Two-scale model results were computed by J. T. Johnson using a code
based on:
Yueh, S. H., ``Modeling of wind direction signals in polarimetric sea
surface brightness temperatures," IEEE Transactions on Geoscience and
Remote Sensing, vol. 35, pp. 1400-1418 Nov 1997.
No atmospheric or foam contributions are included in these calculations
(i.e. F_r in Yueh's equation (6) is set to zero.) The Durden-Vesecky
spectrum with parameter a_0=0.008 is used as described in Appendix
D of the reference, and as described in equation (59) the parameter
b_0 is set to match the long and short wave spectra at k_j=2 rads/m.
The long wave slope pdf is assumed to be Gaussian, and the
hydrodynamic modulation function of the reference in equation
(11) is included.
Input parameters to the code at 19.35 GHz were as follows:
Frequency 19.35 GHz
Sea Water Dielectric Constant 29.04+i35.55
Surface temperature 283 K
Friction Velocity u [0.106 0.191 0.312 0.464 0.626 0.798 0.981] m/s
Resulting U_19.5 [2.98 5.98 8.99 12.00 15.00 17.99 20.99] m/s
Values of d parameter [0.6084 0.5742 0.5040 0.4347 0.3802 0.3361 0.2993]
Incidence angle Theta 55 degrees
Azimuth angle Phi 0 to 180 degrees in 15 degree steps
(0 degrees indicates upwind observation)
k_d long/short wave either 120 or 40 rads/m
separation parameter
Short wave spectrum set to 8100 rads/m
zero above
At 37 GHz parameters were the same as above, except
Frequency 37 GHz
Sea Water Dielectric Constant 14.34+i24.17
k_d long/short wave ether 230 or 77 rads/m
separation parameter
Short wave spectrum set to 15500 rads/m
zero above
Files include tilting over the long wave slope distribution.
Tilted results were obtained by a Gauss-Hermite quadrature
over the long wave slope distribution, using 4x4 points.
Note the two cutoff values chosen are around k0/3 and k0/10
respectively. The latter value is outside the "resonance"
effects of the second order SPM kernel, and therefore could
potentially be put forward as more reasonable.
First and second azimuthal harmonic coefficients were extracted
from these data using a least squares fit algorithm.
These values are plotted in the *.pdf files and compared with
the JPL WindRAD empirical model.
The results show only a minor effect of the cutoff parameter,
although horizontal brightnesses are somewhat better matched by
the k0/10 choice. In general the two-scale model used here
(with the doubled Durden-Vesecky spectrum) shows reasonable
agreement with measured data but there are noticeable discrepancies.
The data files contain data in 7 columns. The data are:
u (m/s) | Theta (deg) | Phi (deg) | Th (K) | Tv (K) | U (K) | V (K) |
Please contact me to discuss how our codes compare.
I expect there may be some internal parameters not listed
above that may cause some differences between codes.
We can develop more detailed comparisons of internal
code variables to assist in locating the sources of
any differences observed. Other cases can also be tried as
the study progresses.
--------------------------------------------------------------------------
Joel T. Johnson, Associate Prof. ------------- The Ohio State University
E-mail: johnson@ee.eng.ohio-state.edu -------- Dept. of Electrical Eng.
Voice: (614) 292-1593 or 1606 ------------- 205 Dreese Laboratories
FAX: (614) 292-7297 ------------- 2015 Neil Ave
URL:http://eewww.eng.ohio-state.edu/~johnson - Columbus, OH 43210