Spaceborne Doppler Wind Lidars
Scientific motivation and impact
studies for ADM/Aeolus
Erland Källén
with help from
David Tan, Carla Cardinali, Paul Berrisford
Slide 1
ECMWF
Wind Lidar working group February 2010
Slide 1
Outline
 ADM/Aeolus
 Scientific motivation
 Present observing system
 Forecast error Sensitivity to Observations
 Re-analysis uncertainties
 ADM/Aeolus impact study
 Conclusions
Slide 2
Wind Lidar working group February 2010
Slide 2
Atmospheric Dynamics Mission ADM/Aeolus
Slide 3
Wind Lidar working group February 2010
Slide 3
ADM-Aeolus
•Doppler Lidar
•Aerosol and molecular
scattering
•Intermittent pulses
[H]LOS
•Only one wind component
•Dawn-dusk polar orbit
•Measurement error < 2 m/s
Slide 4
Wind Lidar working group February 2010
Slide 4
ADM/Aeolus
Slide 5
Wind Lidar working group February 2010
Slide 5
Main scientific objectives of ADM/Aeolus
Improve representation of wind field
in atmospheric analyses
 Tropics: Wind field governs dynamics
 Mid-latitudes: Intense storm developments
and meso-scale circulation systems
Numerical weather prediction
Climate sensitivity
Wind Lidar working group February 2010
Slide 6
Slide 6
Additional objectives
Aerosol information
Cloud properties
Slide 7
Wind Lidar working group February 2010
Slide 7
Outline
ADM/Aeolus
Scientific motivation
 Present observing system
 Forecast error Sensitivity to Observations
 Re-analysis uncertainties
 ADM/Aeolus impact study
Conclusions
Slide 8
Wind Lidar working group February 2010
Slide 8
Present observing system
Radiosondes
Pilot balloons and profilers
Buoys
Satellites
Aircraft data
Slide 9
Wind Lidar working group February 2010
Slide 9
Radiosondes
1 Nov 2004, ECMWF Total: 590
Slide 10
Wind Lidar working group February 2010
Slide 10
Satellite polar orbiting
1 Nov 2004, ECMWF Total: 247309
Slide 11
Wind Lidar working group February 2010
Slide 11
Aircraft data
1 Nov 2004, ECMWF Total 26219
Slide 12
Wind Lidar working group February 2010
Slide 12
Outline
ADM/Aeolus
Scientific motivation
 Present observing system
 Forecast error Sensitivity to Observations
 Re-analysis uncertainties
 ADM/Aeolus impact study
Conclusions
Slide 13
Wind Lidar working group February 2010
Slide 13
Forecast error Sensitivity to Observations
Forecast error J (“dry energy norm” ps, T, u, v)
Forecast error sensitivity
J
J x a J

→
to the analysis xa:
x a
y y x a
Rabier F, et al. 1996.
(y: observations)
Analysis solution:
x a
 K T  R 1HA
xa = xb + K(y - Hxb ) →
y
J
J
 R 1HA
y
x a The tool provides the
Compute the δJ:
J
J
FEC   J 
, y 
(y  Hxb )
y
y
Wind Lidar working group February 2010
Forecast Error Contribution
for each assimilated observation,
which can be accumulated by
observation type, subtype,
variable or level
Slide 14
Slide 14
24 H Forecast Error Contribution of GOS
GOES-Rad
MTSAT-Rad
MET 9-Rad
MET 7-Rad
AMSU-B
MHS
AMSR-E
SSMI
GPS-RO
IASI
AIRS
AMSU-A
HIRS
TEMP-m ass
DRIBU-m ass
AIREP-m ass
SYNOP-m ass
