The Global Ocean Data Assimilation System (GODAS) at NCEP

1
The Global Ocean Data Assimilation System (GODAS)
at NCEP
David W. Behringer, NCEP/EMC Yan Xue,
NCEP/CPC Boyin Huang , NCEP/CPC NOAA Climate
Observation Division 5th Annual System
Review Silver Spring, Maryland 5-7 June 2007
2
Content

• The impact of satellite altimetry data on GODAS
• The impact of deep Argo temperature data on GODAS
  
• The impact of Argo salinity data on GODAS
• Accessing GODAS data and products
• Future plans

3
Seasonal to Interannual Prediction at NCEP
4
Operational GODAS performance
The standard GODAS reproduces the interannual
variability well, notably the El Nino / La Nina
signal.
5
GODAS (MOM V.3) Operational since September
2003 Grid Quasi-global, 1x1 degree horizontal
resolution, enhanced to 1/3 degree in the
tropics, 40 vertical levels. Physics KPP
boundary layer mixing scheme, GM isopycnal
tracer mixing, Smagorinsky velocity mixing, free
surface. Forcing Wind stress, heat flux, E-P
from Reanalysis 2 (R2), SST relaxed to weekly
Reynolds analysis, SSS relaxed to Levitus
monthly climatology. Assimilation method 3D
VAR, limited to upper 750m, univariate in
temperature and salinity, error covariance varies
geographically and temporally. Assimilation
data Temperature profile data from XBTs,
profiling floats (Argo), moorings (TAO),
synthetic salinity constructed from temperature
and local Levitus T-S climatology. Satellite
altimetry from TOPEX / Jason-1 (beginning in
March 2007).
6
Altimetry Assimilation 1993 - 2005
7
(No Transcript)
8
Deep Assimilation 1980 - 2005
9
The changing number and characteristics of
observations
10
Standard vs. Deep assimilation
The standard assimilation extends down to 750 m.
The deep assimilation extends down to 2200 m.
IND PAC ATL
IND PAC ATL
11
Standard vs. Deep assimilation
Independent WOCE CTD section completed in 1988
1989
and repeated in 2003 2005 by PMEL.
Shallow assimilation has a strong cold bias
of 1-3oC below 750 m.
Standard
Deep assimilation eliminates the cold bias.
Deep
12
Assimilating Argo Salinity 2005
13
Equatorial salinity section in the Pacific
(vertical bars show positions of time-series
below).
Assimilating Argo Salinity
GODAS
Salinity variability due to correlation with
temperature.
GODAS-A/S
Salinity variability introduced by observations.
14
Assimilating Argo observations alters GODAS
salinity by as much as 0.5 psu over broad regions.
Assimilating Argo Salinity
Changes in both S and U are largely confined in
and above the thermocline.
15
Assimilating Argo Salinity
Comparison with independent ADCP currents.
ADCP GODAS GODAS-A/S
16
Summary - The impact of observations in GODAS

• Assimilating TOPEX/Jason-1 into an ocean analysis
  improves the SSH and the upper ocean heat content
  outside of the tropical Pacific. Those
  improvements, however, havent led to improved SI
  forecasts. In the tropical Pacific the GODAS is
  already well constrained by data from the TAO /
  TRITON array and it is difficult to improve on
  that. Further improvements in data assimilation
  and coupled modeling may be needed to take
  advantage of satellite altimetry.

• Deep assimilation (2200m) of Argo data eliminates
  the deep temperature drift in the standard GODAS
  and produces an analysis in good agreement with
  independent CTD sections (WOCE A16). In the years
  prior to the availability of Argo profiles, deep
  climatology serves to prevent the temperature
  drift.

• Accurate equatorial currents require accurate
  salinity. Assimilating Argo salinity produces
  the detailed structure in the equatorial mass
  field needed to accurately portray the equatorial
  currents. This has not been possible with
  synthetic salinity.

17
Accessing GODAS products and data
Products http//www.cpc.ncep.noaa.gov/products/GOD
AS Data http//cfs.ncep.noaa.gov/cfs/godas/monthl
y http//cfs.ncep.noaa.gov/cfs/godas/pentad http/
/www.cgd.ucar.edu/cas/catalog/ocean/godas_variable
.shtml
18
(No Transcript)
19
(No Transcript)
20
(No Transcript)
21
Monthly Ocean Briefing - A New CPC Product

• Background
• Air-sea coupled modes such as ENSO, IOD, TAV, PDO
  and AMO have robust impacts on climate.
• The coupled modes vary on interannual to decadal
  time scales due to both natural variability and
  anthropogenic forcings. It is imperative to
  monitor and assess those modes in real time to
  support operational climate predictions at CPC.
• A real-time assessment of CFSs performance in
  forecasting those coupled modes is desired.
• A real-time assessment of GODAS biases and their
  potential impacts on the seasonal forecasts of
  CFS is needed.
• The briefing
• Conference call on the 6th-9th day of every
  month.
• PowerPoint Presentation and Ocean Briefing web
  page
• Outside colleagues are welcome to contribute to
  and participate in the Ocean Briefing.

22
Future Plans

• A new version of the CFS, operational by 2010.
• Fully global with atmosphere, ocean, land and ice
  components
• ODAS based on MOMv4 / 3DVAR (NCEP)
• ODAS based on MOMv4 / KF (JPL)
• ADAS based on GFS / GSI
• Reanalysis (1980-present), 1st guess for each
  component from the CFS
• Retrospective forecasts (1980-present)
• A stand alone operational version of GODAS.
• Frees ocean analysis from long update cycle of
  the CFS and thus
• Serves as a useful platform for introducing new
  methods and data.
• Serves as a tool for CPC state of the ocean
  products and discussions.