bidlot ocean wave analysis - ecmwf · reanalysis workshop slide 3 background : zecmwf has been...
TRANSCRIPT
Slide 1
Ocean Wave Analysisand
use of surface wind observationsover the oceans
Jean-Raymond BidlotPeter JanssenSaleh Abdalla
Hans Hersbach
European Centre for Medium Range Weather Forecasts, Reading, U.K.
ECMWF/GEO Workshop on Atmospheric Reanalysis. 19-22 June 2006
Reanalysis Workshop Slide 2
Outline :
Background.
Wave data available for reanalysis.
Wave data analysis at ECMWF.
ERA-40 wave data problem.
Wave model for the interim reanalysis.
Wave data for the interim reanalysis.
A few comments on in-situ surface wind.
Reanalysis Workshop Slide 3
Background :ECMWF has been running operational wave models since June 1992.
Two-way coupling between the wave model and the atmospheric model was introduced in June 1998 with benefits for both wave model (via high temporal wind field resolution) and atmospheric model (via wave dependent surface roughness).
More than just winds are now exchanged between the two models and more wave dependent processes could be added (i.e. sea-state dependent satellite retrievals, wave-current interactions).
Reanalysis wave data and wind data have many applications, quite often in coastal areas, hence the need for an increase in horizontal resolution.
Reanalysis Workshop Slide 4
Reanalysis wave data :
Wave height 100 year return value as derived from ERA-40as available in the KNMI wave and wind atlas:http://www.knmi.nl/waveatlas
Reanalysis Workshop Slide 5
Lon: 342-351Lat: 54N-45NDir: 0.- 45.Total of:67718. observations,4.%
Tm \ Hs 0.-1. 1.- 2. 2.- 3. 3.- 4. 4.- 5. 5.- 6. 6.-
7.7.-8.
8.-9.
9.-10.
10.- 11.
11.- 12.
12.- 13.
13.- 14.
14.- 33. sum
0.- 3. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.3.- 4. 2. 15. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 17.4.- 5. 24. 883. 122. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 1029.5.- 6. 36. 1016. 1625. 172. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 2849.6.- 7. 3. 647. 1013. 960. 242. 2. 0. 0. 0. 0. 0. 0. 0. 0. 0. 2868.7.- 8. 0. 259. 697. 432. 379. 166. 13. 0. 0. 0. 0. 0. 0. 0. 0. 1946.8.- 9. 0. 43. 355. 223. 92. 92. 74. 21. 1. 0. 0. 0. 0. 0. 0. 901.9.- 10. 0. 1. 69. 148. 38. 19. 12. 18. 12. 2. 0. 0. 0. 0. 0. 319.10.- 11. 0. 0. 1. 16. 27. 14. 3. 1. 1. 2. 1. 0. 0. 0. 0. 67.11.- 12. 0. 0. 0. 0. 0. 3. 0. 0. 0. 0. 0. 0. 0. 0. 0. 3.12.- 13. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.13.- 33. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0. 0.sum 65. 2864. 3882. 1952. 779. 296. 102. 41. 14. 3. 1. 0. 0. 0. 0. 10000.
i 4 90
Reanalysis wave data :
Tabular wave height and mean period bivariate histogramfor an area west of Ireland as derived from ERA-40:http://www.knmi.nl/waveatlas
Reanalysis Workshop Slide 6
Wave data for reanalysis : from satellites
Radar altimeter wave height observations:
- Seasat (1978), Geosat (1985-1986), ERS-1* (1991-1996), Topex(1992-2002),
- ERS-2* (1995-), GFO (2000-), Jason* (2002-), ENVISAT* (2002-)
Note: the wave model predicts the evolution of the wave spectrum. Wave height is connected to the total energy, which isproportional to the integral of the wave spectrum. Hence some assumptions are required to go from analysed wave height to analysed spectrum.
* Currently available in the ECMWF archives
Reanalysis Workshop Slide 7
Wave data for reanalysis : from satellites
Note: radar altimeter are nadir looking instrument, with a very narrow swath.Their global coverage is therefore limited.
ENVISAT in 24 hrs Jason in 24 hrs
Reanalysis Workshop Slide 8
Wave data for reanalysis : from satellites
SAR/ASAR ocean mode data:
- ERS-1* and 2* SAR
- ENVISAT* ASAR
Note: SAR does not unfortunately observe the full two-dimensional wave spectrum. In practice SAR wave spectra are retrieved using a model first guess. For these reasons, the impact of SAR assimilation is much reduced than originally hoped.
