HR: 15:25h
AN: OS23B-08    [Abstracts]
TI: Obtaining Accelerating Geostrophic Flows with a Single HF Radar
AU: * Barrick, D
EM: don@codaros.com
AF: CodarOcean Sensors, Ltd, 1914 Plymouth St., Mountain View, CA 94043 United States
AU: Fitzgerald, R M
EM: rfitzgerald@utep.edu
AF: Physics Dpt.-University of Texas at El Paso, 500 W. University Ave., El Paso, TX 79968 United States
AB: A single HF radar maps only the radial components of surface currents in a polar coordinate system. Normally two or more radars must view the same point on the sea to produce a total horizontal current vector, leading to the desired two-dimensional maps of horizontal flows. In many cases, it is not possible to employ two radars, for example, on an oil platform. Attempts to estimate total horizontal surface flows date back three decades, but none have been successful. There has always been an implied assumption; for example, no horizontal divergence, which fails in common upwelling regions that are of great interest. We describe some successful approaches that we have thus far tested with simulations where the input is known. These involve using the dominant three terms of the Navier-Stokes equations as a constraint in fitting modes to the radial maps. These three terms are the acceleration term; the pressure-gradient term (surface slope); and the Coriolis term. The reason to expect this to work is the success of satellite altimetry in deducing deep-water geostrophic circulation: in that case only two terms are used, i.e., acceleration is neglected. This becomes a first-order partial differential equation in a polar Lagrangian system. It cannot be linearized because of the time/space scales of the fields. Nonetheless, we recover accurately the unknown azimuthal field at any step forward in time if an initial azimuthal field at time t = 0 is known. Our most recent investigation explores methods to estimate the unknown field when a complete initial condition is not available. We report on several techniques that appear promising based on simulations, among them the Kalman filter. Our goal has been to identify the mathematical and physical reasons that cause a non-unique solution to this least-squares minimization problem, and thence to find a minimal number of independent measurements that can stabilize and optimize the solution to this important initial value problem.
DE: 4255 Numerical modeling (0545, 0560)
SC: Ocean Sciences [OS]
MN: Fall Meeting 2005