Publishers version - DTU Orbit

Figure 197: Scattering geometry sketch from Naderi et al. (1991).
A single σ0 measurement is not sufficient to determine both the wind speed and direction.
More measurements are required from different azimuth angles and different polarizations,
obtained with more than one beams. During the wind inversion process, the set of wind
speed and direction that maximizes the probability of the measured σ0 is determined using
a maximum-likelihood estimation (MLE) method. Typically, more than one solutions with
extreme MLE values are obtained, known as “ambiguities”. These ambiguities correspond
to almost the same wind speed but different wind directions. The ambiguity that has the
maximum value is chosen as the “best” estimate. The geometry of each σ0 measurement
or “footprint” depends on the specifications of the scatterometer. Many nearly-simultaneous,
space-collocatedσ0 measurementsare averaged in Wind VectorCells (WVC). The dimensions
of the WVC depend on the mission specifications.
GMFs have evolved during the decades of scatterometer applications, as more measurements
become available and more validation studies are performed. The SASS-2 GMF was
developed by Wentz et al. (1984), using the statistics from 3 months of SASS measurements
and a mean global wind speed from a climatology. The NSCAT mission resulted in a new
GMF, based on the correlation of the radar backscatter with modelled winds from ECMWF
and SSMI wind speeds, as described in Wentz and Smith (1999). The SSM/I GMF was based
on a model for the brightness temperature of the ocean and the atmosphere above, which is
calibrated using buoy and radiosonde data as described in Wentz (1997).
ERS-1 required a new type of GMF due its different operating frequency compared to
SASS. Data collected during several campaigns related σ0 from air-borne instruments with in
situ observations from research ships and buoys. This resulted in the pre-launch GMF known
as CMOD2 (Offlier, 1994). The operational ERS-2 GMF was CMOD4, developed by Stoffelen
and Anderson (1997), using satellite derived σ0 and 10 m winds from the ECMWF analysis.
The CMOD5 function (Hersbach et al., 2007) was released to correct for deficiencies in
the CMOD4 version, fitting measurements of extreme backscatter and winds, obtained from
aircraft and in situ data. CMOD5.N, described in Hersbach (2010), is tuned to wind at 10 m
above the surface assuming neutral atmospheric stratification.
16.3 Equivalent Neutral Wind
GMFs are typically derived using open ocean buoy measurements and relating those to radar
backscattermeasurements.Once the empeirical relationshiphas been established,it is applied
to the scatterometer σ0 values to derive the wind speed and direction through the wind
inversion process. As a convention, σ0 values are related to the wind at 10 meters above the
sea surface assuming a neutral atmospheric stratification, i.e. the Equivalent Neutral Wind
(ENW) (Liu & Tang , 1996).
u = u∗
κ
z
ln −ΨM
z0
(352)
298 DTU Wind Energy-E-Report-0029(EN)