SCAT-w ind
MODIS-AMV
MET-AMV
MTSAT-AMV
GOES-AMV
PILOT-w ind
TEMP-w ind
DRIBU-w ind
AIREP-w ind
SYNOP-w ind
0
2
4
6
8
10
12
14
16
18
20
FEC %
GOES-Rad
MTSAT-Rad
MET 9-Rad
MET 7-Rad
AMSU-B
MHS
AMSR-E
SSMI
GPS-RO
IASI
AIRS
AMSU-A
HIRS
TEMP-m ass
DRIBU-m ass
AIREP-m ass
SYNOP-m ass
SCAT-w ind
MODIS-AMV
MET-AMV
MTSAT-AMV
GOES-AMV
PILOT-w ind
TEMP-w ind
DRIBU-w ind
AIREP-w ind
SYNOP-w ind
Slide 15
0
5
10
15
20
FEC per OBS %
Wind Lidar working group February 2010
Slide 15
25
30
Mass versus Wind contributions
Total
per OBS
70
60
FCE %
50
40
30
20
10
0
Conventional-wind Conventional-mass
Satellite-wind
Slide 16
Wind Lidar working group February 2010
Slide 16
Satellite-mass
Outline
ADM/Aeolus
Scientific motivation
 Present observing system
 Forecast error Sensitivity to Observations
 Re-analysis uncertainties
 ADM/Aeolus impact study
Conclusions
Slide 17
Wind Lidar working group February 2010
Slide 17
Re-analyses of zonal winds
NCEP
ERA-15
Difference
NCEP/ERA-15
Slide 18
Kistler et al., 2001
Wind Lidar working group February 2010
Slide 18
ERA-Interim
Zonal mean wind 1989-2001
>15
<-10
>30
m/s
>25
30
Slide 19
Wind Lidar working group February 2010
Slide 19
Difference ERA-Interim vs. ERA-40
Zonal mean wind 1989-2001
>2
<-4
m/s
Slide 20
Wind Lidar working group February 2010
Slide 20
Outline
ADM/Aeolus
Scientific motivation
 Present observing system
 Forecast error Sensitivity to Observations
 Re-analysis uncertainties
 ADM/Aeolus impact study
Conclusions
Slide 21
Wind Lidar working group February 2010
Slide 21
Assimilation study for ADM/Aeolus
 Assimilation ensembles for data impact assessment
 Use ensemble spread as proxy for short-range
forecast errors (background errors)
 By extension, good data reduce ensemble
spread
 DWL impact
 Radiosonde/profiler impact - provides
calibration
 Tan et al., QJRMS 133:381-390 (2007)
Slide 22
Wind Lidar working group February 2010
Slide 22
OSE
Reference
Observations
NWP-System
Verification
Assimilation/ forecast
Compare to reference
Result
Impact assessment
Real atmosphere
Assimilation/ forecast
Compare to reference
Assimilation Ensemble
Reference
Ensemble
Observations
NWP-System
An & Fc
Diagnostics
Assimilation/ forecast
Ensemble spread
Result
Calibrate
Ref. run
Assimilation/ forecast
Ensemble spread
Slide 23
Assimilation/ forecast
Wind Lidar working group February 2010
Ensemble spread
Slide 23
Impact assessment
Data impact on ensemble forecasts - zonal wind spread at
500 hPa
120 W
Thursday 16 January 2003 12UTC ECMWF Forecast t+12 VT: Friday 17 January 2003 00UTC 500hPa **u-velocity
150°W
120°W
90°W
60°W
30°W
0°
30°E
60°E
90°E
120°E
60 W
60 E
120 E
Thursday 16 January 2003 12UTC ECMWF Forecast t+12 VT: Friday 17 January 2003 00UTC 500hPa **u-velocity
150°E
120 W
0.1
0
60 E
120 E
-0.1
0.1
-0.1
0.1
2.25
60°N
60 W
-0.1
80 N
60°N
0
-0.1
-0.1
-0.1
2.6
80 N
-0.1
2.2
-0.1
-0.1
-0.1
60 N
60 N
2.1
1.95
30°N
1
0.4
40 N
40 N
0.3
30°N
1.8
20 N
20 N
