My Ocean User Meeting, Stockholm, 7-8 April 2011 - Seasonal Forecasting - 1
Delayed Ocean Analysis ~12 days
Real Time Ocean Analysis ~Real time
ECMWF:Weather and Climate Dynamical Forecasts
ECMWF:Weather and Climate Dynamical Forecasts
Medium-Range (10-day)Partial coupling
Medium-Range (10-day)Partial coupling
Seasonal ForecastsFully coupled
Seasonal ForecastsFully coupled
Extended + Monthly Fully coupled
Extended + Monthly Fully coupled
Ocean model
Atmospheric model
Wave model
Atmospheric model
Ocean model
Wave model
My Ocean User Meeting, Stockholm, 7-8 April 2011 - Seasonal Forecasting - 2
• Observations
� Quality controlled: (bias correction, black listing) retrospective and real time
� Profiles of T/S, SL from Altimeter, MDT, SST, sea-ice concentration-thickness, possibly ocean colour (climatology, maps)
• Ocean re-analysis:
� Initialization, verification, calibration of seasonal (monthly, decadal) forecasts
� Long and Consistent records
� Estimates of uncertainty
� Timely
� Compatibility with ocean model in forecasting system
• Software sharing & RD development
� Ocean Model (example NEMO). Software and configurations
o High horizontal/vertical resolution needed. “Off MyOcean Shelf “
� Data assimilation
• Outlook:
� explicit/improved representation of air-sea interaction for all forecast ranges and data assimilation
Seasonal Needs: Outline & Summary
My Ocean User Meeting, Stockholm, 7-8 April 2011 - Seasonal Forecasting - 3
End-To-End Seasonal forecasting System
EN
SE
MB
LE
GE
NE
RA
TIO
N
COUPLED MODEL Tailored Forecast
PRODUCTS
Initialization Forward Integration Forecast Calibration
OCEAN
PR
OB
AB
ILIS
TIC
CA
LIB
RA
TE
D F
OR
EC
AST
JUL2006
AUG SEP OCT NOV DEC JAN2007
FEB MAR APR MAY JUN JUL AUG SEP
-1
0
1
2
Ano
mal
y (d
eg C
)
-1
0
1
2
Monthly mean anomalies relative to NCEP adjusted OIv2 1971-2000 climatologyECMWF forecast from 1 Jan 2007
NINO3.4 SST anomaly plume
Produced from real-time forecast data
System 380°S80°S
70°S 70°S
60°S60°S
50°S 50°S
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
70°N 70°N
80°N80°N
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°E 160°E
160°E 180°
180° 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
14.3 10.39.8 13.325.1 26.22 2.9
No Significance 90% Significance 95% Significance 99% Significance
FORECAST CLIMATE
My Ocean User Meeting, Stockholm, 7-8 April 2011 - Seasonal Forecasting - 4
0 1 2 3 4 5 6Forecast time (months)
0
0.2
0.4
0.6
0.8
1
1.2
Rm
s er
ror (
deg
C)
Ensemble sizes are 5 (0001), 5 (0001) and 5 (0001) 64 start dates from 19870401 to 20021201
NINO3 SST rms errors
Fcast S3 Fcast S2 Fcast S1 Persistence Ensemble sd
•Steady progress: ~1 month/decade skill gain
•How much is due to the initialization, how much to model development?
S1 S2 S3
TOTAL GAIN
OC INI
MODEL
0
5
10
15
20
25
30
35
40
1%
Relative Reduction in SST Forecast ErrorECMWF Seasonal Forecasting Systems
TOTAL GAIN OC INI MODEL
Half of the gain on forecast skill is due to improved ocean initialization
A decade of progress on ENSO prediction
Balmaseda et al 2010
My Ocean User Meeting, Stockholm, 7-8 April 2011 - Seasonal Forecasting - 5
Dealing with model error: Hindcasts
Ocean
reanalysis
Coupled Hindcasts, needed to estimate climatological PDF,
require a historical ocean reanalysis
Real time Probabilistic
Coupled Forecast
time
Consistency between historical and real-time initial conditions is required.
