QuikScat Scatterometer
SeaWinds
is a rotating scatterometer (measurement of the surface
wind speed and direction over the ocean) onboard
QuikSCAT,
launched by NASA
in June 1999. QuikScat/SeaWinds is a rotating antenna
with two differently polarized emitters:
the
H-pol with incidence angle of 46.25° and V-pol with
incidence angle of 54°. The inner beam has a swath of
about 1400km, while the outer beam swath is 1800km
width. The spatial resolution of SeaWinds
σ°
(oval footprint) is of 25×35 km. The latter are binned
over the scatterometer swath into cells of 25´25
km, called Wind Vector Cell (WVC). |
 |
|
There are 76 WVC across the satellite swath, and each
contains the center of 10 to 25 measured
σ°.
The remotely wind vectors are estimated from the
scatterometer
σ°
over each WVC using the empirical model QSCAT-1
relating the measured backscatter coefficients to
surface winds. Every day, about 1.1 million 25-km ocean
surface wind vector observations are retrieved from
QuikScat measurements covering about 90% of the Earth
surface. This study deals with near real time QuikScat
data generated by National
Oceanic and Atmospheric
Administration (NOAA/NESDIS:
http://manati.orbit.nesdis.noaa.gov)
and provided by Météo-France. The work includes data
from L2A product, related to backscatter measurements,
and from L2B product related to wind vector retrievals.
Near real time (NRT) products are extracted from
Mete-France data base in BUFR format.The accuracy of
near time data are investigated through various
comparisons with off-line QuikScat data generated by Jet
Propultion Laboratory (JPL :
http://podaac.jpl.nasa.gov/
). The latter are are extracted from CERSAT/IFREMER
which is JPL mirror for scatterometer data. The main
difference between NRT and offline QuikScat products is
the spatial resolution of the backscatter coefficient (σ°).
In NRT products,
σ°
is an average of all backscatter coefficients measured
by the same beam (fore-inner, fore-outer, aft-inner,
aft-outer) and located within a given WVC. In offline
product, each
σ°
is given at its nominal spatial resolution. Both
L2B products have been calculated using the standard
scatterometer method based on the Maximum Likelihood
Estimator (MLE) (JPL, 2001). The
scatterometer retrieval algorithm estimates several wind
solutions for each wind cell. In general speaking there
are four solutions. The ambiguity removal method is then
used to select the most probable wind solution. The
latter are used in this study. To improve the wind
direction, especially in the middle of swath where the
azimuth diversity is quite poor, an algorithm called
Direction Interval Retrieval with Threshold Nudging (DIRTH)
is used too.
|
Special Sensor Micro
Wave/Imager (SSM/I)
The
SSM/I radiometers onboard the DMSP F13, F14, anf F15
satellites provide measurements of the surface
brightness temperatures at frequencies of 19.35, 22.235,
37, and 85 GHz (hereafter referred to as 19, 22, 37, and
85 GHz), respectively. Horizontal and vertical
polarization measurements are taken at 19, 37, and 85
GHz.
Only
vertical polarization is available from 22 GHz. Due to
the choice of the channels operating at frequencies
outside strong absorption lines [for water vapor] (50-70
GHz), the radiation observed by the antennae is a
mixture of radiation emitted by clouds, water vapor in
the air and the sea surface, as well as radiation
emitted by the atmosphere and reflected at the sea
surface. |
 |
|
For
estimation of the 10-m wind speed from SSM/I brightness
temperatures, we used an algorithm published by Bentamy
et al. (1999). The SSM/I wind speeds are
calculated over swaths of 1394-km width, with a spatial
resolution of 25 km 25 km. Previous studies
investigated the accuracies of the retrieved SSM/I winds
through a comparison with wind speed and direction
measured by moored buoys in several oceanic regions (Bentamy
et al. 2002).
The retrieved wind speed was calculated from brightness
temperature measurements provided by NASA Marshall Space
Flight Center (MSFC). The standard error values of SSM/I
wind speeds with respect to the buoy winds are less than
2 m/s. The bias values do not exceed 0.20 m/s.
In this study we are dealing with near real time SSM/I
brightness temperatures provided by Meteo France in BUFR
format. The latter was changed to NetCDF format
requested by several softwares used in this study. The
NRT brightness temperatures as well as retrieval winds
were compared to MSFC data. No significant differences
were found. Furtheremore, the NRT winds are compared to
Remote Sensing System (http://www.remss.com)
data too.
|
|
