Modeling of Polarimetric Thermal Emission from the Sea Surface

The WindSAT program led by the Naval Research Laboratory plans to place a polarimetric radiometer in orbit in 2003. The radiometer will operate at an incidence angle of approximately 53 degrees, and has channels at 6.8, 10.7, 18.7, 23.8, and 37 GHz. The primary goal of WindSAT is to demonstrate improvements in wind speed and direction remote sensing over the sea surface which can be achieved through the use of polarimetric radiometry.

Physical models for emission from wind generated sea surfaces are of interest in the WindSAT program to assist in developing the most effective wind retrieval methods. However, the problem is complicated by the fact that several effects can be important contributors to sea surface emission, including emission from long ocean waves, resonant emission from short sea waves, contributions of foam on the sea surface, modulation of short waves by long waves, and direct and reflected emissions from the atmosphere. A complete physical model should take all of these contributions into account.

Several models for emission from the sea surface itself (i.e. the long wave and short wave components discused above) have been developed in the literature. One model is based on a ``two-scale" theory, in which a theory for emission from the short wave part of the surface is ``tilted" over the slope distribution of the underlying long wave surface. Evaluating a sea surface brightness temperature from the two scale theory requires calculation of a quadruple integral. A second theory is based on physical or geometrical optics, and should capture emission from long length scale features in the surface spectrum. Calculation of physical optics predictions requires a double integral over the surface slope probability density function (PDF). All theories of sea surface emission are functions of the description of the sea surface used, whether it be the directional spectrum (as in two scale theories) or the slope PDF (as in optical theories.) Differing surface spectra or slope PDF's will cause changes in predictions, so reasonable models for these statistics are also important.

Because these theories are somewhat complex, there are potentially many places to introduce problems when writing a code. It can therefore be very useful to compare results from different codes as a check on implementations and the accuracy of the integration methods used. This web page is intended to provide a central location for posting model results from different organizations who have implemented sea surface emission theories, and to allow intercomparison of results for validation and other purposes.

Two-scale code results (single wind speed):

19 GHz, no tilting
19 GHz, with tilting
37 GHz, no tilting
37 GHz, with tilting

Two-scale code results (multiple wind speeds and harmonic plots):

19 GHz, cutoff 120 data file
19 GHz, cutoff 120 harmonic plots versus windspeed
37 GHz, cutoff 230 data file
37 GHz, cutoff 230 harmonic plots versus windspeed
19 GHz, cutoff 40 data file
19 GHz, cutoff 40 harmonic plots versus windspeed
37 GHz, cutoff 77 data file
37 GHz, cutoff 77 harmonic plots versus windspeed

PO code results:

19 GHz
37 GHz

Small Slope Approximation:

19 GHz
37 GHz

Related publications of RS Theory group at Ohio State:

web page

ElectroScience Laboratory home page:

web page

Joel T. Johnson (