Recent geochemical studies of convergent margin volcanoes have been strongly influenced by the concept that calc-alkaline andesites are primary magmas formed by partial melting of wet mantle peridotite above subducted oceanic crust (Fountain, 1979; Rogers and Ragland, 1980; Suzuki and Shiraki, 1980). Partial melting of the peridotite is thought to occur through access of volatiles released by dehydration of the descending oceanic crust (McBirney, 1969; Best, 1975), or of volatile-rich silicate melts generated by partial fusion of the crustal portion of the subducted plate (Ringwood, 1975). Such an interpretation is critically dependent upon the assumption that the composition of melts erupted at the surface has been little modified by crystal fractionation. This paper examines the petrogenesis of Pleistocene-Holocene andesites from the Garibaldi Lake area of southwestern British Columbia on the light of observed mineralogical and chemical variations. These andesites provide evidence that crystal fractionation processes have significantly modified magma compositions.

Evolution of British Columbia's continental margin during the past 10 m.y. has been dominated by a complex pattern of plate interactions (Riddihough, 1977; Keen and Hyndman, 1979). The present plate-tectonic setting of southwestern British Columbia is depicted in Figure 1. The Queen Charlotte fault system, which parallels the edge of the continental shelf from southeastern Alaska to a poorly defined “ridge-trench-fault” triple junction just north of Vancouver Island, marks the locus of right-lateral transcurrent motion between the Pacific and North American plates. En echelon spreading axes (Juan de Fuca Ridge, Explorer Ridge, Dellwood and J. Tuzo Wilson Knolls) offset by short transform segments form the Juan de Fuca-Pacific and Explorer-Pacific plate boundaries. The Explorer plate, detached from the Juan de Fuca plate as the spreading ridge system rotated to become more nearly orthogonal to the Queen Charlotte fault, has been shown to be moving independently during the past 3 m.y. (Riddihough, 1977). Left-lateral motion between the Juan de Fuca and Explorer plates occurs across the Nootka tranform fault system, which extends north-easterly from the northern end of the Juan de Fuca Ridge to Nootka Island off the west coast of Vancouver Island (Keen and Hyndman, 1979). South of the Nootaka fault, the boundary between the Juan de Fuca and North American plates is a zone of convergence; seismic data show that the oceanic lithosphere dips at 5° toward the margin and reaches depths of 4 to 5 km before disappearing under Vancouver Island (Riddihough, 1979). The absence of a clearly defined eastward-dipping seismic zone and the lack of a major bathymetric trench at the foot of the continental slope have contributed to uncertainty as to whether subduction has taken place recently. Riddihough and Hyndman (1979), Ando and Balazs (1979), and Riddihough (1979), however, have reviewed relevant geological, geodetic, and geophysical data and concluded that the Juan de Fuca plate is now continuously underthrusting western North America. The plate motion is largely aseismic, but a few Benioff-type events at about 50 to 60 km beneath southern Georgia Strait (Keen and Hyndman, 1979) provide some weak constraints on the present position of the subducted plate. Whether the Explorer plate is now being subducted beneath the continent is less certain Riddihough (1977) suggested the possibility that the oceanic lithosphere east of Dellwood Knolls has become locked to the North American plate.

First Page Preview

First page PDF preview
You do not currently have access to this article.