Examining the wind forced velocity structure of the California Current system using observations derived from satellite remote sensing

Methodology is derived to observe mesoscale time-dependent wind-driven ocean velocities. The procedure involves the removal of a geostrophic component from "total flow" velocity observations. Total flow measuring data sets are investigated by statistical analysis, searching for theoretical characteristic signals of wind-driven flow. These signals are found in drifting buoy data, acoustic Doppler current profiler data (ADCP) data, and velocity data extracted from satellite imagery using the maximum cross-correlation technique (MCC), demonstrating that these products observe both the geostrophic and the wind-driven components of the ocean flow. Initial tests, used altimeter mean absolute dynamic topography (MADT) data as the geostrophic signal removed. This resulted in residual velocities that were dominated by vertical geostrophic shear. Methodology was then developed to combine CTD (conductivity, depth, temperature) data, which provides estimates of the geostrophic current relative to the surface, with the MADT product, to produce geostrophic velocity estimates at depth. For MCC derived observations to be used in this analysis, the depth of this product required consideration. Statistical comparison with coincident ADPC and drifter velocity observations suggest that the MCC derived velocities are characteristics of ocean currents at ∼30 m depth. This characteristic is hypothesized to be a result of the inherent average velocity observations produced by the MCC method and the nature of the variability of the ocean currents.;A 12-year time series of wind-driven velocity observations is then produced using this methodology applied to the satellite and in situ data sets. Observations generated demonstrate characteristics consistent with Ekman theory. Strong temporal agreement was found in the fluctuations wind velocity observations derived from satellite scatterometry and the wind-driven observations. Regression models, driven by the wind and the wind-driven current observations, are then used to characterize the response of the ocean to wind-forcing. The vertical response demonstrated strong linear magnitude decay with depth. The horizontal response shows complex structure that demonstrates little connection to the spatial patterns of wind velocity. To attempt to understand these patterns of wind-influence EOF and principal axis analysis are used. The results suggest that the spatial variation of frictional wind influence is strongly modulated by the shape of the coastline. Further analysis suggests that these regions of increased wind influence are driving a significant portion of the variability of the California Current. Future work will involve generating coastal altimetry observations to increase the spatial coverage of this wind-driven velocity product.