-0.1
0.2
1.65
0°
0°
0
-0.3
0
1.5
1.35
30°S
20 S
20 S
30°S
1.2
0.1
-0.1
-0.2
-0.3
-0.1
40 S
40 S
-0.4
1.05
60°S
60°S
-0.1
60 S
0.9
Control
150°W
120°W
90°W
60°W
30°W
0°
30°E
60°E
90°E
120°E
0.75
150°E
80 S
60 S
0.1
0.1
-0.1
-0.1
-0.1
0.1
120 W
-0.1
-0.1
-0.1
-0.1
-0.1
60 W
0
60 E
Sondes
0.10.1
-2.2
-0.1
80 S
-2.6
120 E
Thursday 16 January 2003 12UTC ECMWF Forecast t+12 VT: Friday 17 January 2003 00UTC 500hPa **u-velocity
120 W
80 N
-0.1
-0.1
60 N
60 W
0
60 E
-0.1
120 E
-0.10.1
-0.1
-0.1
-0.1
0.1
2.6
-0.1
80 N
-0.1
-0.1
2.2
60 N
-0.1
1
-0.1
0.4
40 N
40 N
20 N
20 N
0.3
0
0
-0.1
20 S
0.2
-0.1
-0.1
20 S
0.1
-0.1
40 S
-0.4
-0.1
60 S
-0.1
-0.1
60 S
 DWLSlide
over
oceans & tropics
24
-1
-0.1
-0.1
80 S
 Radiosondes and wind
profilers over Japan,
Australia, N.Amer, Europe
-0.2
-0.3
40 S
-0.1
0.1
-0.1
120 W
-0.1
0.1
0.1
-0.1
0.1
-0.1
-0.1
ADM-Aeolus
-2.2
-0.1
-0.1
0.1 0.1
60 W
0
60 E
Wind Lidar working group February 2010
120 W
60 W
0
60 E
80 S
-2.6
120 E
Thursday 16 January 2003 12UTC ECMWF Forecast t+12 VT: Friday 17 January 2003 00UTC 500hPa **u-velocity
120 E
-1
0.1
-0.1
Slide 24
Data impact on ensemble forecasts - zonal wind spread at
200 hPa
RMS of Ensemble fcspread: RMS(fcspread_em9l_em9m_em9n_em9o) - RMS(fcspread_embd_embe_embf_embg)
Lev=200, Par=u, FcDate=20030116-20030228 12Z, Step=12
NH=-0.0390°W
SH= -0.0560°WTrop= -0.09
Eur=-0.06
NAmer=
-0.03 60°E
NAtl= 0 NPac=
-0.04
150°W
120°W
30°W
0°
30°E
90°E
120°E
150°E
Thursday 16 January 2003 12UTC ECMWF Forecast t+12 VT: Friday 17 January 2003 00UTC 200hPa **u-velocity
150°W
120°W
90°W
60°W
30°W
0°
30°E
60°E
90°E
120°E
150°E
1
0.8
2.8
60°N
60°N
30°N
30°N
2.6
60°N
60°N
30°N
30°N
0.4
2.4
2.2
1.8
0°
0°
0°
1.6
30°S
1.4
30°S
1
60°S
30°S
60°S
60°S
-1
150°W
120°W
90°W
60°W
30°W
0°
30°E
60°E
90°E
120°E
150°E
1
0.8
60°N
60°N
30°N
30°N
0°
0°
 Radiosondes and wind
0.4
0.2
profilers over Japan,
0.1
0.04
Australia, N.Amer, Europe
-0.04
0.6
-0.1
30°S
30°S
-0.2
-0.4
60°S
60°S
ADM-Aeolus
150°W
120°W
90°W
60°W
30°W
0°
30°E
60°E
90°E
120°E
Wind Lidar working group February 2010
-0.6
-0.8
-0.6
-0.8
Sondes
0.75
60°W
30°W
0°
30°E- RMS(fcspread_em9l_em9m_em9n_em9o)
60°E
90°E
120°E
150°E
RMS of 150°W
Ensemble120°Wfcspread:90°WRMS(fcspread_emi1_emi2_emi3_emi4)
Lev=200, Par=u, FcDate=20030116-20030228 12Z, Step=12
NH=-0.03
SH=
-0.04 Trop=
-0.1730°WEur=-0.020° NAmer=30°E0.02 NAtl=
-0.02 NPac=
-0.04
150°W
120°W
90°W
60°W
60°E
90°E
120°E
150°E
-0.2
-0.4
0.8
Control
0.04
-0.04
-0.1
30°S
1.2
60°S
0.2
0.1
2
0°
0.6
 DWLSlide
over
oceans and
25
tropics
-1
150°E
Slide 25
Conclusions
Wind data is lacking in present global
observing system
 Tropical analyses suffer
 Climate system re-analyses uncertain in
tropics, polar areas and stratosphere
ADM/Aeolus will provide vertical wind
profiles with global coverage
Slide 26
Wind Lidar working group February 2010
Slide 26
Thank you for your attention–
questions?
Slide 27
Wind Lidar working group February 2010
Slide 27