A parametric representation of polarimetric sea surface brightness temperature suitable for inferring wind speed and direction from microwave measurements

Polarimetric passive measurements of sea surface brightness temperature have been proposed as a means of inferring wind speed and direction. A limited set of circle flight measurements of the wind direction dependence has demonstrated that there may be enough independent information in the polarimetric measurement to make this feasible. A predictive model by S. Yueh (1997) reproduces the observations closely enough that the dominant mechanisms are probably included. Optimizing the fit of this type of model with a growing data set is made difficult by the close coupling of the Yueh approach with a particular wind-wave spectral model. This makes it unclear as to how to parameterize the model, a prerequisite of the well-known optimization techniques. Here, we present an alternate formulation, which uses the Baum-Irisov (2000) model to isolate the particular properties of the wavy ocean surface which affect the radiance, in the form of six discrete parameters. Standard techniques are used to optimize the values of these parameters with respect to any large data set. The computation speed is high enough (∼1 second per emulation in the current non-optimized development code) to make the optimization procedure feasible on a data base of realistic size. The optimized model, with the associated measured error bars, becomes the ideal tool for developing algorithms for solving the inverse-problem, i.e., inferring wind speed and direction from the polarimetric brightness temperature.