Title: Climate sensitivity and variability with models extending into the middle atmosphere
1Climate sensitivity and variability with models
extending into the middle atmosphere
- F. Sassi
- National Center for Atmospheric Research
- Climate and Global Dynamics
2The impact of the middle atmosphere on
tropospheric climateThe climate sensitivity
- Climate sensitivity (CS) is a broadly used tool
to determine the models response to different
forcing (increase of GHGs, solar variability,
aerosols, etc.) - CS is defined as the equilibrium change of global
surface temperature following a doubling of CO2. - Gregory et al (2004) have defined a simple method
to calculate CS based on linear regression
analysis between the change of net heat flux at
the TOA and the change of global surface
temperature using an atmospheric model coupled to
a SOM. - This method has been used recently by Kiehl et al
(2006) to calculate the CS of the CCSM3.
3The impact of the middle atmosphere on
tropospheric climateThe simulations
- Run WACCM and CAM in similar configurations
- CAM3 SOM, present day CO2
- CAM3 SOM, 2xCO2
- WACCM3 SOM, present day CO2, full chemistry
- WACCM3 SOM, 2xCO2, full chemistry
- Although by and large the physical
parameterization that turn CAM into WACCM are
relevant only above the stratopause, WACCM is not
exactly identical to CAM - Efficiency parameter for orographic gravity waves
4Climate Sensitivity
CAM/ CS2.2
WACCM/ CS2.1
5Climate Sensitivity
- No significant changes of CS between CAM and
WACCM. - CS is a gross measure of climate change. No
change of CS between the two models reflects tiny
changes in all globally averaged fields (not
shown). - Regional effects?
6Surface Temperature ANN
7Surface Albedo ANN
- Both models predict an Arctic reduction of
albedo (? surfarce warming ? less sea ice) in the
doubled CO2 scenario. - The Arctic changes are greater in CAM than in
WACCM. - These changes have a seasonal cycle (not shown),
being more pronounced in DJF. - Implications for ocean circulation?
8DJF Sea Level Pressure
- Both models predict a reduction of sea level
pressure over the Arctic. - As before, changes are larger and more
ubiquitous in CAM than in WACCM. - Note that sea level pressure over the north
Atlantic is shallower in WACCM than in CAM.
91x CO2 Momentum Forcing
10DJF Zonal Mean Zonal Wind
- As the strength of the polar vortex decreases,
the sea level pressure is expected to decrease
stronger westward drag in the stratosphere ?
stronger mean meridional circulation ? mass
redistribution between polar and low latitudes. - This similar to the Polvani and Kushner
mechanism.
11Redistribution of mass in the vertical column
- The sea level pressure change WACCM CAM
poleward of 60N is 230 Pa ? a total mass
redistribution of 8E14 Kg. - About 50 of that mass change occurs in the
stratosphere.
Approx. location of tropopause at high latitudes
12DJF 500 hPa Geopotential Height
- Both models predict a thickening of the
atmosphere due to warming associated with the
doubling of CO2. - Model responses are different regionally CAM
shows a wave-2 (roughly) pattern with atmospheric
thickening occurring over the continents WACCM
pattern is weaker.
13DJF Zonal Mean Temperature
- Both models produce an increase of temperature
in the troposphere and a decrease in the lower
stratosphere. - Upper tropical tropospheric warming is larger in
CAM. - Tropical middle stratospheric cooling is larger
in WACCM.
14Cumulative precipitation DJF
151x CO2 Momentum Forcing
16EPD Difference
17NH Annular Modes
- Use daily data of geopotential interpolated to
standard pressure levels. - Take only data northward of 20N, zonally
averaged. - Calculate the composite annual cycle and the
anomalies against it. - Obtain the leading zonal mean pattern (EOF-1)
from the WACCM 1x simulation. - Project the pattern on the time series (? PC).
- Stratospheric influence on the troposphere
- Calculate lag correlation of all points vrs 10
hPa and near surface - Calculate composite of stratospheric weak and
strong jet events.
18Leading EOF
19Correlation with near surface events
- Correlation with near surface events is
amplified in CAM at lag zero Near 1 hPa, the
re-analysis show a correlation less than 0.1,
while it is gt 0.25 in CAM. - WACCM is much closer to the re-analysis.
- Both models overestimate the tropospheric
correlation.
ERA40
Christiansen 2005
20Correlation with stratospheric events
- Downward progression of stratospheric anomalies
is quite similar between the two models in the
stratosphere. - In the troposphere, there is no downward
progression in CAM/WACCM.
ERA40
Christiansen 2005
21Composite of weak stratospheric vortex events
- Both models show a relatively long persistence
of the anomalies in the lower stratosphere. This
is consistent with longer newtonian relaxation
rate. - The WACCM simulations show that tropospheric
anomalies lead the stratospheric events. This is
a realistic feature that is not reproduce in the
CAM. - Downward influence of stratospheric events on
the troposphere wanes rapidly in CAM. WACCM
suggests longer time scales.
22Composite of strong stratospheric vortex events
- There is little distinction between before and
after the events in the CAM.
23Stratospheric warmings
Model Frequency of SSW (no./year)
CAM 1x 0.14 (0.03)
WACCM 1x 0.12 (0.03)
24CONCLUSIONS
- Globally averaged measures of climate change are
identical in both CAM and WACCM ? the presence of
a well resolved middle atmosphere is irrelevant
to metrics like CS. - Regional metrics can be different in the two
models surface temperature, sea ice, sea level
pressure, zonal mean temperature. By and large,
the signatures of change are amplified in CAM
compared to WACCM. - By and large, stratosphere-troposphere coupling
is more realistic in WACCM than in CAM.