUNAR Mg-RICH ROCKS AS ANALOGS OF TERRESTRIAL KOMATIITES: IMPLICA- TIONS OF EARLY OUTGASSING OF EARTH'S VOLATILE ELEMENTS

Komatiites (ultramafic volcanic rocks) are common among Archean terrains, but rare or absent among rocks< 2.8 Ga old, which implies that the Archean upper mantle was fundamentally different from the modem one. The magma( s) parental to a certain group of lunar samples appear to have formed by exceptionally high-degree partial melting, like the magmas parental to komatiites: The Mg-rich suite of pristine nonmare rocks comprises numerous extremely magnesian troctolitic rocks, with Fo87 olivine; a few even have Fo92 olivine, i.e., as magnesian as liquidus olivine from a typical komatiite melt. One crucial difference between the ancient lunar mantle and the modem terrestrial one is that the latter is far richer in volatiles, most notably H20. Extraterrestrial samples indicate that the Earth formed hot, and volatiles were probably outgassed from its outer mantle. Isotopic data from Allegre et al. (1983) confirm tha the MORB source region outgassed > 4.4 Ga ago, but parts of the mantle did not. Eventually, the outer mantle must have regained volatiles, either by admixture of a veneer of carbonaceous chondrite-like material, acquired during a 'late heavy bombardment' > 3.9 Ga ago, or else by slow outgassing from the deeper mantle. Traces of water lower the solidus temperature (T5 ) of mantle peridotite by hundreds of degrees. It is widely assumed that due to greater radiogenic heating, the Archean lithosphere was much thinner than the modem one. Actually, wherever the outer mantle remained dry, the lithosphere was probably thick. The lithosphere is rigid and cools by conduction, unlike the asthenosphere which cools by convection. Convection is inhibited by viscosity, which is proportional to exp (30*T/f). Without water to lower T5 , the lithosphere/asthenosphere boundary must have been at a much higher T, which implies that it was deep. A locally dry lithosphere, comparable in thickness to the modem one, can explain many Archean phenomena. The exceptionally high liquidus temperatures of komatiites imply ascent of diapirs from great depth in a hot, rigid lithosphere: Melting seldom occurs in the asthenosphere, because convective cooling causes dT/dP to be quasiadiabatic, and thus dT/dP > dT/dP. Ultramafic melts are also more likely to come from a deep, dry source region because it would retain more pyroxene and less olivine. Metamorphic mineral assemblages from some Archean terrains indicate that sialic crust survived burial to depths> 30 km. Without insulation by a thick lithosphere, sialic materials at these depths would have been extensively melted.