* Currently available in the ECMWF archives
Reanalysis Workshop Slide 9
Wave data for reanalysis verification: in-situ
40°S40°S
30°S 30°S
20°S20°S
10°S 10°S
0°0°
10°N 10°N
20°N20°N
30°N 30°N
40°N40°N
50°N 50°N
60°N60°N
160°W
160°W 140°W
140°W 120°W
120°W 100°W
100°W 80°W
80°W 60°W
60°W 40°W
40°W 20°W
20°W 0°
0° 20°E
20°E 40°E
40°E 60°E
60°E 80°E
80°E 100°E
100°E 120°E
120°E 140°E
140°E1991-08-01 to 1996-06-03
Number of collocations between ERS-1 altimeter and buoys
0
150
300
450
600
750
900
1050
1200
1350
1500
1650
1747.3
40°S40°S
30°S 30°S
20°S20°S
10°S 10°S
0°0°
10°N 10°N
20°N20°N
30°N 30°N
40°N40°N
50°N 50°N
60°N60°N
160°W
160°W 140°W
140°W 120°W
120°W 100°W
100°W 80°W
80°W 60°W
60°W 40°W
40°W 20°W
20°W 0°
0° 20°E
20°E 40°E
40°E 60°E
60°E 80°E
80°E 100°E
100°E 120°E
120°E 140°E
140°E2003-07-21 to 2006-05-13
Number of collocations between ENVISAT altimeter and buoys
0
150
300
450
600
750
900
1018.4
Buoy reporting wave observationshave a relatively limited geographical coverage.They are used for verification only.
Jun-
1987
Jun-
1988
Jun-
1989
Jun-
1990
Jun-
1991
Jun-
1992
Jun-
1993
Jun-
1994
Jun-
1995
Jun-
1996
Jun-
1997
Jun-
1998
Jun-
1999
Jun-
2000
Jun-
2001
Jun-
2002
Jun-
2003
Jun-
2004
Jun-
2005
Jun-
2006
months
0
10000
20000
30000
40000
50000
Num
ber o
f col
loca
tions
with
mod
el
Comparison with buoy data3 month collocation period
1987 2006The number of wave buoys available on the GTS has steadily increased over the years.
40°S40°S
30°S 30°S
20°S20°S
10°S 10°S
0°0°
10°N 10°N
20°N20°N
30°N 30°N
40°N40°N
50°N 50°N
60°N60°N
160°W
160°W 140°W
140°W 120°W
120°W 100°W
100°W 80°W
80°W 60°W
60°W 40°W
40°W 20°W
20°W 0°
0° 20°E
20°E 40°E
40°E 60°E
60°E 80°E
80°E 100°E
100°E 120°E
120°E 140°E
140°E1996-06-03 to 2003-06-22
Number of collocations between ERS-2 altimeter and buoys
01503004506007509001050120013501500165018001950210022502400
Reanalysis Workshop Slide 10
Wave data analysis:
The wave model data analysis is based on an optimal interpolation scheme for wave height as originally developed by Lionello et al. 1992.
The scheme was extended to assimilate wave systems as derived from SAR wave spectra.
It is not yet directly coupled to the atmospheric scheme.
Reanalysis Workshop Slide 11
ERA-40 :
1.5ºx1.5º wave model coupled to T159 atmospheric model.
Only deep water physics.
Wave spectra discretised with 12 directions and 25 frequencies.
Near-real time operationally obtained ERS-1 (1991-1996) and ERS-2 (1996-2002) altimeter data were used.
An extensive comparison with buoy data and Topex altimeter data was carried by a KNMI team http://www.knmi.nl/waveatlas
Reanalysis Workshop Slide 12
Too low wind in ERA-40 :
Jun-
1987
Jun-
1988
Jun-
1989
Jun-
1990
Jun-
1991
Jun-
1992
Jun-
1993
Jun-
1994
Jun-
1995
Jun-
1996
Jun-
1997
Jun-
1998
Jun-
1999
Jun-
2000
Jun-
2001
Jun-
2002
Jun-
2003
Jun-
2004
Jun-
2005
Jun-
2006
months
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
Win
d Sp
eed
Bias
(mod
el-b
uoy)
(m/s
)
ERA-40 (e4 0001)ops (od 0001)
Comparison with buoy data3 month running average
1987 2006
Wind speed bias
Jun-
1987
Jun-
1988
Jun-
1989
Jun-
1990
Jun-
1991
Jun-
1992
Jun-
1993
Jun-
1994
Jun-
1995
Jun-
1996
Jun-
1997
Jun-
1998
Jun-
1999
Jun-
2000
Jun-
2001
Jun-
2002
Jun-
2003
Jun-
2004
Jun-
2005
Jun-
2006
months
-0.4
-0.2
0
0.2
0.4
Wav
e H
eigh
t Bia
s (m
odel
-buo
y) (m
)
CY30R1 stand alone hindcast (rd era5)
Comparison with buoy data3 month running average
1987 2006
Hs bias
Results from a wave model hindcast run(uncoupled, no data assimilation),with the latest operational model version (CY30R1),forced by ERA-40 winds until August 2002, thenoperational analysis winds.
