Meteorite impacts trigger massive mantle melting; a critical re-examination of the role of impacts in generating LIPs

Jones, A.P.1, G.D. Price1 and K. Wunnemann2
1Dept Earth Sciences, University College London,
2Dept Earth Science and Engineering, Imperial College London.

We are further exploring the potential melting response of terrestrial lithosphere and underlying mantle to large meteorite impact cratering (~ 200 km crater) through the use of hydrodynamic models, as a development of our preliminary models which suggested that large impacts may be expected to initiate Large Igneous Provinces (LIPs). Our new results provide more exact constraints on the key physical parameters which determine the volume, timescale and distribution of mantle melting beneath an impact crater. Our models produce identical results independent of the specific hydrocode model used (SALES, AUTODYNE), and, can reproduce exactly the recent work of Ivanov and Melosh (IM). We recognise, however, that shock processes are inherently non-equilibrium, and therefore unlike Ivanov and Melosh, we calculate the total volume of mantle material which experiences, during the shock, P-T condition above the solidus. This leads to our assessment that the volume of melt generated by a 20 km diameter peridotitic impactor travelling at 20 km/s can more by a factor of up to 100 than recently estimated by IM. The volume of melt critically depends upon the geotherm considered. The "hot" case modelled by IM is too cool to represent young (10-20 Ma) crust considered appropriate for modelling an impact origin of the Ontong Java Plateau. Mantle melting is emphatically not simply constrained to the few seconds related to passage of shock and release waves. Melting is virtually instantaneous (microseconds), whereas the kinetics of crystallisation are much slower (years), and depend critically on the degree of undercooling. Therefore, melts are very long lived in comparison to the shock process. Total melt volumes (~0.8 million cubic km for ~200 km crater with hotter geotherms) may be conservative, as we have (i) only modelled dry melting, (ii) assumed a mantle potential temperature of only 1270°C (rather than the higher values currently being inferred from petrology), and (iii) not yet considered the effect of melt mobility/eruption and lithospheric foundering, nor long-term, reorganisation of asthenospheric mantle flow etc. We conclude that (1) impacts into hot lithosphere produce 1-2 orders magnitude more melt than claimed by Ivanov and Melosh; (2) the melts may persist for ~thousands of years. We now need to consider the post-shock evolution of this melt, and its petrogenetic significance. Reference: Ivanov B A and Melosh H J Geology, 31, 869-872, 2003