An experiment to study buoyancy driving of the poleward flowing coastal current off the west coast of Vancouver Island was carried out from June to November 1984. The experiment consisted of two hydrographic surveys and a three-dimensional instrumented array that included 51 current/temperature sensors distributed on 18 moorings. Moorings were maintained across the mouth of the Strait of Juan de Fuca (to monitor the fluctuating buoyant source/sink) and along three cross-shelf transects roughly 40, 90, and 130 km north of the mouth of the strait. To detect possible incoming propagating coastal-trapped waves (CTWs), moorings were also located ~10 and ~30 km south of the mouth of the strait.

The main thrust of the experiment was to attempt to identify signals specifically associated with buoyant outflow from the Strait of Juan de Fuca. Four such signals were identified. (1) A seasonally varying mean poleward flow over the mid and inner shelf. (2) Velocity and temperature fluctuations associated with buoyant plumes over the shelf. The plumes, which occur at roughly monthly and bimonthly intervals, are driven by anomalously strong freshwater pulses from the strait, and the associated fluctuations in the along-isobath component of velocity accounted for about 20% (10%) of the variance during summer (fall) in the region within about 50 km of the mouth of the strait. (3) Positive temperature anomalies (~3 ¿C) related to the cessation of strait outflow. These anomalies, which are strongest at and south of the mouth of the strait, occur during periods of regional southwesterly winds and have a south-to-north lag consistent with poleward advection of the pre-existing (usually cold) plume water at about 30 km d-1. (4) Phase modification of the wind-driven CTW and local wind-driven velocity signals across the mouth of the strait; a south to north phase difference of about 90¿ is observed across the mouth of the strait at low (<0.12 cpd) frequencies during both summer and fall. Statistical analysis as well as application of a first-mode wind-driven CTW model demonstrated that the dominant variance over the Vancouver Island shelf is related primarily to the first-mode wind-driven CTW (summer, ~55%) or local wind forcing (fall, ~80%). This being the case, and also because of the rugged shelf topography in the area, the analysis proved insufficient to identify definitively more subtle results of plume and strait processes such as the local generation of CTWs. ¿American Geophysical Union 1991