Reanalysis Workshop Slide 13
Problem with wave data in ERA-40 :
Jun-
1987
Jun-
1988
Jun-
1989
Jun-
1990
Jun-
1991
Jun-
1992
Jun-
1993
Jun-
1994
Jun-
1995
Jun-
1996
Jun-
1997
Jun-
1998
Jun-
1999
Jun-
2000
Jun-
2001
Jun-
2002
Jun-
2003
Jun-
2004
Jun-
2005
Jun-
2006
months
-0.4
-0.2
0
0.2
0.4
Wav
e H
eigh
t Bia
s (m
odel
-buo
y) (m
)
ERA-40 (e4 0001)ops (od 0001)
Comparison with buoy data3 month running average
1987 2006Dec. 1991
May 1993
Between Dec. 1991 and May 1993low quality ERS-1 altimeter datawere wrongly assimilated.
Jun-
1987
Jun-
1988
Jun-
1989
Jun-
1990
Jun-
1991
Jun-
1992
Jun-
1993
Jun-
1994
Jun-
1995
Jun-
1996
Jun-
1997
Jun-
1998
Jun-
1999
Jun-
2000
Jun-
2001
Jun-
2002
Jun-
2003
Jun-
2004
Jun-
2005
Jun-
2006
months
-0.5
0
0.5
1
1.5
2
Wav
e M
ean
Perio
d B
ias
(mod
el-b
uoy)
(s) ERA-40 (e4 0001)
ops (od 0001)
Comparison with buoy data3 month running average
1987 2006Dec. 1991
May 1993
Between Dec. 1991 and May 1993low quality ERS-1 altimeter datawere wrongly assimilated.
Hs bias Tz bias
Note: this problem was originally spotted by KNMI. This emphasises theneed for ‘in-house’ monitoring tools.
Reanalysis Workshop Slide 14
Better surface winds for the interim reanalysis :
T159 T255 better (ideally more)
Benefit from using 4dvar.
Improved atmospheric model physics.
The latest physics changes (CY31R1) should be even better.
Jun-
1987
Jun-
1988
Jun-
1989
Jun-
1990
Jun-
1991
Jun-
1992
Jun-
1993
Jun-
1994
Jun-
1995
Jun-
1996
Jun-
1997
Jun-
1998
Jun-
1999
Jun-
2000
Jun-
2001
Jun-
2002
Jun-
2003
Jun-
2004
Jun-
2005
Jun-
2006
months
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
Win
d Sp
eed
Bias
(mod
el-b
uoy)
(m/s
)
ERA-40 (e4 0001)ops (od 0001)preliminary interim reanalysis (e4 0060)interim reanalysis prototype (ei 1001)interim analysis prototype (ei 1103)
Comparison with buoy data3 month running average
1987 2006
Wind speed bias
Reanalysis Workshop Slide 15
Improved wave model for the interim analysis :
1.5º 1.0º
24 directions and 30 freq.
Shallow water physics.
Improved wave model.-Unresolved bathymetry scheme.
-New dissipation source term.
-…
Jun-
1987
Jun-
1988
Jun-
1989
Jun-
1990
Jun-
1991
Jun-
1992
Jun-
1993
Jun-
1994
Jun-
1995
Jun-
1996
Jun-
1997
Jun-
1998
Jun-
1999
Jun-
2000
Jun-
2001
Jun-
2002
Jun-
2003
Jun-
2004
Jun-
2005
Jun-
2006
months
10
12
14
16
18
20
22
24
26
Wav
e M
ean
Perio
d Sc
atte
r Ind
ex (%
)
ERA-40 (e4 0001)ops (od 0001)CY30R1 hindcast (rd era5)preliminary reanalysis (e4 0060)interim prototype (ei 1001)interim reanalysis prototype (ei 1103)
Comparison with buoy data3 month running average
1987 2006
Tz bias
Reanalysis Workshop Slide 16
Wave data for the interim reanalysis :
Because of the inhomogeneity of the ERS-1 and also some other known problems with the NRT ERS data, we have obtained OPR ERS-1 and 2 data from ESA.
Some works was necessary to convert the data from CD’s into data in a format more compatible with ECMWF system (i.eBUFR)!
NRT ENVISAT altimeter data are already of good quality. Note that some orbits are missing, especially in early days of the mission due to NRT transmission problems. Access to off-lindata could prove useful.
NRT Jason altimeter data are also available.
SAR data from ERS-2 and ASAR data from ENVISAT will also be used.
