Evaluation of WBC representation in climate models

1
Evaluation of WBC representation in climate models

• LuAnne Thompson
• University of Washington
• Young-Oh Kwon
• Woods Hole Oceanographic Institution
• Frank Bryan, Steve Yeager, Julie McClean
• Western Boundary Current Workshop
• Phoenix, AZ
• January 2009

2
Climate change, WBCs and societal importance
(I.e. the next IPCC report) ? Decadal
predictability from WBCs Carbon and heat uptake
and storage in STMWs
3

• Questions
• What are the biases in climate models in the
  WBCs?
• What are the consequences of these biases?
• Can we do anything about them?

4
RMS Sea Surface Height Anomaly (cm)
Altimetry
See also OFES runs
CCSM4Yr 5
McClean CCSM Ultra High Resolution Simulation
Team
5
Example CCSM3.0

• 100 years of CCSM3.0 coupled ocean-atmosphere
  model with 10 ocean and T85 atmosphere (CCSM)
• 100 years of ocean-only with repeating forcing
  (POP)
• Review paper, Kwon et al, 2009

6
SST biases in CCSM3 and POP
7
MLD biases in CCSM3 and POP
8
STMW Salinity
Thompson and Cheng (2008) for NP
9

• Example 1
• North Pacific Mean Circulation

10
Mean Curl over North Pacific
CCSM3
POP
11
Sverdrup circulation
CCSM
POP
Sverdrup
Barotropic Streamfunction
12
Sea Surface Height model vs obs
CCSM
POP
Observations Kuroshio Extension at
350N Models Kuroshio Extension at 400N
Obs
13
Mean Path location POP and CCSM agree
POP
CCSM
Obs
14
One-D MLD modelSept-Sept Forcing ___ COADS- -
- Coupled Ocean Only
KE MLDs___ WOA- - - Coupled Ocean Only
WOA IC
POP IC
CPL IC
Thompson and Cheng (2008)
15

• Example 2
• North Pacific Decadal variability
• Thompson and Kwon, AMS, 2009

16
A decadal model of variability CCSM3
17
SST EOFs in North Pacific
EOF1 (43)
EOF1 (43)
EOF2 (11)

• EOF2
• (11)

Observations Deser and Blackmon (1995)
CCSM Alexander et al (2006)
18
SST variability max North of Kuroshio

• CCSM 1440E-1640E,
• 400N-450N
• Mean 100C
• RMS SST 10C
• GHRSST 1420E-1620E, 350N400N
• Mean 180C
• RMS 0.60C

CCSM
GHRSST
19
Model1440E-1640E RMS path 0.50
Path and SST 5 year low pass relationships
Path
SST
Degree C and Latitude
Obs 1420E-1620ERMS path 0.40
Year
20
Regression of annual SST on path index

CCSM

Obs
21
CM2.1 (GFDL) SST evidence in other models
Knutson et al 2006
22

• Example 3
• A fix to the Northwest Corner?
• Weese and Bryan (2006)

23
SST bias in run with dynamically adjusted NAC
Coupled
Ocean-only
Control
Adjusted
24

• Example 4
• Gulf Stream/Labrador Sea connections (Steve
  Yeager poster)

25
B Coupled Ocean-ice hindcast
A Ocean-only hindcast


DWBC
DWBC
WB
WB
MLD
MLD
Enhanced DWBC strength (and WB) off Grand Banks
gt improved NAC in 1o simulations DWBC
strength is set by Labrador Sea surface
transformation and convection
26
February Surface Density Flux (colors) SSD
(contours)

kg/m2/s
Comparison with an observed air-sea flux product
(LY08) shows that more realistic DWBC transport
and NAC path (B) is associated with excessive
Labrador Sea surface transformation. The
sensitivity of subpolar thermal transformation
to high latitude haline forcing (ocean-ice
salinity restoring fluxes) is explained by
feedbacks arising from mixed boundary conditions
Surface Diapycnal transformation rate
27
(Coupled) Climate Model biases Both GS and
KE Separation point and penetration of WBCs
warm upstream, cool downstream Missing eddies
weak WBC, weak isopycnal mixing into STG
interior Wind Stress biases SST/PBL biases
28
Gulf Stream Interaction of GS with topography
(Northwest Corner) Interaction of GS with DWBC
GS penetration and sea-ice cover Freshwater
forcing sensitivity Kuroshio Unrealistic (?)
decadal variability focused on KE Oyashio and
Kuroshio (and mode waters) not separated
29
Diagnostic Metrics for ocean models
relationships between variables
Heat transport convergence/surface heat
fluxes/heat content tendency (Heat Budget) Mode
water thickness/MLD/heat content (i.e. NA Mode
water thickness large when heat content
low) Interannual EKE/Path/Recirculation/Mode
Water Large scale wind-stress/path Path/transpor
t/SST Precip/SST Wind/SST Troposphere
omega/SST/Heat flux including mean seasonal and
synoptic dependence
30
Diagnostic Metrics for air-sea interaction
relationships between variables
Precip/SST Local Wind/SST Troposphere
omega/SST/Heat flux including mean seasonal and
synoptic dependence
31
Modeling studies and CPTs suggested by this
workshop
Representation of eddy (non-linear) effects in
low resolution ocean models separation,
acceleration and penetration vs. mixing
from WBC into subtropical gyre WBC influence on
troposphere in climate models via SST/wind
coupling in low resolution models Comparison/cla
rification of theories/models of air-sea
interaction in the mid-latitudes on all time
scales Intrinsic variability in the ocean
including buoyancy forcing Observing System
Simulation Experiments (OSSEs) Eddy permitting
data assimilation
32
Strawman modeling program (Rothstein)
Global studies Data fusion, synthesis and model
(in)validation Focus sites Numerical
methods Process studies Component and coupled
model evaluations
33
World Modelling Summit for Climate
PredictionReading, UK, 6-9 May 2008Summit
Statement The Climate Prediction Project1.
Climate modeling is constrained by limitations in
computer power and scientific understanding. 3.
Climate prediction is among the most
computationally demanding problems in science.
4. The Summit strongly endorsed the initiation
of a Climate Prediction Project coordinated by
the World Climate Research Programme. The goal
of the project is to provide improved global
climate information to underpin global mitigation
negotiations and for regional adaptation and
decision-making in the 21st century.9.
Sustained, long-term, global observations are
essential to initialize, constrain and evaluate
the models. 10. To estimate the quality of a
climate prediction requires an assessment of how
accurately we know and understand the current
state of natural climate variability, with which
anthropogenic climate change interacts. 11.
Advances in climate prediction will require close
collaboration between the weather and climate
prediction research communities.