Assimilating and determining Assimilating and determining the impact of sea surface winds the impact of sea surface winds measured by WindSat/Coriolis data measured by WindSat/Coriolis data in the Global Forecast System in the Global Forecast System Li Bi Tom Zapotocny James Jung Michael Morgan 31 May 2006
32
Embed
Li Bi Tom Zapotocny James Jung Michael Morgan 31 May 2006
Assimilating and determining the impact of sea surface winds measured by WindSat/Coriolis data in the Global Forecast System. Li Bi Tom Zapotocny James Jung Michael Morgan 31 May 2006. Overview. WindSat is a polarimetric microwave radiometer Measures ocean surface wind speed and direction - PowerPoint PPT Presentation
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Assimilating and determining the Assimilating and determining the impact of sea surface winds impact of sea surface winds
measured by WindSat/Coriolis data measured by WindSat/Coriolis data in the Global Forecast Systemin the Global Forecast System
Li Bi
Tom Zapotocny
James Jung
Michael Morgan
31 May 2006
OverviewOverview
WindSat is a WindSat is a polarimetric polarimetric microwave radiometermicrowave radiometer
Measures ocean Measures ocean surface wind speed surface wind speed and directionand direction
Launched on 6 Launched on 6 January 2003January 2003
circular orbitcircular orbit 830 km altitude830 km altitude 98.7 degrees of inclination98.7 degrees of inclination 1800 Local Time of the 1800 Local Time of the
Ascending Node (LTAN)Ascending Node (LTAN) about 14.1 orbits per dayabout 14.1 orbits per day
1800 LTAN for 1800 LTAN for validation with validation with QuikSCATQuikSCAT
WindSat instrument WindSat instrument has 1025 km swath has 1025 km swath widthwidth
http://www.seaspace.com/news
http://www.seaspace.com/news
Projected CapabilitiesProjected Capabilities
Demonstrate the capability of Demonstrate the capability of polarimetric microwave radiometry to polarimetric microwave radiometry to measure the ocean surface wind vector measure the ocean surface wind vector from space.from space.
How ocean surface physics change with How ocean surface physics change with wind and boundary layer conditions.wind and boundary layer conditions.
WindSat will aid with forecast of short-WindSat will aid with forecast of short-term weather, issuing timely weather term weather, issuing timely weather warnings and gathering general climate warnings and gathering general climate data.data.
How WindSat can measure How WindSat can measure wind speed and direction?wind speed and direction?
Wind roughening the surface of the Wind roughening the surface of the ocean causes an increase in the ocean causes an increase in the brightness temperature of the brightness temperature of the microwave radiation emitted from the microwave radiation emitted from the water’s surface.water’s surface.
Multiple Multiple frequencies and polarizations allow for simultaneous retrievals of different surface and atmospheric parameters.
Downward traveling Downward traveling atmospheric and cold atmospheric and cold space radiation that is space radiation that is scattered back from scattered back from the ocean surfacethe ocean surface
http://www.ofcm.gov
Principles of OperationPrinciples of Operation
WindSat measures full polarizations. WindSat measures full polarizations.
rclc
h
v
TT
TT
T
T
V
U
Q
I
Is4545
Band, GHz Polarizations Horizontal Spatial Resolution, km
6.8 V, H 40 x 60
10.7 V, H, +45, -45, L, R
25 x 38
18.7 V, H, +45, -45, L, R
16 x 27
23.8 V, H 12 x 20
37.0 V, H, +45, -45, L, R
8 x 13
CLW Retrievals
18.7,28.8 and 37GHz
WV Retrievals
18.7,23.8 and 37GHz
Wind Speed Retrievals
10.7,18.7,23.8 and 37GHz
SST Retrievals
10.7GHz
Wind Direction Retrievals
10.7, 18.7 and 37GHz3rd and 4th Stokes
biclwi TaL
biwvi Tav
biui Tau
bisstisst TaT
5
1 ,
2mod,,
)(
))((
imeasViU
VUmeasViU
TVar
TTJ
NOAA WindSat Wind NOAA WindSat Wind Vector Retrieval Vector Retrieval Algorithm – Ver 0Algorithm – Ver 0
http://cioss.coas.oregonstate.edu/
Sample picture of WindSat wind speed from a Sample picture of WindSat wind speed from a preliminary wind vector retrieval algorithm.preliminary wind vector retrieval algorithm.
