Transcript Slide 1

ENSO modelling at
Hadley Centre and NCAS
Sarah Ineson - with thanks to Mike Davey, David Fereday, Tim
Graham, James Lloyd*, Dave Rowell, Andy Turner*
*NCAS
© Crown copyright Met Office
November 2010
Outline
HadGEM3 development
• Literature review – 3 papers on impact of
resolution
• HadGEM3 development cycle
• ENSO results
• Focussed scientific development
• ENSO Issues
© Crown copyright Met Office
Model resolution (1) - ocean
Impact of resolution of the tropical
Pacific circulation in a matrix of coupled
models. Roberts et al. 2009 J.Clim.
The mean state is improved in
the models with higher ocean
resolution
This is due to tropical
instability waves which reduce
the cold tongue bias
© Crown copyright Met Office
Model resolution (2) - atmosphere
Representing El Niño in coupled ocean-atmosphere GCMs:
The dominant role of the atmospheric component. Guilyardi et
al. 2004. J. Clim
Atmosphere model is
setting the frequency –
regardless of the ocean
The various resolutions of
ECHAM4 influence frequency
- T106 has power close to
observations
© Crown copyright Met Office
Model resolution (3) - stratosphere
collapse of stratospheric
vortex
+ve GPH
Rossby wave driving
-ve (easterly) wind
deeper Aleutian low
-ve PMSL
-ve NAO
cold northern European
temperatures in late winter
El Niño
+ve SST
© Crown copyright Met Office
A well resolved stratosphere is needed to
simulate the observed European response
to El Niño. Bell et al. 2009, Cagnazzo and
Manzini 2009, Ineson and Scaife 2009.
HadGEM3 model development –
an ongoing process
Model development process
Parameterisation
and dynamical
changes
Focussed work on
systematic errors
Technical
upgrades
Model Development Team
HadGEM3
HadGAM3-A
CICE
NEMO
PAT
Report to
(Meets 6-monthly.
Scientific
review/challenge;
recommends releases of
model versions)
(Meets monthly and decides what
to include and do next)
N48ORCA2
development
N96ORCA1
development
N216ORCA0.25
development
Seasonal model
development
Assessment of model output
(major assessment ~6-monthly)
Regional model
development
PB
(Meets 6-monthly.
Approves decisions and
deals with resources)
GloSea4 operational seasonal
forecast system - N96ORCA1
© Crown copyright Met Office
ENSO Metrics
Obs
GEM1
GEM2
ahsaf
ahuna
ahuqj
ahsak
aitcb
aicvx
xepec
N96
L38
O1
L42
N96
L63
O1
L42
N96
L38
O1
L42
N96
L85
O1
L42
N96
L85
O1
L42
N96
L85
O1
L75
N216
L85
O025
L75
100 yr
30 yr
40 yr
65yr
60 yr
100 yr
70 yr
2001-2070
Nino3 standard
deviation
0.80
0.65
0.76
0.70
0.65
0.87
0.67
Nino3 power
spectrum
timescale (yrs)
3.5
5.3
4.5
4.1
4.9
6-11
3.3
5.0
3.8
3.8
3.2 5.1
Ann. Mean Nino3
SST
25.9
23.7
24.9
24.1
24.3
24.8
25.2
Ann. Mean Nino4
TAUX
-0.029
-0.064
-0.048
-0.042
-0.045
-0.038
Correlation
Nino3 SST/SOI
-0.62
-0.31
-0.45
-0.38
-0.36
-0.33
© Crown copyright Met Office
0.66
0.67
0.79
5.6
5.0
25.4
25.7
26.1
-0.041
-0.037
-0.038
-0.033
-0.37
-0.39
-0.37
-0.41
Comparison of standard and high
resolution ENSO composites
Composite of DJF El Niño and La Niña events (>1.2σ)
HadISST
© Crown copyright Met Office
N96L85O01L75
N216L85O025L75
westward extension is much
improved in high resolution
Higher resolution improves
rainfall teleconnection response
Dave Rowell
to global SST
OBS
Example: ENSO/southern Africa, DJF
HadGEM3-N96 (~120km, 1°)
HadGEM3-N216 (~60km, 0.25°)
Based on 75 years of CRU observations (1921-1995 ) and model integrations
© Crown copyright Met Office
Testing new GloSea4 using winter
2009/10
David Fereday
Old System
New System
1.5m T
© Crown copyright Met Office
Focussed Scientific Development ENSO and ENSO teleconnections
ENSO mean state and variability:
• understanding the tropical Pacific SST cold bias (Tim Graham)
• the easterly equatorial wind bias (Dan Copsey)
• role of cloud feedback in ENSO (James Lloyd)
• understanding the role of resolution (Sarah Ineson, Mike Davey)
ENSO teleconnections
• Analysis of the Asian monsoon response to ENSO in the Met Office
GloSea4 seasonal hindcasts (Andy Turner)
© Crown copyright Met Office
TIW heat budget
Tim Graham
Following Menkes et al. (2006) – use a moving box composite
to estimate the heat budget of TIWs
ORCA1
© Crown copyright Met Office
ORCA025
Atmospheric feedbacks during
ENSO - James Lloyd
μ (10-3Nm-2C-1)
α (Wm-2C-1)
Taux ' = μ SST ' (Niño3) +ve feedback
Q ' = α SST ' (Niño3)
-ve feedback
obs
13.2 ERA40
- 19.4 OAFlux
- 26.4 ERA40
μ and α are underestimated in the models
aisas
8.0
- 8.07
xemjo
10.5
- 13.7
α NET = α SW + α LH + α LW + α SH
Observations (- 8.9)
© Crown copyright Met Office
CGCM (+ 5.3)
AMIP (0.0)
The shortwave response in the CGCM will tend
to grow rather than terminate an event
Understanding the role of
resolution
standard
X-res (std A, high O)
high
© Crown copyright Met Office
Summary
• HadGEM3 model development occurs within a
cyclic and structured project
• Focussed Scientific Development – lead to
better understanding and model improvements
• High resolution (atmos+ocean) HadGEM3 –
improves ENSO - amplitude and spatial
structure - and ENSO teleconnections
© Crown copyright Met Office
ENSO Issues
• What GCM resolution is required? – can we represent all the
key processes and parametrize those which aren’t resolved
• What initiates an ENSO event?
• What are the mechanisms that allow an event to develop?
• Are there circumstances for which ENSO is more (or less)
predictable?
• Are the ENSO impacts represented properly in models?
© Crown copyright Met Office
© Crown copyright Met Office
Annual mean equatorial SST and
TAUX
© Crown copyright Met Office
ENSO phase locking
Niño3 standard deviation for each month
HadISST
© Crown copyright Met Office
xepec