Current meter data from a trial of moorings located on the continental rise offshore of Point Reyes-Point Arena and from a mooring site located about 250 km offshore and to the northwest of the array are analyzed. The continental rise moorings were deployed from October 1984 to July 1985, and each had five meters at depths from about 150 to about 3560 m. The offshore mooring was deployed from September 1982 to August 1983 with six meters positioned from about 150 to 3800 m. Velocity components were highly vertically coherent in the upper 600 m and were in phase for periods longer than about 10 days; temperature tended to be less coherent. There was some horizontal coherence in the continental rise array. Coherence at 9--13 days between the deep temperature (T at about 3200 m) measurements at M1 and M2 (located about 120 km offshore and separated by about 100 km) coincided with very high, in phase, coherence in the northward velocity component v.

Although there was no coherence between the first baroclinic modal components at M1 and M2, the eastward barotropic velocity component was coherent at periods of 9--11 days, and the northward barotropic velocity component was coherent at 8--15 days and 4.5--6 days. The triad provided a complete data set only at 350 m: the highest coherence was between out of phase v components (and also T) at M2 and M3 (about 100 km seaward of M2) at periods of 45--90 days, with the dominant time domain empirical orthogonal function (EOF) representing strong horizontal shear between M2 and M3. At periods of 9--13 days, there was bottom and surface intensification in the nearshore kinetic energy spectra, suggestive of a superposition of a barotropic mode with the first two baroclinic modes and partially consistent with the presence of topographic Rossby waves. The vertical structure at M2 differed from that at W10 (the offshore mooring), where the deep (about 3000 m) flow tended to be stronger than that near the continental slope and rotary behavior was apparent at 3000 m. The barotropic modal energy was higher at W10 than at M2, but the first baroclinic modal energy was comparable at the two sites.

At M2, when they were coherent, the barotropic and first baroclinic modes enhanced each other in the upper water column (and let to reduced flow at deeper levels); at W10, these modes were mainly out of phase (and so enhanced the flow at deeper levels). These differences, which may be spatial rather than temporal, increase the complications of initialization of predictive models from mass field data alone. There are some indications of current and temperature variability due to fluctuating winds; for example, the northward components of wind stress and barotropic current were coherent at all locations at periods of 8--9 days. At W10 there was also coherence between the eastward components of barotropic current and wind stress at 16--19 days. On the basis of frequency domain EOFs, some of the current variability was forced by large-scale (i.e., about 1000 km) wind stress at periods of between 5 and 10 days; however, less than half the current variability was demonstrably related to the wind stress. At M2 and M3 at 150 m, large-amplitude (about 1 ¿C) temperature variations, which are sustained for several weeks, are influenced by alongshore flow convergence and offshore jets. Coherence between T at 350 m and wind stress curl is suggestive of Ekman pumping at M3, but the temperature was also related to alongshore convergence and offshore flow.