Abstract
A grid of geostrophic wind estimates, sampled at six‐hourly intervals over 32 years, was compared with a long time series of directly measured surface winds for Ocean Weather Station P, Tofino airport and Cape St James lighthouse. A cross‐spectral analysis revealed that coherences were high over the open ocean for periods greater than 4 d. The coherence level decreased for shorter periods. For periods less than 2.5 d the geostrophic wind could not be considered a good predictor of the open‐ocean surface winds. The coherence between the geostrophic winds and the surface winds measured at the coastal stations showed considerable variability compared to that at Station P and was strongly dependent on exposure to the open ocean. At Cape St James, with good exposure, coherences were high for periods greater than 4.0 d. At Tofino, where the surface flow is distorted by mountains, the coherences fluctuated markedly. A regression of wind components versus‐low‐pass sea‐level fluctuations was used to determine a direction corresponding to a maximum sea‐level fluctuation. It indicated that the north‐south winds were fairly coherent while the east‐west winds were incoherent.