GODAE inter-comparisons around Australia

1
GODAE inter-comparisons around Australia

• Peter Oke1 and Gary Brassington2
• Centre for Australian Weather and Climate
  Research
• 1CSIRO Marine and Atmospheric Research
• 2Bureau of Meteorology
• May 2009

2
Regions

• The goals of this study are to
• to understand when, where and why different
  systems perform well, or poorly, so that all
  systems can be improved.
• determine whether some ocean processes are
  consistently well/poorly reproduced by all
  systems and to understand why.
• Inter-comparison period is Feb 2008-May 2008

3
GODAE Inter-comparisons Argo T/S

• OMAPS-fc Bluelink operational 3-4-d forecasts
• OMAPS-hc Bluelink operational hindcast (6-9-d
  behind real-time)
• BRAN Bluelink ReANalysis (1-yr behind
  real-time)
• HYCOM US Navy hindcast (5-d behind real-time)
• Mercator French hindcast (7-14-d behind
  real-time)
• UKMet UK hindcast (3-m behind real-time)

4
GODAE Inter-comparisons Argo T/S
OMAPS-fc Bluelink operational 3-4-d RT
forecasts OMAPS-hc Bluelink operational
hindcast (6-9-d behind RT) BRAN Bluelink
ReANalysis (1-yr behind RT) HYCOM US Navy
hindcast (5-d behind RT) Mercator French
hindcast (7-14-d behind RT) UKMet UK hindcast
(3-m behind RT)
5
GODAE Inter-comparisons Argo T/S
OMAPS-fc Bluelink operational 3-4-d RT
forecasts OMAPS-hc Bluelink operational
hindcast (6-9-d behind RT) BRAN Bluelink
ReANalysis (1-yr behind RT) HYCOM US Navy
hindcast (5-d behind RT) Mercator French
hindcast (7-14-d behind RT) UKMet UK hindcast
(3-m behind RT)
6
GODAE Inter-comparisons Argo T/S
OMAPS-fc Bluelink operational 3-4-d RT
forecasts OMAPS-hc Bluelink operational
hindcast (6-9-d behind RT) BRAN Bluelink
ReANalysis (1-yr behind RT) HYCOM US Navy
hindcast (5-d behind RT) Mercator French
hindcast (7-14-d behind RT) UKMet UK hindcast
(3-m behind RT)
7
GODAE Inter-comparisons XBT
OMAPS-fc Bluelink operational 3-4-d RT
forecasts OMAPS-hc Bluelink operational
hindcast (6-9-d behind RT) BRAN Bluelink
ReANalysis (1-yr behind RT) HYCOM US Navy
hindcast (5-d behind RT) Mercator French
hindcast (7-14-d behind RT) UKMet UK hindcast
(3-m behind RT)
8

