Geochemical processes during long-distance lateral magma transport within upper crust

Recent geophysical observations demonstrate that magma is transported long distances from volcanic centers by dyke propagation (e.g., Miyakejima volcano: Geshi et al., 2002; Hachijojima volcano: Ishizuka et al., 2008). However, studying the physical and chemical properties of upper crustal fissures and intrusions in active volcanic systems is difficult. We have chosen to utilize the extensive dyke networks of the British Tertiary Volcanic Province that are exposed within the upper crust to investigate the processes occurring during magma transport. Lineations within dykes formed by bubble and crystal alignment indicate that the direction of magma flow gradually changed with distance from the volcanic center. Adjacent to the Mull volcanic center the dominant flow direction is near vertical to high-angle, while beyond c. 50km the flow direction is essentially horizontal. These observations indicate that the dykes extending southeast from Mull were fed by laterally-transported magma from this volcanic center. Geochemically, the dykes show two distinct trends: one representing magmatic differentiation and the other crustal assimilation. Assimilation is most clearly recognized in the correlations between radiogenic isotopes and major elements (e.g., SiO2 vs. 87Sr/86Sr, 143Nd/144Nd), and in the relationships between trace element ratios or ΔNb. Potential assimilants in the studied area are the lower crust, which is likely to be dominated by Lewisian gneiss, and the upper crust, featuring schists of sedimentary origin. Most of the dykes lie along a distinct isotopic trend from the Mull Plateau Group (MPG) lavas; these being a likely equivalent composition to the initial magmas before displacement along the fissures. Isotopically, these lavas reflect a contribution of lower crustal material (with low143Nd/144Nd, 206Pb/204Pb and Th/Ce) to the MORB-type asthenospheric mantle. In contrast, the dykes, except for some proximal samples, can be explained by addition of upper crustal material to a similar asthenospheric source. Contribution of this crustal material to the dykes appears to vary with distance, with the maximum143Nd/144Nd decreasing, and the minimum Th/Ce increasing away from Mull. Magmas with lesser crustal contribution (i.e., high 143Nd/144Nd, low Th/Ce) only occur close to the Mull volcanic center, and all the distal dykes have strong influence of upper crustal assimilation. These data imply that the dyke magma gradually assimilated upper crust on its journey away from Mull. A lateral variation in extent of crustal assimilation such as this has not been recognized in the Mackenzie Dyke Swarm (Baragar et al., 1996). Variable Ti/Zr between dykes implies that they formed during different stages of Mull volcanism, reflecting the equivalent variation in Ti/Zr in the Mull lavas. Accordingly, isolating geochemical signatures constrained by high-precision 40Ar/39Ar dating of dyke rocks enable us to correlate individual intrusions with known effusive events.