The detection and tracking of oceanic frlmts and eddies is problematic using infrared sensors such as .AVHRR due to persistent cloud cover over many frontal systems. Passive microwave data can overcome this problem although their application is limited by coarse spatial resolution. A novel approach has been developed to extrapolate highresolution predictions of front location from a-priori estimates of current velocity and acceleration derived from sequences of infrared imagery during periods of cloud-free coverage. The prediction is subsequently matched to the smoothed signature of the frontal boundary in passive microwave datasets to achieve sub-pixel co-location accuracy.