Impact of whitecap coverage derived from a wave
model on the assimilation of radiances from
microwave imagers
Louis-François Meunier1
Stephen English2
Peter Janssen2
1 Météo-France
2 European
and CNRS, CNRM-GAME/GMAP
Centre for Medium-Range Weather Forecasts
27th March 2014
Impact of the whitecap coverage on microwave
imagers observations
Microwave imagers are very
sensitive to the surface emission
and consequently the surface
emissivity
In RTTOV, over seas, FASTEM is
used
Emissivity calculation in FASTEM
1
Flat specular sea water
2
Small-scale correction
3
Large-scale correction
4
Foam partial cover
5
Azimuthal angle correction
L-F. Meunier, S. English, P. Janssen
19Ghz, Ts =283K, θ=53°
0.65
0.60 ev
0.55
Flat sea
0.50
0.45
0.40
0.35
0.30 eh
0.25
0 5 10 15 20 25 30
10m wind speed (m/s)
Whitecap coverage, MW imagers
27/03/2014
1 / 13
Impact of the whitecap coverage on microwave
imagers observations
Microwave imagers are very
sensitive to the surface emission
and consequently the surface
emissivity
In RTTOV, over seas, FASTEM is
used
Emissivity calculation in FASTEM
1
Flat specular sea water
2
Small-scale correction
3
Large-scale correction
4
Foam partial cover
5
Azimuthal angle correction
L-F. Meunier, S. English, P. Janssen
19Ghz, Ts =283K, θ=53°
0.65
0.60 ev
0.55
Flat sea
0.50
Small scale
0.45
0.40
0.35
0.30 eh
0.25
0 5 10 15 20 25 30
10m wind speed (m/s)
Whitecap coverage, MW imagers
27/03/2014
1 / 13
Impact of the whitecap coverage on microwave
imagers observations
Microwave imagers are very
sensitive to the surface emission
and consequently the surface
emissivity
In RTTOV, over seas, FASTEM is
used
Emissivity calculation in FASTEM
1
Flat specular sea water
2
Small-scale correction
3
Large-scale correction
4
Foam partial cover
5
Azimuthal angle correction
L-F. Meunier, S. English, P. Janssen
19Ghz, Ts =283K, θ=53°
0.65
0.60 ev
0.55
Flat sea
0.50
Small scale
0.45
Large scale
0.40
0.35
0.30 eh
0.25
0 5 10 15 20 25 30
10m wind speed (m/s)
Whitecap coverage, MW imagers
27/03/2014
1 / 13
Impact of the whitecap coverage on microwave
imagers observations
Microwave imagers are very
sensitive to the surface emission
and consequently the surface
emissivity
In RTTOV, over seas, FASTEM is
used
Emissivity calculation in FASTEM
1
Flat specular sea water
2
Small-scale correction
3
Large-scale correction
4
Foam partial cover
5
Azimuthal angle correction
L-F. Meunier, S. English, P. Janssen
19Ghz, Ts =283K, θ=53°
0.65
0.60 ev
0.55
Flat sea
0.50
Small scale
0.45
Large scale
foam p. cover
0.40
0.35
0.30 eh
0.25
0 5 10 15 20 25 30
10m wind speed (m/s)
Whitecap coverage, MW imagers
27/03/2014
1 / 13
Impact of the whitecap coverage on microwave
imagers observations
Microwave imagers are very
sensitive to the surface emission
and consequently the surface
emissivity
In RTTOV, over seas, FASTEM is
used
Emissivity calculation in FASTEM
1
Flat specular sea water
2
Small-scale correction
3
Large-scale correction
4
Foam partial cover
5
Azimuthal angle correction
L-F. Meunier, S. English, P. Janssen
19Ghz, Ts =283K, θ=53°
0.65
0.60 ev
0.55
Flat sea
0.50
Small scale
0.45
Large scale
foam p. cover
0.40
0.35
0.30 eh
0.25
0 5 10 15 20 25 30
10m wind speed (m/s)
Whitecap coverage, MW imagers
27/03/2014
1 / 13
Whitecap coverage parametrisation
The energy dissipation flux (Φoc )
is calulated by the wave model
and can be used to diagnose the
whitecap fraction
ECWAM: 02/10/12 12UTC
60°N
40°N
20°N
0°
20°S
40°S
60°S
180°
6
5
4
3
2
1
0
Whitecap f. (fws) (%)
Bottom friction set aside, wave
breaking is the dominant energy
dissipation mechanism for ocean
waves
120°W
60°W
0°
60°E
120°E
180°
Whitecap fraction (W) (see [Kraan et al., 1996])
E is the total wave energy,
W=
−Φoc
γρw gωp E
L-F. Meunier, S. English, P. Janssen
ωp is the peak angular frequency,
γ is the fraction of the total wave energy
dissipated per whitecapping event
Whitecap coverage, MW imagers
27/03/2014
2 / 13
Assimilation experiments setup
Two experiments:
fwas (control): FASTEM uses the current formulation of the whitecap
fraction (based on 10 m wind speed)
fw76 (new): FASTEM uses the whitecap fraction calculated by the
wave model
Both experiments are based on cy38r2 of IFS. Microwave imagers
are asssimilated in allsky conditions
The two experiments have ran from 1 October 2012 00UTC to the
14 December 2012 12UTC. The results shown later on are
computed from the 10 October 2012 00UTC in order to allow the
variational bias correction to adjust
Heavy precipitation areas are excluded from the statistics shown
later on
L-F. Meunier, S. English, P. Janssen
Whitecap coverage, MW imagers
27/03/2014
3 / 13
TMI 11Ghz: First guess departures w.r.t. windspeed
TMI/TRMM. Ch #1 (11V).
