HR: 13:40h
AN: V52D-01    [PDF]
TI: Nb-U-Th-La Tracers of Basalt Sources Revisited
AU: * Hofmann, A W
EM: hofmann@mpch-mainz.mpg.de
AF: Max-Planck-Inst. f. Chemie, Postfach 3060, Mainz, 55020 Germany
AB: Oceanic basalts from mantle upwellings (MORB, OIB) have typically much higher Nb abundances relative to other highly incompatible elements than do island arc and continental rocks, and this serves as one of the chief geochemical discriminants between the two tectonic environments. Hofmann et al. (1986) quantified this and claimed, on the basis of 28 MORB and 41 OIB analyses (from 11 oceanic islands), that Nb/U ratios are surprisingly uniform at 47 $\pm$ 10 (=$\pm$ 21%) in MORB and OIB. They further argued that mantle melting in the MORB-OIB environment does not change Nb/U, so that this ratio also characterizes the mantle source(s) of these rocks, in contrast with a much lower average Nb/U =10 of the continental crust. Finally, they concluded that this rules out recycling of continental crust as an explanation for the trace element enrichment of OIB sources. Subsequently, local correlations between Nb/U and 87Sr/86Sr in EM-type OIBs indicated at least some admixture of recycled continental material to those sources (Chauvel et al., 1992, Hofmann, 1997), so that specific source contributions of OIBs can be identified. Sims \& DePaolo (1997) reanalyzed the 1986 database and concluded that the Nb-U relationships (as well as Ce-Pb, not discussed here) are consistent with recycling of continental material to explain the entire isotopic heterogeneity of OIB sources. Seventeen years later, the amount of relevant data has increased dramatically, and their assessment is facilitated by the GEOROC and PetDB databases (http://www.earthchem.org/). Even now, a major difficulty lies in filtering out data of poor quality, due to either alteration or analytical errors. This can be done using criteria of analytical technique and data scatter, but a subjective judgment is unavoidable. I present a survey of about 450 OIB analyses from 20 (non-EM-2 type) hotspots and 170 MORB analyses from the three major oceans. Elements forming a globally uniform concentration ratio should be linearly correlated on a logA-logB plot with a slope of 1.0. Thorium, U, and La in MORBs form tightly linear correlations with Nb, with slopes changing systematically from 1.06 through 0.99 to 0.71, respectively. For OIBs, the respective slopes are 1.11, 1.06, and 0.93. Thus, for both MORBs and (non-EM-2) OIBs, the logNb-logU correlations have slopes closest to unity, with mean values of Nb/U = 47$\pm$24% (MORB) and 52$\pm$25% (OIB). R.S.D. values of Nb concentrations are 100 % (MORB) and 123% (OIB). Although no attempt has so far been made to weight the data, e.g. by location, this evaluation of a greatly expanded dataset provides remarkable confirmation of the original conclusions, with the important modification that most EM-2 (and at least some EM-1) type OIBs have significantly lower Nb/U ratios, which correlate with 87Sr/86Sr ratios, consistent with significant input of recycled sediment or other continental material in their sources. Other OIB sources contain little or no recycled continental source components. Chauvel, C. et al. (1992) Earth Planet. Sci. Lett. 110, 99-119. Hofmann, A.W. et al. (1986) Earth Planet. Sci. Lett. 79, 33-45. Hofmann, A.W. (1997) Nature 385, 219-229. Sims, K.W.W. \& DePaolo, D.J. (1997) Geochim. Cosmochim. Acta 61, 765-784.
DE: 1025 Composition of the mantle
DE: 1065 Trace elements (3670)
DE: 4825 Geochemistry
DE: 4875 Trace elements
SC: Volcanology, Geochemistry, Petrology [V]
MN: 2003 Fall Meeting