The Feasibility of Estimating Ocean Surface Currents on an Operational Basis Using Satellite Feature Tracking Methods

Abstract

The feasibility of using a relatively new technique, often referred to as satellite feature tracking, for estimating ocean surface currents is described. Sequential satellite imagery is used to determine the displacements of selected ocean features over the time intervals between successive images. Both thermal infrared (IR) imagery from the Advanced Very High Resolution Radiometer (AVHRR) and ocean color imagery have been used to conduct feature tracking. Both subjective and objective techniques related to feature tracking exist to estimate surface low fields. Because of the requirement for accurate earth location and coregistration of the imagery used in feature tracking, the technique has been primarily restricted to coastal regions where landmarks are available to renavigate the satellite data. The technique is identical in concept to the approach that has been used in meteorology for the past 25 years to estimate low-level winds from geostationary satellite data. Initially, a description of the feature tracking technique is given, followed by the history of satellite feature tracking in oceanography. Next, the limitations associated with this technique are discussed. Also, only a few validation studies have been conducted to verity the results of satellite feature tracking. These studies are summarized together with some new results. Although this technique produces surface flow patterns that generally agree with the expected patterns of flow, discrepancies in speed and direction are often found when detailed comparisons with in situ observations are made. With respect to current speeds in particular, serious underestimates have occasionally been observed. A case study is given illustrating the technique for the slope water region off the U.S. East Coast. Finally, an example of a surface current analysis that is being produced experimentally for one region off the Past Coast is presented. In spite of certain limitations, this technique offers the potential for acquiring synoptic-scale coverage of the surface circulation in coastal areas on a quasi-continuous basis. Such information will be vital in supporting hydrodynamic circulation models that are currently being developed for U.S. coastal waters.