Petrochemistry of the British Old Red Sandstone volcanic province

Abstract

The lavas and minor intrusions associated with the sediments of the Old Red Sandstone were extensively sampled over a broad area of Scotland, northern England and the north of Ireland. The main areas of outcrop are Lorne, Glencoe and Ben Nevis (the SW Highlands); the Sidlaw, Ochil and North Fife Hills (the North Midland Valley); Ayrshire and the Pentland Hills (the South Midland Valley); the Cheviot Hills; St. Abb's Head and Shetland. Most of the samples collected have been examined petrographically, and 597 have been analysed for 10 major and 17 trace elements. The phenocryst compositions of many of the lavas has also been investigated by microprobe. The large number of samples to be analysed required the development of a rapid and accurate system of analysis by X-ray fluorescence spectrometry, and a wide range of data processing and data handling computer programs, described in the Appendices.

In general, the rocks have strong affinities with modern calcalkaline volcanic suites, although those from Shetland and from Straiton, Ayrshire, show some characteristics transitional to tholeiitic types. Although the rocks of each outcrop area are petrographically and chemically diverse, the variation is generally inexplicable by fractional crystallisation processes, except for Shetland and (possibly) the Cheviot Hills. Chemical variation in most areas can only be explained by complex contamination models.

Rocks with >100 ppm Ni are unusually abundant for calc-alkaline suites, and may be found in most outcrop areas. These rocks show pronounced geographical variation in their concentrations of Sr, P, light rare earth elements, Ba, K, Y and, to a lesser extent, Nb and Sc. All except Y and Sc show a marked increase in a traverse from the South Midland Valley to the SW Grampian Highlands, and a similar relationship also exists for rocks poorer in Ni. This geographical variation is similar to that seen, in modern subduction-related calcalkaline suites, and constitutes strong evidence that the rocks of the province were genetically related to a subduction zone. Little satisfactory evidence is available for the age of the Scottish Lower Old Red Sandstone, and it is believed that- the rocks were deposited during the Silurian, rather than the Lower Devonian as has been generally accepted, prior to final closure of the Iapetus Ocean. The rocks of the Cheviot Hills are thought to have been erupted a little later, after final closure. The bulk of the volcanic rocks were however probably directly related to the subduction responsible for closure.

In order to explain the geographical variation in concentrations of the above elements in terms of a single subduction zone, it is necessary to postulate a major change in strike of the zone from ENE in southern Scotland to nearly N in the North Sea. This strike change, in conjunction with almost east-west relative plate motions can explain the contrast in end-Silurian deformation styles between Britain and Scandinavia. It is believed that motion was largely transcurrent on the British-Irish sector of the subduction zone. The possible origins of the geographic variation in incompatible element concentrations are constrained by the failure of Rb and Zr to correlate with other incompatible elements, and by the decrease in Y with increasing depth to the inferred Benioff zone. These variations are best explained by a model involving (i) generation of melts in the vicinity of the subducted lithosphere (ii) their ascent through the mantle and the zone-refining of dispersed incompatible elements (but not those located in minor crystalline mantle phases such as phlogopite, ilmenite or apatite). The concentration of an incompatible element in a particular magma is then mainly a-function of the path length to the surface and the distribution of minor phases along that path.