Using a transport balance model with a lower and upper mantle, an oceanic crust and a continental crust, we have simulated Earth history commencing with accretion 4.55 Ga ago, aiming to reproduce present day Pb and Nd isotope compositions and U, Th, Pb, Sm and Nd concentrations of the known reservoirs. In particular, modelling terrestrial Pb isotope evolution must account for the 'future paradox' (average upper continental crust and upper mantle Pb plots to the right of the meteorite isochron in the 207Pb/204Pb vs 206Pb/204Pb diagram) and the Th/U mantle paradox (the Th/U ratio of the upper mantle is too low to have produced its 208Pb/206Pb ratio over time). Important new constraints are: Hf-W data define a short effective accretion and core formation time interval of c. 20 Ma (Jacobsen and Harper, 1995) and K isotope data preclude Pb loss by volatilization during Earth accretion (Humayun and Clayton, 1995). Therefore a 'conservative' reservoir with a low U/Pb ratio to account for the 'future' paradox has to be located in the silicate Earth. Further, noble gas data require a lower mantle reservoir with limited exchange to the upper mantle since c. 4.4 Ga ago (Tolstikhin and O'Nions, 1995). This negates models in which the upper mantle is generated from the lower mantle over time as the continental crust grows (McCullogh and Bennett, 1994).
To accommodate the heterogeneity in age and geochemistry of the continental crust, it is subdivided into four domains of equal size: 'younger' and 'older' as well as 'upper' and 'lower'. Newly formed sial is added to the 'younger' crust reservoirs and corresponding bulk mass fluxes to the 'older' reservoirs keep the domains equal in size. In recycling by erosion, 'younger' crust is more likely to be eroded than 'older' crust (2x in most scenarios). Recycling of sialic crust into the mantle is by erosion of upper crust and delamination of lower crust, in variable proportions, whereby bulk fluxes from 'lower' to 'upper' crust reservoirs and vice versa keep the masses of the crustal domains equal. After 2 Ga b.p., erosional recycling of U is enhanced relative to Pb, Th and REE (Staudigel et al., 1995). An important set of variables is the curve giving the mass of sialic crust existing at time t vs. t. Together with a given present day age distribution curve, this determines rates of crustal growth and recycling into the mantle. Sm-Nd data alone do not uniquely determine crustal growth and recycling, but for a given amount of continental crust in the very early Archean, present day mantle and crust Sm and Nd concentrations as well as an eNd value of +10 for the upper mantle further constrain the recycling history. Pb modelling was done using scenarios successful in terms of these present day Sm-Nd characteristics.
Good fits to present day Pb isotope compositions were only obtained with growth models in which the crustal mass was < 2E24 g by 4.4 Ga ago, then grew to reach 1.55E25 g by 1.7 Ga ago, followed by a slower increase to 2E25g at the present time. Further, the rate of sialic crust recycling into the mantle was 10% or less of the rate of crust formation up to c. 2 Ga ago, and then increases to reach c. 50% at present. The mantle Th/U paradox is solved with U recycling enhanced by a factor of c. 1.8 compared to Th and Pb. No fits to data were obtained with models in which a large mass of crust (> 4E24 g) was produced in the first 400 Ma of Earth history: The large amount of recycling required by these models to reproduce the approximate present day age distribution of the continental crust destroys the low U/Pb 'old lower crust' reservoir. Further, all models in which lower crust delamination is more than 10% of total recycling fail to produce fits, as the lower crust is again not 'conservative' in these. Crust histories successful for Pb modelling all produce similar eNd vs. time curves for the upper mantle, which stay very close to zero until c. 3 Ga ago, to increase to 10 almost linearly with time after that. This is in agreement with most eNd data derived from mafic rock suites throughout Earth history, but in conflict with results obtained by the back-correction of data from ancient crust provinces (Bennett et al., 1993; Bowring and Housh, 1995). The low rate of continent recycling found for the Archean is in agreement with conclusions from Hf isotopes (Patchett et al., 1984). Our results reflect a non-steady-state Earth in which the continental crust is still growing at present.
Bennett, V.C., Nutman, A.P. & McCulloch, M.T., Earth Planet. Sci. Lett. 119, 299-317 (1993).
Bowring, S.A. & Housh, T.B., Science 269, 1535-1540 (1995).
Humayun, M. & Clayton, R.N., Geochim. Cosmochim. Acta 59, 2131-2148 (1995).
Jacobsen, S.B. & Harper, C.C., In Isotopic Studies of Crust-Mantle Evolution (Basu, A.R. & Hart, S.R., eds.) A.G.U. Monograph (1995).
McCullogh, M.T. & Bennett, V.C., Geochim. Cosmochim. Acta 58, 4717-4738 (1994).
Patchett, P.J., White, W.M., Feldmann, H., Kielinczuk, S. & Hofmann, A.W., Earth Planet. Sci. Lett. 69, 365-378 (1984).
Staudigel, H., Davies, G.R., Hart, S.R., Marchant, K.M. & Smith, B.M., Earth Planet. Sci. Lett. 130, 169-185 (1995).
Tolstikhin, I.N. & O'Nions, R.K., Abstract AGU Spring Meeting (1995).