Spatial distribution of trends in heat content

1
Spatial distribution of trends in heat content
1982-2006 mean minus 1959-1981 mean
SST (deg C)
Taux (x 0.01N/m2)
Tauy (x 0.01N/m2)
T300 (deg C)
How reliable are the trends in ERA40 winds?
2
Comparison with ocean observations
IPCC-AR4 (LEVITUS)
ORA-S3
CI0.05 deg/decade

• Similarities
• Equatorial cooling
• Warmer subtropics
• Cooling at 60N

• Comments
• Trends in ERA40 winds seem robust
• Stronger features in ORA-S3, more structure
• Circulation changes as well as mixed layer changes

3
Trends in the Pacific Ocean

• Cooling of the Equatorial ocean is a robust
  feature, captured by both sub-surface
  observations and ocean model forced by
  ERA40-winds
• Shallower and wider equatorial meridional cell
• Warming in the Subtropical Pacific is robust
• Cooling in the North Pacific

Needs for decadal predictions

• What signals/modes are predictable at decadal
  time scales?
• Which observations are needed for decadal
  prediction?

4
ENSO and changing climate

• How to provide robust assessment and routine
  monitoring of changes?
• Need for improved ocean and atmospheric
  reanalyses
• Should/can we do attribution of the observed
  changes?
• Start validating the AR4 predictions?
• Is there a climate shift in 76/77? Or part of a
  continuous process?
• What are the implications of the observed changes
  for ENSO and ENSO teleconnections?
• Apply the BJ index to existing reanalysis for the
  past and recent period?

5
Status of ENSO forecasting

• Progress in forecasting SST is steady
• Improved forcing fluxes yield better forecast
  skill
• All the components of the ocean observing system
  have a positive impact on the forecast skill.
• Model error/initialization shock is a serious
  problem in the Eastern Pacific.
• Low amplitude in the intraseasonal time scales
  can degrade the prediction of the interannual
  variability
• The response of the atmosphere to the SST forcing
  is deficient and limits the seasonal forecast
  skill

6
How to improve further?

• More sophisticated assimilation methods are
  needed for a seamless prediction system
• A balanced initialization does not mean using
  less information about the real world, but
  adequate mapping between the observed state and
  the model state.
• The term coupled intialization is too vague.
  Need to qualify different coupled
  initialization strategies.
• Need for a workshop on Coupled Model
  INITIALIZATION?
• Better Coupled models are needed to improve the
  Seasonal Forecasts of T2m and precipitation.
• Improved atmospheric and ocean reanalyses are
  needed
• Model improvement
• Initialization
• Calibration of model output for reliable seasonal
  forecasts

7
Progress in ENSO forecast skill
Rms error of forecasts has been systematically
reduced (solid lines) .
8
Contribution of model improvement versus initial
conditions
S2 S2ic_S3model S3
Improvements in the coupled model (green versus
blue) are comparable to improvements in the
initial conditions (blue versus purple)
9
Perceived Paradigm for initialization of coupled
forecasts
Real world
Model attractor
Medium range Being close to the real world is
perceived as advantageous. Model retains
information for these time scales. Model
attractor and real world are close?
Decadal or longer Need to initialize the model
attractor on the relevant time and spatial
scales. Model attractor different from real
world.
Seasonal? Somewhere in the middle?
At first sight, this paradigm would not allow a
seamless prediction system.

• Experiments
• Uncoupled SST Wind Stress Ocean Observations
  (ASSIM)
• Uncoupled SST Wind Stress (WIND)
• Coupled SST (COUP)

10
Impact of external real world information
ASSIM WIND COUP

• Need better (more balanced) initialization
• More information corrects for model error, and
  the information is retained during the fc.
• Model errors that can not be corrected by
  initialization (intraseasonal variability)

• Relation between drift and Amplitude of
  Interannual variability.
• Possible non linearity is the warm drift
  interacting with the amplitude of ENSO?
• Other source of errors even with the correct
  mean state the I.V amplitude is small. MJO?

11
Impact of real world information on skill
NINO3.4 RMS ERROR ASSIM WIND COUP
The additional information about the real world
improves the forecast skill, execept in the
Equatorial Atlantic
However optimal use of the observations may
require more sophisticated assimilation
techniques, able to map the observation space
into the model space
12
Impact of Observing system on Seasonal Forecast
skill

• Moorings only the effect of anomalies is
  measured, since the effect of the mean state is
  included indirectly in the altimeter
  assimilation.
• Observing systems are complementary
• Altimeter has larger effect on Atlantic and
  Eastern Pacific
• Argo has larger effect on Indian Ocean and
  Western Pacific

13
Statistics for the period 1993-2007

• Impact of data assimilation depends on the
  region
• Central/Eastern Pacific assimilating data
  improves the seasonal forecast skill, by
  correcting the mean state and anomalies (NINO4)
• In the Eastern Pacific (NINO1/2, NINO3) the
  impact of better initialization of the anomalies
  is hampered by degradation of the mean state. The
  assimilation of data results in warmer bias (not
  shown), which is detrimental for ENSO prediction

14
Summary of ENSO prediction

• Most common ocean initialization strategy is the
  uncoupled initialization
• Ocean observations are assimilated into an ocean
  model forced by atmospheric fluxes.
• In general, this strategy improves the forecast
  skill in the prediction of SST (if the coupled
  model is good/discerning enough).
• If there are serious model errors this strategy
  can lead to large initialization shocks and
  degradation of the skill (Equatorial Atlantic).
• The skill of seasonal forecasts of SST is
  steadily improving due to
• Improved quality of coupled models
• Improved quality of atmospheric reanalysis
• Improved ocean observing system (contribution of
  ARGO and Altimeter add to the moorings)
• Improved ocean assimilation systems.
• More sophisticated assimilation methods are
  needed for a seamless prediction system
• A balanced initialization does not mean using
  less information about the real world, but
  adequate mapping between the observed state and
  the model state.