BLUElink

1
Towards merging GHRSST and GODAE for SST
forecasting
Brassington, Pugh, Beggs and Oke Bureau of
Meteorology CSIRO Marine and Atmospheric Research
2
Outline

• Pathways for GHRSST gt GODAE
• GODAEs role in SST
• Bureau systems
• Ocean forecasting for Australia, OceanMAPS
• Version 1.0
• First forecasts
• Assimilating SSHA and SST

3
Pathways for GHRSST gt GODAE
GHRSST gt GODAE Validation of GODAE
products GHRSST gt NWP gt GODAE Analysed surface
fluxes GHRSST gt GODAE ltgt NWP Assimilated SST,
improved analysed and forecast currents Forecaste
d SST GHRSST gt GODAE ltgt WAM ltgt NWP Forecast
currents for wave refraction GHRSST gt BRAN ltgt
GODAE Reanalysed SST, feedback on GODAE system
design GHRSST gt MCC gt GODAE Filling the data
gap in altimetry
4
GODAEs role in SST
GODAE gt GHRSST Dynamical background
analysis Methodology for SSHA GODAE offers a
forecast capability for SST NWP relies on a
static SST analysis for forecast b.c.s Skill
threshold some way off for NWP uptake Improved
coupled NWP fluxes benefit back to GODAE GODAE
ideally should be able to match GHRSST
analyses Several impediments
5
Dynamical interpolation uncertainties
Sources of uncertainty NWP fluxes Mixed layer
scheme and other paramterisations Predictability
of ocean dynamics Model resolution Assimilation
method Specification of covariances
cov(SSHA,SST) Pluses Non gaussian
covariances e-folding scale defined by model
variability Multi-variate covariances BODAS
continues to prove it is a good strategy for
ocean forecasting
6
Bureau systems
Ocean, Analysis and Prediction System
(OceanMAPS) Brassington et al High Resolution
Sea Surface Temperature (HRSST) Beggs et
al Australian Wave Model (AusWAM) Greenslade et
al Global atmospheric prediction system
(GASP) Seaman et al Moving to UKMet Office
UM Coupled limited area model TC-LAPSltgtAusWAMltgt
OceanMAPS
7
Ocean Model, Analysis and Prediction System
(OceanMAPSv1)
OFAM MOM4p0d 1/10ºx1/10º (90E-180E,
70S-16N) 10m (0-200m) BODAS Multi-variate
optimal interpolation (T, S, eta) Model error
covariances gt 72 member ensemble of
anomalies /- 5days altimetry Localisation
8ºx8º Background, daily average Surface
fluxes GASP Observations Jason1, Envisat SSHA
products AMSR-E descending track GTS, GDAC Argo
in situ
8
OceanMAPS schedule
9
Current status and performance
10
Case study monster eddy
11
Case study EAC drifter experiment

• EAC drifter experiment
• 8 drifters deployed from the PX30 line

12
SSHA SST gt SST nowcasts/forecasts
BLUElink has demonstrated the advantage of GHRSST
products to an ocean re-analysis (a) Removes
obvious biases (b) Multi-variate does modify the
sub-surface structure (Oke) (c) Modifies near
surface currents, quantitative improvement and
indications of skill over persistence Positive
impact has accelerated implementation into
OceanMAPSv1.0 Results have translated to removal
of bias Availability of GHRSST products made
this feasible AMSR-E 25km resolution matches
OFAM and coverage Microwave data gap is a concern
13
Surface ocean currents
14
Impact of HRSST for LAPS
15
Impact of HRSST for LAPS
16
GHRSST requirements
OceanMAPS L2P or L3P foundation for direct
assimilation L4 foundation and skin for
validation Error bars - normalised Documentation
Timeliness (Real-time to 10 days
behind) Resolution (1/16 degree) Diurnal model
(model foundation to skin) Data
gap BRAN Reanalysed L2P or L3P
17
Conclusions
GHRSST-PP is serving the ocean prediction
community well Unlocked the power of the
observation to the non-specialist Plans and
funding for continuity of GHRSST products very
positive Thank you Very rapid implementation
of AMSR-E OceanMAPS and BRAN demonstrating clear
improvements Requirements for GHRSST products
will continue to grow BODAS dynamical based
analysis scheme very positive results need to
be optimised to control uncertainties provide
many advantage