Measuring Directional Wave Spectra and Wind Speed with a Scanning Radar Altimeter

4 Pages·1999·0.21 MB·English

by  Walsh, E. J.

At,r" Pacje 3/5
q75c37
Measuring DlrectlonM Wave Spectra and Wind Speed with aScanning Radar Altimeter
E..1. Walsh _,D. Vandemark, C. W. Wright
NASA Goddard Space Flight Center, Code 972
Laboratory for Hydrosphec|c Processes
Wallops Flight Facility, Wallops Island, VA 23337, LISA
303-497-6357, fax [phone], [email protected]
_presently on assignment at NOAA Envlromnental Technology Laboratory, Boulder, Colorado
R. N. Swift, J. F. Scott, D. E. Hines
EG&G Services, Inc.
N159, Wallops Flight Facility
Wallops Island, VA 23337, USA
757-824-1019, fax [phone], swift @osb.wff nasa.soy
Fig. 1 shows the geometry for the NASA Scanning Radar variation. But during the 1998 hurricane season the SRA was
Altimeter (SRA). It transmits a 8-ns duration pulse at Ka-band installed on one of the NOAA P-3 hurricane hunter aircraft.
(8.3 ram) and measures time of flight as it scans a 1° (two-way) The NOAA minimum height for the hurricane flights was 1500
beam from left to right across the aircraft ground track. The m, significantly higher than the SRA had previously flown.
most re_2ent configuration determines the surface elevation at And the flight patterns were such that the waves were generally
6..I.points spacM at uniform angular intervals of about 0.7' propagating across the aircraft ground track rather that along it.
across a swath whose width is about 0.8 times the aircraft T_e largest waves wer¢ generally obsctwed near the edge of the
altttude. The system generates these raster lines of the surface - swath during the hurricanue flights.
topography beneath the aircraft at about a 10 Hz rate. In post- A 3-dimensional simulation wa.s carried out and it was
flight processing the SRA wave topographic data _e determined that. even with a beamwidth as small as l*, wave tilt
transformed with a two.-dimensit_nal FFT and Dopple-r- modulation of the radar cross section within the antenna
corrected to produce directional wave spectra fI, 21. footprint cart cause a significant, _ystematic increase in the
The SRA is not absolutely caJibrated in power, but by apparent m_gnhude of wave components in the cross-track
measuring the relative fall-off of backscatter with increaxing direction. The higher the altitude and the shorter the
incidence angle, the SRA can also determine the mean square wavelength, the larger the effect.
slope (rnas) of the sea surface, a surrogate for wind speed [3]. Fig. 2 shows the geometry for a 1500 m aircraft height and
For the slol:m-dependent specular point model of radar sea ocean waves of 50 m length and 1.25 rn height. The two top
surface scattering [4, 5, 6], an expression approximated by a panels indicate the SRA 8-ns pulse appruaching the sea sttrl'a_:e
geomeme optics form, for the relative variation with incidence
angle of the normalized backscatter radar cross section would
be
O_,,.i= Se_:4:0 exp ( -tan: 0 / ross ) it)
where 0 is the off-nadir incidence angle
Ifthe logarithm or the range-corrected hackscattered power
is plotted against tan 1 0, the resulting curve_ approximate
straight lines whose slope is inversely proportiond to ross. But
\
when a straight line is fitted through actual data, the resulting
value of rrtss is dependent Ontile span of itlcidence angles used
[3]. This problem can be eliminated by fitting aquadratic to the
data. The ratio of the hnear to the quadratic coeffictent is
/
found to be determined by the rrt,;s, so a general model for
/
relating ross to the non-Gaussian variation of backscattered
power with incidence angle can be determined. This nx>del was /
u_ed to determine the distortions expected in the SRA
measurements of sea surface tolmgraphy to be used in the
computation of directional wave spectra. o.ah_/ /
The SRA had traditionally flown either upwave or downwave
to acquire topographic data in the direction of most rapid Fig 1. NASA Scanning Radar Altimeter (SR,\) geometry.
!;,._t t/y" _tP Ldt;et'Jet :]t00; 303 497 GISf; Ai)r" 'Ju !1 :.1/; P,._{je ,1.,5
with the two-way halF-power antenna beamwidth indicated by
the dashed lines and the O.! power beamwith indicated by the
dotted lines. The :Jolid line indical_ the boresight, with awave
20 crest at the bore,sight in the top panel and a wave trough in the
I I second panel. Spatial filtering by the antenna tuotprint would
: I 1 be expected for waves this short.
£
t In the bottom two panels of Fig. 2. the boresight incidence
• :_ angle is22° off-nadir, the nominal edge of the SRA swath. In
0 .... ' ' " this instance the simulation indicated that the measured wave
-- height would actually be enhanced. The wavc.x shown are
_I: 40 _ I I t i propagating in the cross-track direction, and when the antenna