Proposed workProposed work Work with the JCSDA (Joint Center for Work with the JCSDA (Joint Center for
Satellite Data Assimilation) to evaluate the Satellite Data Assimilation) to evaluate the impact of assimilating WindSat data in the impact of assimilating WindSat data in the NCEP GFS (Global Forecast System) model. NCEP GFS (Global Forecast System) model. A November 2005 version of the SSI and GFS A November 2005 version of the SSI and GFS were used and run at T254L64.were used and run at T254L64.
Compare the forecast impact with QuikSCAT Compare the forecast impact with QuikSCAT data from the same time period. QuikSCAT data from the same time period. QuikSCAT data have provided a positive impact on data have provided a positive impact on forecastsforecasts
Data time period: 1 Jan – 15 Feb 2004Data time period: 1 Jan – 15 Feb 2004
First Goal First Goal
Run GFS with QuikSCAT (cntrl254)Run GFS with QuikSCAT (cntrl254)
Run GFS without QuikSCAT (noqscat254)Run GFS without QuikSCAT (noqscat254)
Study forecast impact for QuikSCAT windsStudy forecast impact for QuikSCAT winds
22 2
1 1 1( )( ) ( ) /{( ) }
NN N
i ii i i ii i i
C AD A C AFI 100
N N N
Error in experimentError in controlError in control
The 500 hPa geopotential height anomaly correlation at day 5 for 20-80N (7 Jan–15 Feb 2004)
The 500 hPa geopotential height anomaly correlation at day 5 for 20-80S (7 Jan–15 Feb 2004)
N. Hemisphere 500 hPa AC Z 20N - 80N Waves 1-20
1 Jan - 15 Feb '04
0.6
0.65
0.7
0.75
0.8
0.85
0.9
0.95
1
0 1 2 3 4 5 6 7
Forecast [days]
An
om
aly
Co
rrel
atio
n '
Control Noqscat
S. Hemisphere 500 hPa AC Z 20S - 80S Waves 1-20
1 Jan - 15 Feb '04
0.6
0.65
0.7
0.75
0.8
0.85
0.9
0.95
1
0 1 2 3 4 5 6 7
Forecast [day]
An
om
aly
Co
rrel
atio
n '
Control Noqscat
Second Goal Second Goal Run GFS with WindSat (windsob1°254)Run GFS with WindSat (windsob1°254)
Run GFS with WindSat (windsob0.5°254)Run GFS with WindSat (windsob0.5°254)
Run GFS with WindSat no QuikSCAT (windnoq254)Run GFS with WindSat no QuikSCAT (windnoq254)
Compare forecast impact for WindSat winds with Compare forecast impact for WindSat winds with the forecast impact for QuikSCAT windsthe forecast impact for QuikSCAT winds
Develop and improve QC techniques and verify the Develop and improve QC techniques and verify the current retrieval algorithmcurrent retrieval algorithm
Data used only at 6 hour synoptic time Data used only at 6 hour synoptic time with a plus/minus 3 hour windowwith a plus/minus 3 hour window..
If the absolute value of the observed wind If the absolute value of the observed wind component is more than 6 mscomponent is more than 6 ms-1 -1 from the from the corresponding background wind corresponding background wind component the observation is eliminated. component the observation is eliminated. This only removed the extreme outliers. This only removed the extreme outliers.