• SLA from along-track altimeter

9
GODAE Inter-comparisons along-track SLA
OMAPS-fc Bluelink operational 3-4-d RT
forecasts OMAPS-hc Bluelink operational
hindcast (6-9-d behind RT) BRAN Bluelink
ReANalysis (1-yr behind RT) HYCOM US Navy
hindcast (5-d behind RT) Mercator French
hindcast (7-14-d behind RT) UKMet UK hindcast
(3-m behind RT)
10
GODAE Inter-comparisons along-track SLA
OMAPS-fc Bluelink operational 3-4-d RT
forecasts OMAPS-hc Bluelink operational
hindcast (6-9-d behind RT) BRAN Bluelink
ReANalysis (1-yr behind RT) HYCOM US Navy
hindcast (5-d behind RT) Mercator French
hindcast (7-14-d behind RT) UKMet UK hindcast
(3-m behind RT)
11
GODAE Inter-comparisons along-track SLA
OMAPS-fc Bluelink operational 3-4-d RT
forecasts OMAPS-hc Bluelink operational
hindcast (6-9-d behind RT) BRAN Bluelink
ReANalysis (1-yr behind RT) HYCOM US Navy
hindcast (5-d behind RT) Mercator French
hindcast (7-14-d behind RT) UKMet UK hindcast
(3-m behind RT)
12
GODAE Inter-comparisons along-track SLA
OMAPS-fc Bluelink operational 3-4-d RT
forecasts OMAPS-hc Bluelink operational
hindcast (6-9-d behind RT) BRAN Bluelink
ReANalysis (1-yr behind RT) HYCOM US Navy
hindcast (5-d behind RT) Mercator French
hindcast (7-14-d behind RT) UKMet UK hindcast
(3-m behind RT)
13
GODAE Inter-comparisons along-track SLA
OMAPS(fc) Bluelink operational 3-4-d RT
forecasts OMAPS(ban) Bluelink operational
hindcast (6-9-d behind RT) HYCOM US Navy
hindcast (5-d behind RT) Mercator French
hindcast (7-14-d behind RT) UKMet UK hindcast
(3-m behind RT)
14
GODAE Inter-comparisons along-track SLA
OMAPS(fc) Bluelink operational 3-4-d RT
forecasts OMAPS(ban) Bluelink operational
hindcast (6-9-d behind RT) HYCOM US Navy
hindcast (5-d behind RT) Mercator French
hindcast (7-14-d behind RT) UKMet UK hindcast
(3-m behind RT)
15
GODAE Inter-comparisons along-track SLA
OMAPS(fc) Bluelink operational 3-4-d RT
forecasts OMAPS(ban) Bluelink operational
hindcast (6-9-d behind RT) HYCOM US Navy
hindcast (5-d behind RT) Mercator French
hindcast (7-14-d behind RT) UKMet UK hindcast
(3-m behind RT)
16
GODAE Inter-comparisons along-track SLA
OMAPS(fc) Bluelink operational 3-4-d RT
forecasts OMAPS(ban) Bluelink operational
hindcast (6-9-d behind RT) HYCOM US Navy
hindcast (5-d behind RT) Mercator French
hindcast (7-14-d behind RT) UKMet UK hindcast
(3-m behind RT)
17