Reanalysis Workshop Slide 17
Wave data for the interim reanalysis :
Because the OPR ERS-1 and 2 have never been used by our data assimilation system, we have run passively the altimeter data through a long wave model hindcast with the latest operational wave model version (CY30R1) using ERA-40 and operational winds.
A comprehensive buoy data set was generated by merging NDBC (US) and MEDS (Canada) data sets with GTS data from the ECMWF archives.
This comprehensive buoy data set was collocated with the altimeter data and the model counterparts.
Triple collocation techniques can be used to infer relative biases and error estimates for each data set.
Reanalysis Workshop Slide 18
Wave data for the interim reanalysis :
ERS-1 OPR wave heights are stillbiased low when compared to buoys
ERS-2 OPR wave heights are slightlybiased low when compared to buoys
ENVISAT wave heights are slightlybiased high when compared to buoys
0 1 2 3 4 5 6 7 8 9 10 11 12 13GTS wave heights
0123456789
10111213
Cor
rect
ed a
lt. w
ave
heig
hts ENTRIES:
1 - 33 - 66 - 12
12 - 26
26 - 59
59 - 133
133 - 300
ERS-1
1991-08-01 to
1996-06-03
Buoy wave heights
0 1 2 3 4 5 6 7 8 9 10 11GTS wave heights
0
1
2
3
4
5
6
7
8
9
10
11
Cor
rect
ed a
lt. w
ave
heig
hts ENTRIES:
1 - 33 - 77 - 15
15 - 37
37 - 90
90 - 220
220 - 540
ERS-2
1996-06-03to
2003-06-22
Buoy wave heights
0 1 2 3 4 5 6 7 8 9 10 11GTS wave heights
0
1
2
3
4
5
6
7
8
9
10
11
Cor
rect
ed a
lt. w
ave
heig
hts ENTRIES:
1 - 33 - 55 - 10
10 - 22
22 - 46
46 - 98
98 - 210
ENVISAT
2003-07-21to
2006-05-13
Buoy wave heights
Reanalysis Workshop Slide 19
Wave data for the interim reanalysis :
Using a non-parametric bias estimation technique, we have determined the relative bias with respect to the buoys (work still under way).
Initial results are quite satisfactory. More test will follow.0 1 2 3 4 5 6 7
wave height (m)
-0.8
-0.6
-0.4
-0.2
0
0.2
BIA
S fo
r alti
met
er w
ave
heig
ht a
s pr
oces
sed
by W
AM
ERS-1ERS-2ENVISAT
BIAS as derived from non-parametric estimate
Reanalysis Workshop Slide 20
Wave data for interim reanalysis :
Jun-
1987
Jun-
1988
Jun-
1989
Jun-
1990
Jun-
1991
Jun-
1992
Jun-
1993
Jun-
1994
Jun-
1995
Jun-
1996
Jun-
1997
Jun-
1998
Jun-
1999
Jun-
2000
Jun-
2001
Jun-
2002
Jun-
2003
Jun-
2004
Jun-
2005
Jun-
2006
months
14
16
18
20
22
24
26
28
30
Wav
e H
eigh
t Sca
tter I
ndex
(%)
ERA-40 (e4 0001)CY30R1 stand alone hindcast (rd era5)interim reanalysis prototype (ei 1103)
Comparison with buoy data3 month running average
1987 2006
with bias corrected ERS-1 data
Reanalysis Workshop Slide 21
In-situ surface wind data over the ocean :
The quality of any wave reanalysis is largely determined by the quality of the wind fields used to force the model.
More emphasis should be put on making a better use of wind observations over the oceans.
Reanalysis Workshop Slide 22
In-situ surface wind data over the ocean :
-0.50
0
0
0 0
0
0
0
0
0
00
0
0
0 0
0
0
0
41001
41004
410064100941010
41018
42002 4200342019
42020
42035 420364203942040
44004
4400544007
44009
44011
44014
44025
44140
46001
46002
4600346004
46005
46006
4601346014
46023
46030
4603646041
46042
46050
46059
46132
4614546184
46204
4620646207
46208
51001
5100251003
51004
51026
10°N 10°N
15°N15°N
20°N 20°N
25°N25°N
30°N 30°N
35°N35°N
40°N 40°N
45°N45°N
50°N 50°N
55°N55°N
60°N 60°N160°W
160°W 155°W
155°W 150°W
150°W 145°W
145°W 140°W
140°W 135°W
135°W 130°W
130°W 125°W
125°W 120°W
120°W 115°W
115°W 110°W
110°W 105°W
105°W 100°W
100°W 95°W
95°W 90°W
90°W 85°W
85°W 80°W
80°W 75°W
75°W 70°W
70°W 65°W
65°W 60°W
60°W 55°W
55°W 50°W
50°W
from 19960201 to 19960229Mean analysed minus forecast 10m wind speed for expver:18
-1
-0.5
-0.5
-0.5
0
0
0
0
00
0
0
0
0
0
0
0
0
0
0
0
0.5
0.5
0.5
0.5
1
62026
62081
62103
62105
62106
62107
62108
62109
62112
62162
62163
62165
6230362304
62305
63103
63111
63113
64045
LF3F
LF3J
LF3N
LF4C
LF5T
LF5U
LDWR
40°N 40°N
45°N45°N
50°N 50°N
55°N55°N
60°N 60°N
65°N65°N
70°N 70°N
25°W
25°W 20°W
20°W 15°W
15°W 10°W
10°W 5°W
5°W 0°
0° 5°E
5°E 10°E
10°E
from 19960201 to 19960229Mean analysed minus forecast 10m wind speed for expver:18
During the ERA-40 production, thefollowing mean analysis incrementsin surface winds were found.For a large part, they collocated withlocations where meteorological buoyswere deployed !