Control ----- T254L64 to 7 days (no AMSU, no AQUA AIRS, no QuikSCAT)NoQuikSCAT ----- T254L64 to 7 days (no AMSU, no AQUA AIRS, with QuikSCAT)WindSat 1.0 ----- T254L64 to 7 days (control + WindSat QC check and superobing to 1 deg)WindSat 0.5 ----- T254L64 to 7 days (control + WindSat QC check and superobing to 0.5 deg)
Wind DirectionWind Direction Std Std ≤≤ 20 20° for wspd>3m/s° for wspd>3m/s
LimitationLimitation Wind speed and direction Wind speed and direction
retrievals degrade in retrievals degrade in rainy areasrainy areas
http://www.ofcm.gov
WindSat and QuikSCAT Wind WindSat and QuikSCAT Wind FieldsFields
WindSat QuikSCAT
http://www.npoess.noaa.gov/polarmax
N. Hemisphere 500 hPa AC Z 20N - 80N Waves 1-20
1 Jan - 15 Feb '04
0.6
0.65
0.7
0.75
0.8
0.85
0.9
0.95
1
0 1 2 3 4 5 6 7
Forecast [days]
An
om
aly
Co
rrel
atio
n '
Noqscat Windsat
S. Hemisphere 500 hPa AC Z 20S - 80S Waves 1-20
1 Jan - 15 Feb '04
0.6
0.65
0.7
0.75
0.8
0.85
0.9
0.95
1
0 1 2 3 4 5 6 7
Forecast [day]
An
om
aly
Co
rrel
atio
n '
Noqscat Windsat
WindSat (1° superob) and QuikSCAT (0.5° superob)WindSat (1° superob) and QuikSCAT (0.5° superob) data counts and RMS error (20040125) data counts and RMS error (20040125)
The 1000 hPa wind speed anomaly correlation at day 5 for 20-80N (7 Jan–15 Feb 2004)
The 1000 hPa wind speed anomaly correlation at day 5 for 20-80S (7 Jan–15 Feb 2004)
Geopotential height anomaly correlation of Wind_sob_1.0deg and QuikSCAT_0.5deg at day 5 for 20-80N and 20-80S (7
Jan–15 Feb 2004)
0
100
200
300
400
500
600
700
800
900
1000
0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95
NH_WindSatSH_WindSatNH_QuikSCATSH_QuikSCAT
Future GoalsFuture Goals Examine the direct assimilation of the Examine the direct assimilation of the
WindSat radiances into the GFS by either WindSat radiances into the GFS by either updating the spectral statistical updating the spectral statistical interpolation (SSI) assimilation system or interpolation (SSI) assimilation system or switching to the GSI. switching to the GSI.
Conduct the same study with higher Conduct the same study with higher operational resolution (T382) and an operational resolution (T382) and an alternate season. alternate season.
Study the regional impact of WindSat Study the regional impact of WindSat assimilation and compare the difference assimilation and compare the difference between WindSat and QuikSCAT data. between WindSat and QuikSCAT data.
ConclusionsConclusions Preliminary results indicate that QuikSCAT Preliminary results indicate that QuikSCAT
data improved the forecast more than WindSat data improved the forecast more than WindSat data for a majority of the cases examined.data for a majority of the cases examined.
The combination of WindSat and QuikSCAT The combination of WindSat and QuikSCAT provided the largest positive forecast impact provided the largest positive forecast impact by day 7.by day 7.
At T254 WindSat demonstrated larger AC At T254 WindSat demonstrated larger AC gains with a 1° superob than a 0.5° superob. gains with a 1° superob than a 0.5° superob. (QuikSCAT uses a 0.5° superob by default).(QuikSCAT uses a 0.5° superob by default).
AcknowledgementsAcknowledgements
• Prof. Michael Morgan (UW-AOS) for Prof. Michael Morgan (UW-AOS) for giving insightful advice and for local giving insightful advice and for local computer resources.computer resources.
• Stephen Lord (NCEP) for computer Stephen Lord (NCEP) for computer resources and tape space.resources and tape space.
• Stacie Bender and Dennis Keyser for Stacie Bender and Dennis Keyser for collecting and processing our various collecting and processing our various data streams.data streams.