boresight is at a crest (third panel), the near side of the crest
t i back.scatters more power because theincidence angle issmaller
I I than for the far side of the crest. The centroid of the
20 I I
! 1
0 ..i. _L__. _
-20 0 20 40
CROSS-TRACK DISTANCE FROM NADIR (m) _ -1l- _-_ "
,to[ , \ ', '. "'_ L ..... , L .
t \ ,, ., o loo .2oo 300
20 \\ "'. i
'1 \ q,
o |
I_ l t l , L
\\ _ \\ ".
20 =. \\ -1 "__
0 100 200 300
.___'. \\ "'. j .t. 1
WAVE PHASE (DEGREES)
Fig. 3. Sea surface elevation me'.,_ured by the SRA at nadir (.)
I 1 t I _ and 4_(+), 8_(x), 12"(o). and 16'(*) off-nadir, normalized by the
580 600 620 640
peak of the actual surface elevation for waves of 50 m length
CROSS-TRACK DISTANCE FROM NADIR (m) and !.25 m height propagating in either the along-track
directinn (azimuth = 90". top panel) or the cross-track direction
Fig. 2. SRA 8-ns pulse approaching _ea surface from 1500 m (azimuth --0". bottom panel) for various phase angles of the
height at nadir (top two panels') and 22* off-nadir (bottom two wave train at the SRA antenna boresight. The "assumed sea
panels) for waves of 50 m wavelength and 1.25 m height surface height variation is indicated by the solid sinusotdai _[[!
curve, which has been normalized by its 0 625 m anapl_tudc) .......
prapagating i,xthe cross-track direction.
backscattered power is biased tosho_ter range. The short range
az=90
is ascribed to the boresight angle and the resulting calculated
1(
----o o o o o o o' o--O o o O
crest position is higher. The reverse occurs at wave troughs
× x x x × x x x x x x x
('Fig. 2, bottom), where'the erroneously larger range, ascribed
4- + + + + + 4- + + + + +
'0.5 to the antenna boresight, makes the trough appear deeper.
Fig. 3summarizes wave topography distortions for various
phase angles of the ocean waves at the antenna boresight and
0 I I _1 1 ,1 various off-nadir angles for 50-m waves propagating in the
along-track (top panel) and cro,_s-track (Ix_ttom panel)
az = 60
directions. An ross of 0.08 was assumed, corresponding to _25
1( m/s wind speed, The situation worsens al lower wind speed.
Fig. 4 summarizes (for ross = 0.08) the variation of the ratio
of meaxured to actual wave height for various ocean wave
0.5 lengths and azimuth propaga6on directions relative to the cross-
track plaale of incidence, All wave lengths are assumed to be
40 times the respective wave heights, for constant wave
steepness approximately corresponding to full development.
At nadir, waves propagating in all directions have amplitudes
reduced by the same amount. The reduction is over 60% for 50
m wavelength and negligible for 400 m wavelength.
For waves propagating in the cross-track direction, shorter
waves are significantly enhanced at the edge of the swath. In
"1-
general, the off-nadir incidence angle for the transition between
_o o
ILl _ L I _ .. I I + + wave amplitude reduction and enhancement depends on both
the wavelength and the azimuthal direction of propagation.
4- 1.¢.,x
tu 1.5 4- x" Because the SRA simultanexmsly measures the backscattered
>
az=30 • _,._.x,." power variation when it mea._ure,_ the wave topography, these
distortions can be correcled in the directional wave spectra.
X X
REFERENCES
•
[1] E.J. Walsh, D.W, Hancock, DE. Hines, R.N. Swift, and
J.Scott, "An observation of the dh'ectional wave spectrum
I | .-J.---_ ___ | + evolution from shoreline to fully developed," J.Phys.
4-
+ Oceanogr., vol. t9, pp. 670-690, 1989.
[2] E.J. W'alsh, L.K. Shay, H.C, Graber. A. Guillaume. D.
Vandemark, DE. Hines, R.N Swift, and J.F. Scott,
1.5 az:0 • hh "'Observations of surface wave-current interaction during
o
o o o SWADE." The Global Ammsphere and Ocean System,
vol. 5, pp, .q9-124, Gordon and Breach Science Publishers
SA, 1996.
[31 E J. Wal._h. D Vandemark, C,A. Friehe, SP Run_, D.
/
0.5 Khelif, R,N Swift, and J.F. Scott, "Measuring sea surface
mean square slop with a 36 Gl-lz scanning radar odtin'_ter,
J. Geophys. Res,, vol.103, pp. 12.587-12,601, lune 1998,
0 I I I . 1, I [4] D.E. Bardck, "Rough surface scattering based on the
0 5 10 15 20 25
specular point theory," IEEE Trans. Antennas Propag..
OFF-NADIR INCIDENCE ANGLE
vo1 AP-16, pp 449-454. 1968.
[51 D.E. Barrick, "Wind dependence of quasi-specular
Fig. 4. Wave height measured by the SRA, normalized by the microwave sea scatter." IEEE Trans. Antennas Propag,
actual wave height, for various propagation directions of the vol. AP-22. pp. 135-136. 1974.
wave train from along-track (azimuth =90°,top panel) tocross- [61 G.R. V,'tlenzuela. "Theories for the interaclion of
track (azimuth : 0", bottom panel) for ocean wavelengths of50 electromagnetic and oceanic wave,_ -a review", Bound.
m (dashed), 75 m (+). 100 m (x), 20) m (oJ, ,and400 m() for Layer Meteor, vol. 13. pp. 61.85, 1978,
various incidence angles of the SRA antenna boresight
J