Hindcasts are also needed for skill estimation
My Ocean User Meeting, Stockholm, 7-8 April 2011 - Seasonal Forecasting - 6
Real Time Ocean Observations
ARGO floats XBT (eXpandable BathiThermograph)
Moorings
Satellite
SST
Sea Level
My Ocean User Meeting, Stockholm, 7-8 April 2011 - Seasonal Forecasting - 7
Skill: Ocean observations and Winds
Increase (%) in MAE of SST forecasts from removing external information
(1-7 months)
-10
-5
0
5
10
15
20
25
30
NIN
O3
NIN
O3.
4
NIN
O4
TR
PA
C
EQ
IND
IND
1
IND
2
NS
TR
AT
L
EQ
AT
L
%
OC DATA
WINDS
DATA+WINDS
•Both Wind and ocean observations contribute to the initialization
•No Observing system is redundant
•The winds from ERA-Interim produce better forecast over the Equatorial Atlantic.
ERA40+OPS
ERA-INTERIM
EQATL SST Forecasts correlation
1989-2006
Balmaseda et al 2010
Dee et al 2011
My Ocean User Meeting, Stockholm, 7-8 April 2011 - Seasonal Forecasting - 8
Producing Reliable Forecasts: Calibration
Persistence
ECMWF
ensemble spread
RMS error of Nino3 SST anomalies
Bayesian Calibration
EUROSIP
Usually, the forecast systems are not reliable: RMS > Spread
Can we reduce the error? Or Can we increase the spread?
Multimodel
Calibration. Long records of initial conditions are needed for calibration
My Ocean User Meeting, Stockholm, 7-8 April 2011 - Seasonal Forecasting - 9
Which SST product to use?
• OIV2_025_AVHRR: bias cold in the global mean (regional differences) respect cmip5-proto
•Bias decreases with time. Weaker interannual variability
•Fit to insitu Temperature: bias cold in tropics, better in mid latitudes.
• Clear need for GLOBAL, CONSISTENT, UNBIASED, LONG records of HIGH SPATIAL and TEMPORAL RESOLUTION (1/4 of Degree, daily)
• Consistent and GLOBAL SEA-ICE concentrationOSTIA SST from 2008, similar to oiv2_025_AVHR_AMSR
My Ocean User Meeting, Stockholm, 7-8 April 2011 - Seasonal Forecasting - 10
Impact of NEMOVAR on Seasonal Forecasts
Prototype of S4: latest NEMOVAR+36r4. Anomaly Correlation
0 1 2 3 4 5 6 7Forecast time (months)
0.4
0.5
0.6
0.7
0.8
0.9
1
Ano
mal
y co
rrel
atio
n
wrt NCEP adjusted OIv2 1971-2000 climatologyEQ2 034a anomaly correlation
0 1 2 3 4 5 6 7Forecast time (months)
0.4
0.5
0.6
0.7
0.8
0.9
1
Ano
mal
y co
rrel
atio
n
wrt NCEP adjusted OIv2 1971-2000 climatologyATL3 034a anomaly correlation
0 1 2 3 4 5 6 7Forecast time (months)
0.4
0.5
0.6
0.7
0.8
0.9
1
Ano
mal
y co
rrel
atio
n
wrt NCEP adjusted OIv2 1971-2000 climatologyEQIND 034a anomaly correlation
0 1 2 3 4 5 6 7Forecast time (months)
0.4
0.5
0.6
0.7
0.8
0.9
1
An
oma
ly c
orr
elat
ion
wrt NCEP adjusted OIv2 1971-2000 climatology
NSTRATL 034a anomaly correlation
0 1 2 3 4 5 6 7Forecast time (months)
0.4
0.5
0.6
0.7
0.8
0.9
1
Ano
mal
y co
rrel
atio
n
wrt NCEP adjusted OIv2 1971-2000 climatologyNSTRPAC 034a anomaly correlation
0 1 2 3 4 5 6 7Forecast time (months)
0.4
0.5
0.6
0.7
0.8
0.9
1
Ano
mal
y co
rrel
atio
n
wrt NCEP adjusted OIv2 1971-2000 climatologySSTRATL 034a anomaly correlation
NEMOVAR NEMO-NoObs
Software products: ECMWF uses NEMO (ORCA1 configuration)
ECMWF uses NEMOVAR (collaborative project)
CENTRAL EQ. PACIFIC CENTRAL EQ. ATLANTICEQ. INDIAN
N SubTrop PACIFIC N SubTrop ATLANTIC S SubTrop ATLANTIC
My Ocean User Meeting, Stockholm, 7-8 April 2011 - Seasonal Forecasting - 11
Time scales for ocean-atmosphere interaction
ATM delay:
days-weeks
OCN delay:
Hours-days-decades
ATM forcing
OCN response
OCN forcing
ATM response
days weeks Months/years Decades and beyond
Tropical cyclones
Surface waves
Diurnal Cycle
Madden-Julian Oscillation
Tropical Instability Waves
Equatorial Ocean Dynamics:
ENSO, IOD
Seasonal ML variations: NAO?
