A numerical study of ocean polarimetric thermal emission

Abstract

A numerical model for polarimetric thermal emission from penetrable ocean surfaces rough in two directions is presented. The numerical model is based on Monte Carlo simulation with an iterative version of the method of moments (MOM) known as the sparse matrix flat surface iterative approach (SMFSIA), extended to the penetrable surface case through a numerical impedance boundary condition (NIBC) method. Since the small U/sub B/ brightnesses obtained from ocean surfaces (usually less than 1.5 K, or 0.5% of a 300-K physical temperature) require extremely accurate simulations to avoid large errors, a parallel version of the algorithm is developed to allow matrix elements to be integrated accurately and stored. The high accuracy required also limits simulations to near flat surface profiles, so that only high-frequency components of the ocean spectrum are modeled. Variations in nadir polarimetric brightness temperatures with spectrum low- and high-frequency cutoffs show the Bragg (or shortwave) portion of the spectrum to contribute significantly to emission azimuthal signatures, as predicted by the small perturbation or composite surface approximate theories. Quantitative comparisons with approximate methods show perturbation theory to slightly overestimate linear brightness temperatures, but accurately predict their azimuthal variations, while physical optics (PO) significantly underestimates both linear brightness temperatures and their azimuthal variations. Further simulations with the numerical model allow sensitivities to ocean spectrum models to be investigated and demonstrate the importance of an accurate azimuthal description for the ocean spectrum.