We present a comprehensive set of Sr, Nd, and O isotope data and trace element concentrations from tholeiitic and alkaline lavas of the neovolcanic zones of Iceland (picrites, olivine and quartz tholeiites, transitional and alkali basalts, differentiated rocks). Variations in the oxygen isotope results allow us to distinguish two groups. The first, which comprises quartz tholeiites and more differentiated rocks usually associated with central volcanoes, has low r5180 values (+5 to + 1%o) resulting from interaction with the hydrothermally altered Icelandic crust. The second group, which contains picrites, olivine tholeiites, and alkali basalts, has normal mantle oxygen isotopic compositions (r5180 --+5 to +6%o) which are thought to represent those of the mantle source. Nd isotopic compositions vary greatly, from •43Nd/•44Nd -0.51314 in picrites to 0.51295 in alkali basalts. To produce such a variation for rocks with the chemical compositions of Icelandic volcanics (147Sm/144Nd = 0.12-0.28) requires >200 m.y., a period that greatly exceeds the maximum age of Icelandic crust. Previous models, in which the Sr isotopic variations were explained in terms of evolution of crustal reservoirs, are invalidated, and mantle reservoirs with different Nd and Sr isotopic compositions are indicated. The Iceland data define a linear array in the Sr-Nd isotope diagram which overlaps both mid-ocean ridge basalt and oceanic island basalt fields and indicates mixing between depleted and enriched end-members. Alkali basalts come preferentially from an isotopically and chemically enriched component of the Iceland plume, and picrites come from a more refractory, more depleted portion. Positive Sr, Rb, and Ba anomalies are present in picrites and other lavas with low trace element contents. These anomalies are not correlated with isotopic differences but are nevertheless believed to result from interaction between the parent magmas of these rocks and altered Icelandic crust. This indicates that even the most primitive Icelandic lavas have been contaminated with some crustal material. Pb [Hart et al., 1973; Sun and Jahn, 1975; O'Nions et al., 1977; Zindler et al., 1979]. Models proposed to explain the chemical data fall into two groups. The first requires the existence of isotopic heterogeneities in the mantle [Zindler et al., 1979]: the chemically •Now at G6osciences Marines, Ifremer-Brest, Plouzan6, France. 2Now at G6osciences, Universit6 de Rennes 1, Rennes, France. and isotopically enriched lavas come from a source within the plume; the less enriched lavas come either from another component of the plume [Elliott et al., 1991] or from the depleted source of North Atlantic MORB [Schilling, 1973; Hart et al., 1973 ; Langmuir et al., 1978]. In the second type of model a homogeneous mantle source yields magmas that interact with old, hydrothermally altered Icelandic crust during their ascent and eruption [Oskarsson et al., 1982; Steinthorsson et al., 1985; H•mond et al., 1988; Nicholson et al., 1991]. The basis of the latter type of model w...

[1] We present new trace element and isotopic data on a selection of least evolved lavas in Iceland. Pb isotopic compositions are variable: 206 Pb/ 204 Pb ranges from 18.05 in picrites to 19.15 in alkali basalts. Large excesses of Sr and Pb and a deficiency of Hf and Zr relative to rare earth elements are characteristic of the picritic lavas (Ce/Pb % 10, Sr/Nd % 35, and Sm/Hf % 1.75), whereas more normal ratios are observed in the alkali basalts (Ce/Pb % 40, Sr/Nd % 10, and Sm/Hf % 1.4). These variations are correlated to systematic changes in Sr, Nd, and Pb isotopic compositions. Using these new data, the relationship between the Iceland plume and the Mid-Atlantic Ridge can be reexamined. The 87 Sr/ 86 Sr versus 206 Pb/ 204 Pb relationship demonstrates clearly that all isotopic variations in Iceland have their origin in the plume source. In addition, the composition of both Reykjanes and Kolbeinsey Ridges results from contamination of the depleted mantle by the Iceland plume. Trace element and isotopic heterogeneities within Iceland are related to rock type and not to location in the island. This is contrary to what would be expected if a uniform plume mixed with surrounding mid-ocean ridge basalt (MORB), in which case the geochemical values would be more MORB-like at the perimeter. Instead, it suggests that the Iceland plume itself was highly heterogeneous. We propose that the source of the Iceland plume was old recycled oceanic crust that was stored in the mantle for a long time (potentially since the end of the Archean). Melting of the basaltic portion of this crust led to the formation of the alkali basalts, and melting of the clinopyroxeneplagioclase gabbroic portion of the same oceanic crust yielded the picritic basalts. In both cases, melting of the harzburgitic portion of the recycled lithosphere provided the necessary Mg and Ni. The tholeiites represent mixtures from the two sources. Iceland is therefore, together with Hawaii, a clear example of melting of a complete section of recycled oceanic lithosphere.

No abstract

No abstract