NB: maps of mean analysis increments are powerful tools
Reanalysis Workshop Slide 23
In-situ surface wind data over the ocean :
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
00
0
0
0
0
0
0
0
0.1
0.1
41001
41004
410064100941010
42002 4200342019
42020
42035 420364203942040
44004
4400544007
44009
44011
44014
44025
46001
46002
4600346004
46005
46006
4601346014
46023
46030
4603646041
46042
46050
46059
46132
4614546184
46204
46206
4620746208
51001
5100251003
51004
5102620°N 20°N
25°N25°N
30°N 30°N
35°N35°N
40°N 40°N
45°N45°N
50°N 50°N
55°N55°N
160°W
160°W 155°W
155°W 150°W
150°W 145°W
145°W 140°W
140°W 135°W
135°W 130°W
130°W 125°W
125°W 120°W
120°W 115°W
115°W 110°W
110°W 105°W
105°W 100°W
100°W 95°W
95°W 90°W
90°W 85°W
85°W 80°W
80°W 75°W
75°W 70°W
70°W 65°W
65°W 60°W
60°W 55°W
55°Wfrom 19960201 to 19960228
Mean analysis difference in 10m wind speed: 314 - 315
0
0
0
0
0
0
0
0
0
0
0 0
0
00
0
0
0
0
0
0
0
0
0.1
0.1
0.1
0.2
62026
62081
62103
62105
62106
62107
62108
62109
62112
62162
62163
62165
62303 6230462305
63103
63111
63113
64045
LF3F
LF3J
LF3N
LF4C
LF5T
LF5U
LDWR
40°N 40°N
45°N45°N
50°N 50°N
55°N55°N
60°N 60°N
65°N65°N
70°N 70°N
35°W
35°W 30°W
30°W 25°W
25°W 20°W
20°W 15°W
15°W 10°W
10°W 5°W
5°W 0°
0° 5°E
5°E 10°E
10°E 15°E
15°Efrom 19960201 to 19960228
Mean analysis difference in 10m wind speed: 314 - 315
Further analysis, indicated that thescheme used to provide the dataanalysis with the actual height of theanemometer did not work !
Most buoys observes winds at height ~4-5 m. By default the analysis assumes a height of 10m(25m for ships).Once corrected, the expected impact was visible (ERA-40 was modified ~1997).Is it the end of it ?