Blk=Ctl(fwas), Red=New(fw76). Bias & StdDev(dash).
3.0
TMI/TRMM. Ch #2 (11H).
Blk=Ctl(fwas), Red=New(fw76). Bias & StdDev(dash).
3.5
3.0
First guess departure (K)
First guess departure (K)
2.5
2.0
1.5
1.0
0.5
0.0
1.5
1.0
0.5
0.0
0.5
600000
400000
200000
00 2 4 6 8 10 12 15
20
10m windspeed (new) (m/s)
L-F. Meunier, S. English, P. Janssen
2.5
2.0
25
0.5
600000
400000
200000
00 2 4 6 8 10 12 15
20
10m windspeed (new) (m/s)
Whitecap coverage, MW imagers
27/03/2014
25
4 / 13
TMI 37Ghz: First guess departures w.r.t. windspeed
8
4
3
2
1
0
6
4
2
0
1
600000
400000
200000
00 2 4 6 8 10 12 15
20
10m windspeed (new) (m/s)
L-F. Meunier, S. English, P. Janssen
TMI/TRMM. Ch #7 (37H).
Blk=Ctl(fwas), Red=New(fw76). Bias & StdDev(dash).
First guess departure (K)
First guess departure (K)
TMI/TRMM. Ch #6 (37V).
Blk=Ctl(fwas), Red=New(fw76). Bias & StdDev(dash).
5
25
600000
400000
200000
00 2 4 6 8 10 12 15
20
10m windspeed (new) (m/s)
Whitecap coverage, MW imagers
27/03/2014
25
5 / 13
SSMIS: First guess departure bias (abs. difference)
37GHz V:
L-F. Meunier, S. English, P. Janssen
Mean first guess dep. (K)
Mean first guess dep. (K)
19GHz V:
SSMIS/All. Ch #13 (19V). abs(New=fw76) - abs(Ref=fwas).
0.40
60°N
0.32
40°N
0.24
0.16
20°N
0.08
0°
0.00
0.08
20°S
0.16
0.24
40°S
0.32
60°S
0.40
180° 120°W 60°W
0°
60°E 120°E 180°
SSMIS/All. Ch #16 (37V). abs(New=fw76) - abs(Ref=fwas).
0.40
60°N
0.32
40°N
0.24
0.16
20°N
0.08
0°
0.00
0.08
20°S
0.16
0.24
40°S
0.32
60°S
0.40
180° 120°W 60°W
0°
60°E 120°E 180°
Whitecap coverage, MW imagers
27/03/2014
6 / 13
SSMIS: First guess departure bias (abs. difference)
37GHz V:
L-F. Meunier, S. English, P. Janssen
Mean first guess dep. (K)
Mean first guess dep. (K)
19GHz V:
SSMIS/All. Ch #13 (19V). abs(New=fw76) - abs(Ref=fwas).
0.40
60°N
0.32
40°N
0.24
0.16
20°N
0.08
0°
0.00
0.08
20°S
0.16
0.24
40°S
0.32
60°S
0.40
180° 120°W 60°W
0°
60°E 120°E 180°
SSMIS/All. Ch #16 (37V). abs(New=fw76) - abs(Ref=fwas).
0.40
60°N
0.32
40°N
0.24
0.16
20°N
0.08
0°
0.00
0.08
20°S
0.16
0.24
40°S
0.32
60°S
0.40
180° 120°W 60°W
0°
60°E 120°E 180°
Whitecap coverage, MW imagers
27/03/2014
6 / 13
Tentative conclusion
Overall the impact of the new parametrisation is negative at low
frequiencies but slightly positive at higher frequencies
The need for a bias correction predictor based on the wind speed
is less with the new parametrisation (not shown)
The amount of whitecap generated by the new parametrisation in
closed sea, some sheltered area and shallow waters seems to be
excessive
Lack of validation data for the whitecap fraction parametrisation
This led us to calculate a whitecap fraction retrieval (see
[Anguelova and Webster, 2006])
L-F. Meunier, S. English, P. Janssen
Whitecap coverage, MW imagers
27/03/2014
7 / 13
Whitecap fraction retrieval method
FASTEM-5 calculates the sea emissivity as follows:
e = (1 − W) efoamfree + W efoam
How to retieve W ?
