PETROGENESIS OF THE FLOOD BASALTS FORMING THE NORTHERN KERGUELEN ARCHIPELAGO - IMPLICATIONS FOR THE KERGUELEN PLUME

The thick, >20 km, crust of the Kerguelen Archipelago formed as the te ctonic setting of the Kerguelen Plume changed from the oceanic ridge-c entered location at 43 Ma to its present location beneath the Antarcti c plate. The uppermost crust is dominantly flood basalt with a thickne ss of up to 10 km. Inverse isochron Ar-40/Ar-39 ages for upper and low er lavas in a 630 m section of basalt flows from Mont Bureau with 30.4 and 29.0 Ma; Re-Os isotopic systematics are consistent with this age. Most of the lavas in two stratigraphic sections (Mont Bureau and Mont Rabouilere) from the northern part of the archipelago have Sr, Nd and Pb isotopic characteristics similar to the youngest (Upper Miocene to Pleistocene) lavas erupted in the southeast part of the archipelago, i.e. initial Sr-87/Sr-86 >0.7050. Nd-143/Nd-144 <0.5127 and Pb-206/Pb- 204 <18.3. The dominance of this isotopic signature in archipelago lav as for 30 my and its presence in similar to 40 Ma gabbros is consisten t of the Kerguelen Plume. Although this component occurs in high (>10% ) MgO alkalic lavas in the Southeast Province of the archipelago, in t hese northern sections it is confined to transitional lavas with <6% M gO. A low plume flux and extensive crustal processing are inferred. In contrast to the plume-derived lavas, similar to 15% of the flood basa lts in these sections have lower initial Sr-87/Sr-86 (to 0.70396), hig her Nd-143/Nd-144 (to 0.51289), and they have some compositional chara cteristics of plagioclase-rich cumulates, i.e. high Sr/Nd and Ba/Th an d positive Eu anomalies. However, plagioclase phenocrysts are absent i n most of these lavas; therefore a plagioclase-rich component is requi red in their source. A plausible interpretation is that plagioclase-ri ch cumulates formed in the lower oceanic crust when the Southeast Indi an Ridge was coincident with the plume at similar to 43 Ma; subsequent ly these cumulates were melted by the plume and the melts contributed to a small proportion of the flood basalts. Previously it was proposed that as the distance between the archipelago and Southeast Indian Rid ge increased, there was a systematic decrease in the proportion of mid -ocean ridge basalt (MORB)-related component in the source of archipel ago lavas. The new data show that: (1) there is not systematic tempora l trend in the proportion of MORB to plume source components and (2) t he MORB component was derived from cumulate rocks in the oceanic crust rather than as melts derived directly from the asthenosphere. Finally , there is no evidence of a continental lithosphere component in the s ource of Kerguelen Archipelago lavas.