Reanalysis Workshop Slide 24
In-situ surface wind data over the ocean :
NHOV 00
BL
NTSG 00
BL
CHLV2
BL
CLKN7
BL
FBIS1BL
IOSN3
BL
DSLN7
BL
TPLM2
BL
FPSN7
BL
IOSN3
BL
CHLV2
BL
CLKN7
BL
FBIS1BL
ALSN6
BL
DBLN6BL
KIBK
TPLM2
BL
IOSN3
BL
CHLV2
BL
CLKN7
BL
FBIS1BL
ALSN6
BL
DBLN6BL
DSLN7
BL
TPLM2
BL
FPSN7
BL
IOSN3
BL
44013
BL
44011
44014
BL
44004
41001
44025
BL
41004
BL
44009
BL
41002
44005
44004
41001
44025
BL
41004
BL
44009
BL
41002
44005
44004
41001
44011
44009
BL
44014
BL
CG268
BL
DQFV
BL
ZCAM9
BL
44025
BL
41004
BL
44009
BL
41002
44005
44014
BL
44025
BL
41004
BL
44011
44014
BL
44004
41001
NJCS 00
BL
DSLN7
BL
TPLM2
BL
FPSN7
BL
ALSN6
BL
DBLN6BL
CHLV2
BL
CLKN7
BL
FBIS1BL
ALSN6
BL
DBLN6BL
DSLN7
BL
TPLM2
BL
FPSN7
BL
IOSN3
BL
DSLN7
BL
FPSN7
BL
FBIS1BL
ALSN6
BL
DBLN6BL
DSLN7
BL
TPLM2
BL
FPSN7
BL
IOSN3
BL
CHLV2
BL
CLKN7
BL
FBIS1BL
ALSN6
BL
DBLN6BL
44025
BL
41004
BL
41002
44005
44009
BL
44013
BL
44011
44014
BL
44013
BL
44011
44014
BL
44004
41001
44013
BL
44011
44004
41001
44013
BL
41002
44005
44009
BL
41001
44014
BL
41004
BL
4401144028
BL
41001
44014
BL
41004
BL
4401144028
BL
4400444009
BL
41002
4401144028
BL
41001
44014
BL
41004
BL
44005
BL
44013
BL
44025
BL
41001
44014
BL
41004
BL
44005
BL
44013
BL
44025
BL
41001
44014
BL
41004
BL
44005
BL
44013
BL
44025
BL
4400444009
BL
41002
44005
BL
44013
BL
44025
BL
4400444009
BL
41002
44005
BL
44013
BL
44025
BL
41001
44014
BL
41004
BL
44005
BL
44013
BL
44025
BL
4400444009
BL
41002
4401144028
BL
4400444009
BL
41002
4401144028
BL
4400444009
BL
41002
4401144028
BL
40°N 40°N
for 19960301 0 UTCERA40 SHIP and DRIBU observations
SYNOP: 114 DRIBU: 0 AIREP: 0 SATOB: 0 TEMP: 0 PILOT: 0 SATEM: 0 CORRECT 29
POSS ERR 0 PROB ERR 0 ERROR: 85 SYNOP: 0 DRIBU: 66 AIREP: 0 SATOB: 0 TEMP: 0
It was not realised then that some buoy data are available both as SHIP and DRIBU, because of the blend of ECMWF (SHIP) and NCEP (DRIBU) data sets.
The anemometer height scheme only worked on SHIP data !
It now works for DRIBU dataas well (since end of August 2005 in operations) and in the interim reruns (finally)!
Reanalysis Workshop Slide 25
In-situ surface wind data over the ocean :
9N140W9N140W9N140W
8S180W8S180W8S180W8S165E8S165E8S165E 8S155W8S155W8S155W 8S125W8S125W8S125W 8S110W8S110W8S110W
8N170W8N170W8N170W8N165E
FG
8N165E
FG
8N165E 8N125W8N125W8N125W 8N110W8N110W8N110W
5S180W5S180W5S180W5S156E
BL
5S156E
BL
5S156E
BL
5S155W
BL
5S155W
BL
5S155W
BL
5S140W5S140W5S140W 5S110W5S110W
5S110W
5N180W5N180W5N180W 5N170W5N170W5N170W5N165E5N165E5N165E5N155W5N155W5N155W
5N140W5N140W5N140W 5N125W5N125W5N125W 5N110W5N110W5N110W2N180W2N180W2N180W 2N170W2N170W2N170W2N165E2N165E2N165E 2N155W2N155W2N155W2N147E2N147E2N147E 2N125W2N125W2N125W
2S95W2S95W2S95W2S180W2S180W2S180W 2S170W
BL
2S170W
BL
2S170W
BL
2S155W2S155W2S155W 2S140W2S140W2S140W 2S125W2S125W2S125W 2S110W2S110W2S110W0N180W0N180W0N180W 0N170W0N170W0N170W0N165E0N165E0N165E0N156E0N156E0N156E 0N155W0N155W0N155W0N143E
BL
0N143E
BL
0N143E
BL
0N140W0N140W0N140W 0N125W0N125W0N125W 0N110W0N110W0N110W2256
4DDEEK
91338
DZOD
BL
JADC
BL91222
BL
91352
DVRF
BL
CPCL
BL
121691338
91251
BL
SANF1
BL
91222
BL
1136
91377
BL
9132891251
BL91222
BL
FNOM
BLYJZH9
C6IC
BL
1466
KTDQ1225
NCJE
29891395DUMF
NCDWDZOD
P3XM4ELNG9JRKN
JADCKRJFKGXA
ELIG8BL6ZXGBL
DVXENCVVWJLX
9134391355
BL
1217
CBGR
BL
91377
BL
91338
VVGC
9134391355
BL
2976
MYRF
91377
BL
91338
91251
BL91222
BL
5100251004
SKFQJJRQ