• SST fields from AMSRE

18
GODAE Inter-comparisons SST
?
Oke, P. R., G. B. Brassington, J. Cummings, M.
Drevillon, F. Hernandez and M. Martin 2009 GODAE
Inter-comparisons in the south western Pacific
Ocean, in preparation (see http//www.cmar.csiro.a
u/staff/oke/GODAE-SWPAC.htm)
19
GODAE Inter-comparisons SST
?
Perhaps there are more small-scale filaments that
are misplaced
Oke, P. R., G. B. Brassington, J. Cummings, M.
Drevillon, F. Hernandez and M. Martin 2009 GODAE
Inter-comparisons in the south western Pacific
Ocean, in preparation (see http//www.cmar.csiro.a
u/staff/oke/GODAE-SWPAC.htm)
20
GODAE Inter-comparisons AMSRE SST
OMAPS-fc Bluelink operational 3-4-d RT
forecasts OMAPS-hc Bluelink operational
hindcast (6-9-d behind RT) BRAN Bluelink
ReANalysis (1-yr behind RT) HYCOM US Navy
hindcast (5-d behind RT) Mercator French
hindcast (7-14-d behind RT) UKMet UK hindcast
(3-m behind RT)
21
GODAE Inter-comparisons AMSRE SST
OMAPS-fc Bluelink operational 3-4-d RT
forecasts OMAPS-hc Bluelink operational
hindcast (6-9-d behind RT) BRAN Bluelink
ReANalysis (1-yr behind RT) HYCOM US Navy
hindcast (5-d behind RT) Mercator French
hindcast (7-14-d behind RT) UKMet UK hindcast
(3-m behind RT)
22
GODAE Inter-comparisons AMSRE SST
OMAPS-fc Bluelink operational 3-4-d RT
forecasts OMAPS-hc Bluelink operational
hindcast (6-9-d behind RT) BRAN Bluelink
ReANalysis (1-yr behind RT) HYCOM US Navy
hindcast (5-d behind RT) Mercator French
hindcast (7-14-d behind RT) UKMet UK hindcast
(3-m behind RT)
23
GODAE Inter-comparisons AMSRE SST
OMAPS-fc Bluelink operational 3-4-d RT
forecasts OMAPS-hc Bluelink operational
hindcast (6-9-d behind RT) BRAN Bluelink
ReANalysis (1-yr behind RT) HYCOM US Navy
hindcast (5-d behind RT) Mercator French
hindcast (7-14-d behind RT) UKMet UK hindcast
(3-m behind RT)
24
GODAE Inter-comparisons AMSRE SST
OMAPS(fc) Bluelink operational 3-4-d RT
forecasts OMAPS(ban) Bluelink operational
hindcast (6-9-d behind RT) HYCOM US Navy
hindcast (5-d behind RT) Mercator French
hindcast (7-14-d behind RT) UKMet UK hindcast
(3-m behind RT)
25
GODAE Inter-comparisons AMSRE SST
OMAPS(fc) Bluelink operational 3-4-d RT
forecasts OMAPS(ban) Bluelink operational
hindcast (6-9-d behind RT) HYCOM US Navy
hindcast (5-d behind RT) Mercator French
hindcast (7-14-d behind RT) UKMet UK hindcast
(3-m behind RT)
26
GODAE Inter-comparisons AMSRE SST
OMAPS(fc) Bluelink operational 3-4-d RT
forecasts OMAPS(ban) Bluelink operational
hindcast (6-9-d behind RT) HYCOM US Navy
hindcast (5-d behind RT) Mercator French
hindcast (7-14-d behind RT) UKMet UK hindcast
(3-m behind RT)
27
GODAE Inter-comparisons AMSRE SST
OMAPS(fc) Bluelink operational 3-4-d RT
forecasts OMAPS(ban) Bluelink operational
hindcast (6-9-d behind RT) HYCOM US Navy
hindcast (5-d behind RT) Mercator French
hindcast (7-14-d behind RT) UKMet UK hindcast
(3-m behind RT)
28
SST standard deviation for Feb (top) Mar (bot)
29
GODAE Inter-comparisons U, V from surface
drifting buoys
OMAPS-fc Bluelink operational 3-4-d RT
forecasts OMAPS-hc Bluelink operational
hindcast (6-9-d behind RT) BRAN Bluelink
ReANalysis (1-yr behind RT) HYCOM US Navy
hindcast (5-d behind RT) Mercator French
hindcast (7-14-d behind RT) UKMet UK hindcast
(3-m behind RT)
Surface drifting buoys 49 buoys 1816 daily
average values lt0.1m/s removed
30

• Surface U,V fields from drifters

31

• Surface U,V fields from drifters

32
GODAE Inter-comparisons U, V from surface
drifting buoys
OMAPS-fc Bluelink operational 3-4-d RT
forecasts OMAPS-hc Bluelink operational
hindcast (6-9-d behind RT) BRAN Bluelink
ReANalysis (1-yr behind RT) HYCOM US Navy
hindcast (5-d behind RT) Mercator French
hindcast (7-14-d behind RT) UKMet UK hindcast
(3-m behind RT)
33
Summary conclusions

• Theres no stand-out good or bad performers
• different systems perform best in different
  regions and for different variables ?reason to
  consider super-ensembles for operational
  activities
• Inter-comparisons highlighted some known
  limitations
• HYCOM initial conditions were poor for these
  experiments (? sub-surface T/S)
• UK Met SLA assimilation not performed in regions
  of weak stratification (e.g., ACC)
• There is often evidence of a systematic
  degradation in performance from Reanalysis to
  NRT hind-cast to NRT forecast in the Bluelink
  system as we expect.
• links to all of the Australian inter-comparisons
• http//www.cmar.csiro.au/staff/oke/GODAE.htm
• SPA ? http//www.cmar.csiro.au/staff/oke/GODAE-SW
  PAC.htm
• IND ? http//www.bom.gov.au/bmrc/ocean/staff/gbb/
  GODAE_IND.htm

34
(No Transcript)