Rift relocation — A geochemical and geochronological investigation of a palaeo-rift in northwest Iceland

Abstract

A dominant process in the evolution of Iceland is the repeated eastward relocation of the spreading axis in response to westward migration of the plate boundary relative to the plume centre. Two major former rifts can be identified in western Iceland: the Snaefellsnes rift zone, which last erupted tholeiitic lavas at about 7 Ma, and an older spreading system, lava flows from which can be traced some 100 km along a SW-NE strike in the extreme northwest of Iceland. The extinction of the latter is marked by a 14.9 Ma unconformity with a laterite-lignite horizon representing a maximum 200 k.y. hiatus in the lava succession. Lavas below the unconformity dip northwest towards the older axis from which they were erupted, whereas lavas above the unconformity dip southeast towards their source in the younger Snaefellsnes axis. Thus, two nearly complete rift relocation cycles are preserved in western Iceland, each lasting about 8 m.y. as measured between rift extinction events, and for around 12 m.y. from initial propagation to extinction. In this paper we present major- and trace-element analyses, Sr, Nd and Pb isotope data, and 40Ar/ 39Ar dates on basalt samples from above and below the unconformity in northwest Iceland. The Icelandic Tertiary and Quaternary plateau basalts are remarkably homogeneous in composition, in contrast to the much more diverse compositions found in the presently active rift zone. However, basaltic lava flows beneath the unconformity in northwest Iceland show a wider range of incompatible element and radiogenic isotope ratios than do the younger plateau basalts. At least two mantle components, one depleted and the other less depleted with respect to bulk Earth, are required to explain the composition of post-15 Ma Icelandic basalt. The depleted end-member is chemically and isotopically distinct from the N-MORB source. Basalt from the northwest palaeo-rift, however, contains a significant North Atlantic N-MORB component, suggesting that depleted upper mantle can influence the composition of Icelandic basalt in a dying rift that is too far from the plume centre to be dominated by plume mantle. This may account for the periods of low magma productivity represented by troughs between the V-shaped ridges on the Reykjanes Ridge. We suggest that temporal variation in the composition of Icelandic basalt is better explained by crustal accretion and rift relocation processes than by variations in plume composition and temperature.