BL
DVRF
BL
51002510045100251004
VRGW
BL
PGDX
VNGK
BL
C6KV2BL
5100251004
C6HL8DNKL
ELKD6
4XGVVPIB
C6KF2
UHVW
JJRQ
BL
8JPX
PJNL
C6HK9
BL
C6IC
C6MD6
WNJG
BL
PFPT
BL
GCCCPGLA
ELJS5
BLOYMO2
CPCL
BL
HRRF
BL
5100251004
51001BL
52009
BL
9N140W9N140W9N140W
8S180W8S180W8S180W8S165E8S165E8S165E 8S155W8S155W8S155W 8S125W8S125W8S125W 8S110W8S110W8S110W
8N170W8N170W8N170W8N165E
FG
8N165E8N165E 8N125W8N125W8N125W 8N110W8N110W8N110W
5S180W5S180W5S180W5S156E
BL
5S156E
BL
5S156E
BL
5S155W
BL
5S155W
BL
5S155W
BL
5S140W5S140W5S140W 5S110W5S110W5S110W
5N180W5N180W5N180W 5N170W5N170W5N170W5N165E5N165E5N165E5N155W5N155W5N155W
5N140W5N140W5N140W 5N125W5N125W5N125W 5N110W5N110W5N110W2N180W
FG
2N180W
FG
2N180W 2N170W2N170W2N170W2N165E2N165E2N165E 2N155W2N155W2N155W2N147E2N147E2N147E 2N125W2N125W2N125W
2S95W2S95W2S95W2S180W2S180W2S180W 2S170W
BL
2S170W
BL
2S170W
BL
2S155W2S155W2S155W 2S140W2S140W2S140W 2S125W2S125W2S125W 2S110W2S110W2S110W0N180W0N180W0N180W 0N170W0N170W0N170W0N165E0N165E0N165E0N156E0N156E0N156E 0N155W0N155W0N155W0N143E
BL
0N143E
BL
0N143E
BL
0N140W0N140W0N140W 0N125W0N125W0N125W 0N110W0N110W0N110W
139891377
BL
9132891251
BL
SANF1
BL
JRKN
9134391355
BL2837
CBGR
BL
91377
BL
91328
J8PD
BL
913439135291355
BL
2534
2931
2984
91338
1400
SANF1
BL
913439135291355
BL
1124
1227
FAREBL1218
C6KB6
2811ELEU
259191377
BL
9133891328
WFQE
BL
LACP4LACP4
91251
BL
NUOT
BL
ELLE2JCOY
DVCKBL3EPU4
NPFDBLNAON
91222
BL
91352
1124
DVRF
BL
CPCL
BL
9132891251
BL
SANF1
BL
91222
BL
91352
1223
2984
2837
HRRF
BL
91328
SANF1
BL
913439135291355
BL
32302
51026
BL
51003
51001BL
JPTFJCOY
JBTQDZOD
BL
CPCL
BL
51026BL
51003
51001BL 51026BL51001BL
VRSPBL
YJYY2
BL
VRSB
BL
VROC32302
V2EFC6KB4
51026BL51001BL
KCGH
VVGCJKPS
JPTF
JBTQ VRIY
WXWL
6YSA
ELMH5
ZHET5
BLZCKUBL
SHIP
BL
DVRF
BL
MYRF
51026BL51001BL
5100251004
32302
51026
BL
5100451003
51001BL
3232232323
3231851308
51009
51302
51010
52313
52306
BL
51011
51307
51311
51006
51023
51303
513055231152321
51002
3232232323
3231851308
51009
51302
51010
5231352007
52010
BL52306
BL 3231543001
51015
51016
5100751020
51021
51309
5231052311
52006
FG
52003
32302
51026BL
51002
32319
32317510175101451019
BL52306
BL 3231543001
51015
51016
5100751020
5231752301
32302
5100451003
51001BL
32322
32319
32317510175101451019
BL
51306
BL
52313
52306
BL
51011
51307
51311
51006
51023
51303
51305
52309
52006
FG
52003
52317
5100451003
51001BL
32319
3231751017
51302
5231352007
52010
BL 51011
51307
51311
5130352309
5232152001
32302
51026BL
5100451003
51001BL
3232351010
5231352007
52010
BL52306
BL 32315
51311
51303
51305
52309
5232152001
52009
BL
52007
52301
5200352008FG
52313
5131052316
51022
52007
52525
52310
FG
5130551023
5200352006
FG
51302
510095101051306
BL
5101451022 51009
5101451019
BL
3231932318
51308
51006
51007
5130551023
51006
5102351016
43001
51002
52525
32319
32317510175101451019
BL
5102251306
BL
52312
52316
3231543001
51015
51016
5100751020
51021
5130952309
52317
32302
51026
BL
5100451003
51001BL
32319
32317510175101451019
BL
5102251306
BL
52312
52316
51011
51307
51311
51006
51023
51303
51305
52309
5232152001
5231752301
52525
5100451003
51001BL
3232232323
3231851308
51009
51302
52010
BL 51011
51307
51311
51006
5102152321
51026BL
51002
3232332318
51308
510095102251010
523165200752010
BL
3231543001
51015
51016
5100751020
51021
51309
52310523215200152301
32302
51026BL
51002
3232232323
3231851308
52312
52316
52306
BL 3231543001
