Document 13636678

advertisement
The assimilation of surface-sensitive microwave
sounder radiances at ECMWF
Enza Di Tomaso and Niels Bormann
European Centre for Medium-Range Weather Forecasts, Reading, UK, E-mail: Enza.DiTomaso@ecmwf.int and Niels.Bormann@ecmwf.int
Active assimilation of MHS channel 5 (the lowest-peaking in the water vapour band) over land
Impact on the humidity field
• The assimilation of just one extra MHS channel over land has a relevant impact on the mean humidity analysis in both
the 2006 and 2011 period (see Figure 2 and Figure 3).
120°W
60°W
0°
60°E
(kg/m2)
120°E
8
120°W
60°W
0°
60°E
(kg/m2}
120°E
1.104
2.229
8
16
16
60°N
16
60°N
60°N
24
1.6
1.2
8
40
32
40
32
48
0°
40
32
0°
40
24
0°
-0.4
42
16
48
36
24
48
54
30
30
-0.4
-0.6
30°S
24
-0.8
18
-1
12
12
60°S
6
6
-1.2
6
6
-1.4
-2
120°W
MHS channel 5 first guess departures
120°W
90°W
60°W
30°W
0°E
30°E
60°E
90°E
120°E
0°
60°E
120°W
120°E
Figure 2 Mean difference between the TCWV analysis in the channel
5 experiment and the control experiment averaged over August 2006.
The red dots indicate the location of GPS stations.
• MHS channel 5 departures from model estimates do not show too large
land/sea differences (see Figure 1).
150°W
60°W
-0.2
48
18
36
30
18
12
60°S
0.2
0°
48
30
30
24
-1.534
60°S
-1.6
8
0.6
42
42
36
42
36
30°S
18
8
24
30
42
-0.8
30°S
-1.2
8
30
12
30
4236
48
42
30°N
0.4
36
18
36
24
16
24
24
16
16
60°S
18
30
42
0.4
48
40
32
32
40
32
8
24
18
36
48
40
12
18
24
6
24
30°N
0.8
40
48
30°S
18
30°N
48
6
6
6
24
48
60°N
0.8
12
12
32
24
30°N
1
6
2
16
16
8
Here we discuss the impact in NWP of enhancing the use of MHS observations
over land. A set of 3 month experiments were run with/without the
assimilation of MHS channel 5 (channel 5 experiment/control experiment):
– over low orography, skin temperature >278 K,
– and with dynamically retrieved emissivities [1]
(i.e. retrieved from observations at 89GHz, surface emission and reflection).
Experiments were run in two periods (following the work in [2]):
– summer (July, August, September) 2006,
(i.e. the AMMA [3] period, with the additional assimilation of MERIS data);
– summer (June, July, August)/winter(January, February, March) 2011
(with a richer observational system than in the 2006 experiments).
60°W
0°
60°E
120°E
Figure 3 Mean difference between the TCWV analysis in the channel
5 experiment and the control experiment averaged over August
2011.
150°E
(K)
30°N
30°N
0°N
0°N
30°S
30°S
60°S
60°S
150°W
120°W
90°W
60°W
30°W
0°E
30°E
60°E
90°E
120°E
2.77
1.08
0.96
0.84
0.72
0.60
0.48
0.36
0.24
0.12
0.00
-0.00
-0.12
-0.24
-0.36
-0.48
-0.60
-0.72
-0.84
-0.96
-1.08
-4.89
Validation against GPS TCWV
• The effect on the TCWV is significant both in terms of correlation and mean TCWV difference at a few AMMA stations
(see Figure 4).
• The systematic drying over Central Africa during the whole diurnal cycle is consistent with the measurements at the IGS
[4] station in Gabon in the 2006 period (see Figure 5). Both the channel 5 experiment and the control experiment show
variations of TCWV during the day which follow well what is measured by the GSP stations.
0.92
Correlation
0.9
Figure 1 Mean differences between NOAA-19 MHS channel 5 observations
and model estimates in the channel 5 experiment before bias correction.
0.88
0.86
0.84
0.82
Fit to MHS observations
0.8
• The assimilation of channel 5 improves the fit to MHS channel 4
observations (see Figure 6).
• The small increase in the standard deviation for channel 5 in the channel
5 experiment is due to the additional data over land for which the estimate
for the surface emission is still poorer than that over sea.
a
0.78
42
0
0.94
150°E
41.5
-0.5
-1
-1.5
-2
-2.5
1
2
3
(a)
4
GPS stations
5
6
7
Mean TCVW (kg/m^2)
60°N
TCWV difference (kg/m^2)
60°N
41
40.5
40
39.5
39
1
2
3
(b)
4
GPS stations
5
6
7
Figure 4 Correlation (a) and average differences (b) between the experiment analysis in the
channel 5 experiment (blue asterisks) and in the control experiment (black circles) versus the
GPS TCWV at six AMMA stations (Djougou (Benin), Niamey (Niger), Gao (Mali), Tamale
(Ghana), Ouagadougou (Burkina Faso), Tombouctou (Mali)), and at the Gabon IGS station.
0
3
6
9
12
15
Time of the day (UTC)
18
21
24
Figure 5 TCWV diurnal cycle at Gabon station for
the channel 5 experiment (green crosses), the
control experiment (blue asterisks), and GPS
measurements (red pluses).
b
Channel frequency [GHz]
183±7
Fit to MERIS TCWV
• The assimilation of channel 5 improves the fit to MERIS TCWV in part of Central Africa (see Figure 7 and Figure 8), and
greater number of MERIS observations are assimilated in the same area.