Subtropical Gyre, Rossby Waves, THC, MOC
Pacific/ Atlantic Decadal Variability
Boundary layer processes
Heating/cooling
Evaporation/precip
Momentum transfer
My Ocean User Meeting, Stockholm, 7-8 April 2011 - Seasonal Forecasting - 12
Monthly Forecasts needs
• Madden Julian Oscillation (MJO) is corner stone for monthly forecasting (as ENSO is for seasonal)
• It influences NAO regimes (Cassau et al 2008) and predictability over Europe (Vitart)
• MJO forecasts needs interactive ocean, good representation of ocean mixing (high vertical
resolution)
Woolnough et al, MWR 2007
Anomaly Correlation
Persisted SST anomalies
OGCM (10 m vertical res)
Mixed layer (1 m vertical res)
My Ocean User Meeting, Stockholm, 7-8 April 2011 - Seasonal Forecasting - 13
O-A interaction over SST fronts
Air-Sea Interaction also occurs at small scales, such as that of the Western Boundary currents (above) and Tropical Instability Waves TIW (left). The small scales are set up by the ocean
Need of high resolution ocean models
Minobe et al, Nature, 2008
Chelton and Song
My Ocean User Meeting, Stockholm, 7-8 April 2011 - Seasonal Forecasting - 14
Winds, Waves, Currents
Bidlot, 2010
Wind
Neutral Wind ~stress
Wave Height
Wind, Waves and ocean currents need to be treated in coupled mode
Additionally, the effect of waves on ocean mixing should be considered.
My Ocean User Meeting, Stockholm, 7-8 April 2011 - Seasonal Forecasting - 15
Saha et al BAMS 2010
Tropical Precipitation and SST relationship
OBS
CFSR
NCEP R1
NCEP R2
Atmospheric models forced by observed SST do not capture the tropical SST/precip lag relationship
This problem was also evident int the uncoupled NCEP atmospheric re-analysis.
The new CFSR is an improvement in this respect.
My Ocean User Meeting, Stockholm, 7-8 April 2011 - Seasonal Forecasting - 16
• Observations
� Quality controlled: (bias correction, black listing) retrospective and real time
� Profiles of T/S, SL from Altimeter, MDT, SST, sea-ice concentration-thickness, possibly ocean colour (climatology, maps)
• Ocean re-analysis:
� Initialization, verification, calibration of seasonal (monthly, decadal) forecasts
� Long and Consistent records
� Estimates of uncertainty
� Timely
� Compatibility with ocean model in forecasting system
• Software sharing & RD development
� Ocean Model (example NEMO). Software and configurations
o High horizontal/vertical resolution needed. “Off MyOcean Shelf “
� Data assimilation
• Outlook:
� explicit/improved representation of air-sea interaction for all forecast ranges and data assimilation
Seasonal Needs: Outline & Summary