51015
51016
51020
5231052311
5200652003
5231752301
51002
32322
51302
51306
BL
52312
52316
51011
5102051309
5231052311
5200652003
5231752301
52012
BL
52010
BL 52310
FG
52011 52001
52006
FG
523125101051306
BL51303
52311
5230951309
52001
5231252311
51019
BL
51310
510185101851022 51017 32317
51021
5130951007
51021
5130751015
10°S10°S
0° 0°
10°N10°N
20°N 20°N
140°E
140°E 160°E
160°E 180°
180° 160°W
160°W 140°W
140°W 120°W
120°W 100°W
100°W
for 19940301 18 UTCERA40 SHIP and DRIBU observations
SYNOP: 468 DRIBU: 0 AIREP: 0 SATOB: 0 TEMP: 0 PILOT: 0 SATEM: 0 CORRECT 357POSS ERR 5 PROB ERR 0 ERROR: 106 SYNOP: 0 DRIBU: 311 AIREP: 0 SATOB: 0 TEMP: 0
PILOT: 0 SATEM: 0 CORRECT 266 POSS ERR 5 PROB ERR 2 ERROR: 38
The anemometer height scheme relies on a list connecting the station identifier (buoys, ship and platforms) with the height of the anemometer. It has been built by gathering information fromdifferent sources. It will requires more effort for older data sets.
e.g. earlier TAOdata both appear as SHIP and DRBU but with different station identifier:e.g.51014 <=> 5S140W
Reanalysis Workshop Slide 26
In-situ surface wind data over the ocean :
Jun-
1992
Jun-
1993
Jun-
1994
Jun-
1995
Jun-
1996
Jun-
1997
Jun-
1998
Jun-
1999
Jun-
2000
Jun-
2001
Jun-
2002
Jun-
2003
Jun-
2004
Jun-
2005
Jun-
2006
months
0
5000
10000
15000
20000
Num
ber o
f colloca
tions
with
mod
el
ERA-40 (e4 0001)ops (od 0001)
Comparison with TAO buoys as obtained by GTS3 month running average
Number of data
Limited past GTS dissimination of certain data should be complemented by off-line data as they are very useful data.e.g. TAO data for more recentyears before the upgrade in earlier 2005.
Jun-
1992
Jun-
1993
Jun-
1994
Jun-
1995
Jun-
1996
Jun-
1997
Jun-
1998
Jun-
1999
Jun-
2000
Jun-
2001
Jun-
2002
Jun-
2003
Jun-
2004
Jun-
2005
Jun-
2006
months
-1.5
-1.25
-1
-0.75
-0.5
-0.25
0
Win
d Sp
eed
Bias
(mod
el-b
uoy)
(m/s
)ERA-40 (e4 0001)ops (od 0001)
Comparison with TAO buoys as obtained by GTS3 month running average
bias
1992 2006
Jun-
1992
Jun-
1993
Jun-
1994
Jun-
1995
Jun-
1996
Jun-
1997
Jun-
1998
Jun-
1999
Jun-
2000
Jun-
2001
Jun-
2002
Jun-
2003
Jun-
2004
Jun-
2005
Jun-
2006
months
10
15
20
25
30
Win
d Sp
eed
Scat
ter I
ndex
(%)
ERA-40 (e4 0001)ops (od 0001)
Comparison with TAO buoys as obtained by GTS3 month running average
S.I.
1992 2006
Reanalysis Workshop Slide 27
In-situ surface wind data over the ocean :
There is a need to sort out in-situ observations over the oceans to avoid unnecessary duplications.
It is unfortunate that potentially very valuable data from moored buoys are mixed with other type of data of the same nature (e.g.with ship and fixed platforms (SHIP) or with drifting buoys (DRIBU)). This is reflected by the relatively large error assigned to these data. A separate data type should be created (moored buoys) and bit of data mining should take place to bring the data to ECMWF.
Reanalysis Workshop Slide 28
Conclusions :
Wind and waves should be better analysed in the interim analysis following improvements in both atmospheric and wave models.
A more consistent altimeter wave height data base has been created with the acquisition of OPR ERS data and the homogenization of the data following bias estimation from a triple collocated data set.
A bit more work (coordination) on a surface wind data base could be beneficial to any future reanalysis.