183±3
150°W
120°W
90°W
60°W
30°W
30°E
0°E
60°E
90°E
120°E
(K)
150°E
60°N
60°N
30°N
30°N
0°N
0°N
30°S
30°S
60°S
60°S
183±1
1.2
1.3
1.4
1.5
1.6
Std. dev. [K]
1.7
1.8 -1.0
-0.5
0.0
Mean [K]
0.5
ch5 exp
Background departure (o-b)
ctl exp
Bias correction of obs
1.0
Figure 6 MetOp-A MHS brightness temperature departure statistics in terms of standard
deviation (a) and mean (b) of the differences between channel 3, 4 and 5 (i.e. 183+/-1 GHz,
183+/-3 GHz and 183+/-7 GHz) observations and model estimates for the channel 5 experiment
(black) and the control experiment (red) for the Southern Hemisphere. The sample of channel
5 observations is 15% greater in the channel 5 experiment than in the control experiment.
150°W
120°W
90°W
60°W
30°W
30°E
0°E
60°E
90°E
120°E
150°W
7.07
2.43
2.16
1.89
1.62
1.35
1.08
0.81
0.54
0.27
0.00
-0.00
-0.27
-0.54
-0.81
-1.08
-1.35
-1.62
-1.89
-2.16
-2.43
-7.80
150°E
120°W
90°W
60°W
0°E
30°W
60°E
30°E
90°E
120°E
(K)
150°E
60°N
60°N
30°N
30°N
0°N
0°N
30°S
30°S
60°S
60°S
150°W
Figure 7 Mean differences between MERIS TCWV observations and
model estimates in the channel 5 experiment in the 2011 period.
120°W
90°W
60°W
0°E
30°W
60°E
30°E
90°E
120°E
150°E
7.20
2.45
2.18
1.91
1.63
1.36
1.09
0.82
0.54
0.27
0.00
-0.00
-0.27
-0.54
-0.82
-1.09
-1.36
-1.63
-1.91
-2.18
-2.45
-6.66
Figure 8 Mean differences of the MERIS TCWV observations
departures from model estimates between the channel 5
experiment and the control in the 2011 period.
Conclusions MHS channel 5 will be operationally assimilated at ECMWF also over land from May 2012 onwards. The impact on the forecast skill of assimilating MHS channel 5 is neutral to positive for different
variable and forecast ranges, which is consistent with a relatively small impact of these observations on the analysis.
Active assimilation of MHS water vapour channels over sea ice
a
Results
150°W
120°W
90°W
60°W
30°W
0°E
30°E
60°E
90°E
120°E
MHS channel 3 and 4 are good candidates to be assimilated at ECMWF
also over sea ice, while the active assimilation of MHS channel 5 requires
a better estimation of surface emissivity and cloud screening.
60°N
60°N
30°N
30°N
0°N
0°N
30°S
30°S
60°S
60°S
150°W
120°W
90°W
60°W
30°W
0°E
30°E
60°E
90°E
120°E
183±3
150°E
(K)
Conclusions
b
183±7
• Dynamic emissivities retrieved at 157GHz provide reasonably
good emissivities, without any frequency-dependent correction,
for the high peaking water vapour channels (see Figure 9).
• The additional observations over sea ice and cold sea improve
the fit to MHS observations (with the exception of channel 5 in
the Southern Hemisphere) (see Figure 14).
Channel frequency [GHz]
A set of data assimilation experiments were run with/without the active
assimilation of MHS channel 3 and 4 (the high peaking channels in the water
vapour band) over sea ice (sea ice experiment/control experiment):
– dynamic emissivities [4] retrieved at 157GHz (channel 2) are used
for the other MHS channels over sea ice,
– the use of MHS water vapour channels is extended also over cold
sea (i.e. skin temperature < 278 K).
5.25
1.34
1.19
1.04
0.89
0.74
0.59
0.45
0.30
0.15
0.00
-0.00
-0.15
-0.30
-0.45
-0.59
-0.74
-0.89
-1.04
-1.19
-1.34
-4.01
150°E
Figure 9 Mean differences between NOAA-18 MHS channel 4 observations
and model estimates in the sea ice experiment before bias correction.
183±1
1.2
1.3
1.4
1.5 1.6 1.7
Std. dev. [K]
1.8
1.9 -1.0
-0.5
0.0
Mean [K]
0.5
sea ice exp
Background departure (o-b)
ctl exp
Bias correction of obs
1.0
Figure 10 MetOp-A MHS brightness temperature departure
statistics in terms of standard deviation(a) and mean (b) of the
differences between channel 3, 4 and 5 (i.e. 183+/-1 GHz, 183+/-3
GHz and 183+/-7 GHz) observations and model estimates for the sea
ice experiment (black) and the control experiment (red) for the
Southern Hemisphere. The sample of channel 4 (5) observations is
60% (25%) greater in the sea ice experiment than in the control.
References
Acknowledgments
[1] Karbou, F., E. Grard, F. Rabier (2005), Microwave land emissivity and skin temperature for AMSU-A and -B assimilation over land. Q. J. R. Meteorol. Soc., 132, 2333-2355.
[2] Karbou, F. et al. (2010), Global 4DVAR Assimilation and Forecast Experiments Using AMSU Observations over Land. Part II, Wea. Forecasting, 25, 2036.
[3] Bock, O. et al.(2008), The West African Monsoon observed with ground-based GPS receivers during AMMA. J. Geophys. Res., 113, D21105.
[4] Dow, J.M. et al. (2009), The International GNSS Service in a changing landscape of Global Navigation Satellite Systems, J. Geodesy, 83, 191198.
Enza Di Tomaso was funded by the EUMETSAT fellowship programme.
Olivier Bock is gratefully acknowledged for his help with the AMMA
GPS data.
Download