Off-rift Axis Channelized Melt and Lithospheric Metasomatism along Mid-ocean Ridges—A Case Study from Iceland on the Limits of Melt Channelling

Abstract

Abstract Magmatism in Iceland is classically explained by the interaction of the mid-Atlantic ridge with the Iceland plume. The growth of Iceland through time is the result of volcanic activity at the rift axis. However, Holocene volcanism (0–11.5 ka) is not restricted to the rift zone (RZ) but also occurs off-axis, specifically in the western Snæfellsnes Volcanic Belt (SVB) and in the Southern Flank Zone (SFZ). The rift and off-axis postglacial volcanic zones are separated by a gap of ~60–80 km. While the volcanic activity of the SFZ seems correlated with the actual location of the Iceland plume, the plume relation to the SVB is uncertain. To address the origin and relationships between rift and off-rift magmas, we present new data from two transects perpendicular to the Reykjanes rift. The lavas in the SVB are characterized by transitional to alkaline compositions, with elevated incompatible trace element content. In contrast, the RZ volcanic rocks have tholeiitic compositions with trace element signatures slightly more enriched than Mid-Ocean Ridge Basalt (MORB). Rift and off-rift Iceland lavas are all characterized by positive Ba and Nb anomalies, particularly in alkaline rocks. Tholeiitic and alkaline lavas show distinct differentiation sequences, with the main difference being the delayed crystallization of plagioclase in the fractionating assemblage of alkaline magmas. We apply these sequences to calculate primary magma compositions, which are then used to constrain melting conditions. Geochemical modelling indicates that Iceland rift and off-rift magmas can be produced from a peridotitic mantle source if lithospheric processes are involved. We demonstrate that recycled crust in the form of pyroxenite is not required to generate Snæfellsnes alkaline lavas. The low solidus temperature and high productivity of pyroxenite favour early and more extensive melting producing primary magmas that are not sufficiently enriched in incompatible trace elements to explain the compositional variation of Snæfellsnes magmas. An alternative mechanism to involve pyroxenite in the source of Snæfellsnes lavas relates to the hybridization of recycled oceanic crust with peridotite, but such reacted pyroxenite requires specific compositions to reproduce the Ba and Nb anomalies. As an alternative, we suggest that Snæfellsnes alkaline lavas are the result of channelized low-degree melts produced on the periphery of the melting column at distances exceeding 65 km from the ridge axis. These melts accumulate and percolate into the lithosphere producing amphibole ± phlogopite-bearing hydrous cumulates. Incongruent melting of these cumulates via renewed magmatic activity and melt-peridotite reaction can reproduce the alkaline compositions observed in the SVB, including the Ba and Nb anomalies. Numerical simulations of melt extraction below mid-ocean ridges suggest that low-degree melts produced as far as ~65 km from the central ridge axis rise vertically to the base of the lithosphere and are then focussed towards the ridge axis in decompaction channels. We propose that these melts interact with hydrous cumulates previously formed during the development of decompaction channels at the lithosphere–asthenosphere boundary and acquire specific Ba and Nb anomalies. The mixing of these distal enriched melts with more depleted melts extracted from the central part of the melting regime explains the composition of RZ lavas. Alkaline lavas observed in the SFZ show numerous analogies with the Snæfellsnes magmas, suggesting that similar lithospheric processes control their chemistry. The high thermal regime in Iceland and thick lithosphere explains the difference between Icelandic tholeiite and typical MORB. Our results highlight the importance of mantle dynamics below mid-ocean ridges and lithospheric interaction to produce off-axis magmatism with enriched alkaline signatures.