Tropical Atlantic Variability as Simulated by CAM3CCSM3

1
Tropical Atlantic Variabilityas Simulated by
CAM3/CCSM3

• R. Saravanan
• National Center for Atmospheric Research

2
TAV
3
Intercomparison of Simulated Equatorial
SST(STOIC Davey et al., 2002)
4
Topics

• Effect of model improvements
• Role of atmospheric model resolution
• Potential predictability

5
Model integrations

• Coupled Experiments
• CCSM1 (T42), CCSM2 (T42), CCSM3 (T85)
• Uncoupled experiments (1950-2000)
• Vanilla AMIP CCM3 (T42), CAM3 (T42/T85)
• IPCC AMIP CAM3 (T85)

6
March-April-May mean SST
OBS
CCSM1
CCSM2
CCSM3
7
March-April-May SST std. Dev.
OBS
CCSM1
CCSM2
CCSM3
8
March-April-May mean PRECIP
OBS
CAM3/T42
CAM3/T85
CCSM3
9
Potential Predictability

• P E/T - I/N
• E ensemble-mean
  variance
• I internal variance
• T total variance (
  E I)
• N no. of ensemble
  members
• Effect of model version
• Effect of model resolution
• Effect of forcings

10
Potential Predictability of PS
MAM
JJA
CCM3 Vanilla AMIP
CAM3 Vanilla AMIP
CAM3 IPCC AMIP
11
Conclusions

• Some improvement in the coupled model simulations
  of Tropical Atlantic SST
• Significant biases remain, both in the mean and
  in the variance
• Coupled CCSM3 seems to have less of a double ITCZ
  structure than uncoupled AMIP runs!
• Potential predictability in the Atlantic Sector
• Not affected by model resolution
• Not affected by presence of presence of
  Greenhouse Gas forcings etc.
• But tropical SST signal is important

12
Tropical Atmospheric Bridge(Klein et al, 1999)
13
NAO
14
CCSM2 SST Bias
DJF
JJA
15
Questions

• Evidence for remote influence of Tropical
  Atlantic SST anomalies on the Tropical Pacific
• Seen in CAM2slab, CAM2POP, CCM3slab, and
  CCM3MOM. How about other models?
• Observational data dominated by ENSO signal

16
(No Transcript)