A wide-angle, multichannel seismic study of the continental lithosphere
Previous experiments to record seismic data at wide angle on the continental shelf have generally been unsuccessful in determining velocity structure in the lower crust; either the lines were too short or shot-receiver density too sparse to identify lower crustal arrivals. In contrast, deep normal incidence profiles show good structural resolution in the crust and uppermost mantle. This dissertation describes a new sea-bottom multichannel instrument which was developed to record datasets containing closely spaced traces to improve the resolution of reversed wide-angle experiments on the continental shelf. The Pull-Up Multichannel Array (PUMA) is a 1200 m, 12 channel, hydrophone array for remotely recording seismic data on the sea-bed. It consists of 12 short hydrophone sections linked by 100 m long passive sections. A pressure case is attached at one end in which recording electronics, cassette tape recorders and a battery power supply are housed. The PUMA is deployed in less than 200 m of water from a research ship and moored to buoys for recovery. The instrument, which was successfully used in an experiment west of Lewis, Outer Hebrides, U.K., was specifically designed to determine a well constrained velocity structure for the crust and uppermost mantle over part of the BIRPS WINCH deep, normal incidence profile. Data recorded by the PUMA show a high signal to noise ratio and it is easy to correlate phases across the record section and to monitor changes in amplitude because traces are closely spaced. A velocity structure for the continental crust and uppermost mantle has been devised using amplitude modelling. The model is interpreted to show that: the upper crust consists of Lewisian gneiss metamorphosed in the amphibolite facies, and contains a low velocity zone which is probably granitic in composition; the middle crustal layer has virtually no velocity gradient; and the lower crust, represented by second arrivals, contains a high velocity gradient and probably consists of granulites. The Moho is at 27.0 ± 0.5 km depth. This is shown to be a layered boundary and it is suggested that this layering is caused by crustal underplating. There is little velocity gradient in the uppermost mantle. This model shows good agreement with the BIRPS WINCH section, although the Moho appears to coincide with the top rather than the base of a band of reflections at 8.3 s two-way travel time on unmigrated WINCH data.