efoamfree and efoam are calculated within FASTEM-5
e can be estimated from the observed brightness temperature
(see [Prigent et al., 2005], [Karbou et al., 2005])
This method has been applied on an assimilation experiment similar to
our “control” experiment (except that the wind speed predictor was
turned off in the bias correction). The same period is considered (from
10 October 2012 00UTC to 14 December 2012 12UTC)
L-F. Meunier, S. English, P. Janssen
Whitecap coverage, MW imagers
27/03/2014
8 / 13
Mean retrieved whitecap fraction at low frequency
60°N
40°N
20°N
0°
20°S
40°S
60°S
180°
Retrieved whitecap f.(%)
SSMIS/All. Ch #13 (19V). Expver=fwl6.
2.4
2.1
1.8
1.5
1.2
0.9
0.6
0.3
0.0
120°W
L-F. Meunier, S. English, P. Janssen
60°W
0°
60°E
Whitecap coverage, MW imagers
120°E
180°
27/03/2014
9 / 13
60°N
40°N
20°N
0°
20°S
40°S
60°S
180°
2.4
2.1
1.8
1.5
1.2
0.9
0.6
0.3
0.0
120°W
60°W
0°
60°E
120°E
SSMIS/All. Ch #13 (19V). Expver=fwl6.
180°
Diagnosed whitecap f. (%)
60°N
40°N
20°N
0°
20°S
40°S
60°S
180°
Retrieved whitecap f.(%)
Mean retrieved whitecap fraction at low frequency
SSMIS/All. Ch #13 (19V). Expver=fwl6.
2.4
2.1
1.8
1.5
1.2
0.9
0.6
0.3
0.0
120°W
L-F. Meunier, S. English, P. Janssen
60°W
0°
60°E
Whitecap coverage, MW imagers
120°E
180°
27/03/2014
9 / 13
Retrieved whitecap fraction using TMI
6
4
WAM
11V
19V
11H
19H
The emissivities prescribed
by FASTEM-5 during the
retrieval (efoamfree and efoam )
may not be correct...
FASTEM-5 foam emissivity:
2
0
20 2 4 6 8 10 12 15
20
10m windspeed (m/s)
L-F. Meunier, S. English, P. Janssen
25
Foam Emissivity
Whitecap fraction (%)
8
How to explain the
differences ?
TMI/TRMM. Expver=fwl6.
1.00
0.98
0.96
0.94
0.92
0.90
Whitecap coverage, MW imagers
10Ghz V
19Ghz V
22Ghz V
37Ghz V
10Ghz H
19Ghz H
22Ghz H
37Ghz H
0 10 20 30 40 50 60
Satellite Zenith Angle (degrees)
27/03/2014
10 / 13
Conclusion
Our first results show:
At low frequencies, a degradation of the standard deviation of the
first guess departures when the new parametrisation is used. At
higher frequency the impact is neutral to slightly positive.
A positive impact on the bias correction.
Perspectives
The whitecap parametrisation should be improved along the
coasts, in the closed sea and shallow water areas.
The emissivity model of the foam should be modified in order to
correct the inconsistencies between H and V polarisations.
The retrieved whitecap fraction may be used to tune the whitecap
fraction parametrisation.
L-F. Meunier, S. English, P. Janssen
Whitecap coverage, MW imagers
27/03/2014
11 / 13
Bibliography I
Anguelova, M. D. and Webster, F. (2006).
Whitecap coverage from satellite measurements: A first step
toward modeling the variability of oceanic whitecaps.
Journal of Geophysical Research: Oceans, 111(C3):n/a–n/a.
Karbou, F., Prigent, C., Eymard, L., and Pardo, J. (2005).
Microwave land emissivity calculations using AMSU
measurements.
Geoscience and Remote Sensing, IEEE Transactions on,
43(5):948–959.
Kraan, G., Oost, W. A., and Janssen, P. A. E. M. (1996).
Wave Energy Dissipation by Whitecaps.
Journal of Atmospheric and Oceanic Technology, 13:262–267.
L-F. Meunier, S. English, P. Janssen
Whitecap coverage, MW imagers
27/03/2014
12 / 13
Bibliography II
Prigent, C., Chevallier, F., Karbou, F., Bauer, P., and Kelly, G.
(2005).
AMSU-A Land Surface Emissivity Estimation for Numerical
Weather Prediction Assimilation Schemes.
Journal of Applied Meteorology, 44(4):416–426.
L-F. Meunier, S. English, P. Janssen
Whitecap coverage, MW imagers
27/03/2014
13 / 13