Towards stability metrics for cloud cover variation under climate change

1
Towards stability metrics for cloud cover
variation under climate change

• Rob Wood, Chris Bretherton, Matt Wyant, Peter
  Blossey
• University of Washington

2
Stability and low clouds, history

• Slingo (1980)
• Used model potential temperature lapse rate
  between 700 and 850 hPa as predictor of
  subtropical low cloud cover
• Cloud data from satellites (Miller and Feddes
  1971)

3
Stability and low clouds, history

• Klein and Hartmann (1993)
• Used potential temperature difference between 700
  hPa and the surface (lower tropospheric
  stability) as predictor of subtropical low cloud
  cover
• Cloud data from volunteer ship observations
  (Warren cloud atlas)

4
Stability, low cloud, and climate change

• Miller (1997) thermostat hypothesis
• Static stability expected to increase in warmed
  climate
• Cloud amount vs stability relationships indicate
  more cloud
• Negative cloud feedback
• Is LTS a suitable predictor of cloud responses to
  climate change?

5
Inversion strength and stability
Wood and Bretherton (2006)

• In the free-troposphere, d?/dz?FT follows a
  moist adiabat from 700 hPa to the MBL top.
  
• Well mixed surface layer below the LCL, i.e.
  d?/dz0
• In the cloud layer, d?/dz?CL follows a moist
  adiabat from the top of the LCL to the MBL top.

??
6
EIS, a measure of inversion strength ??

• ??(?700 ?0) ??FT?(z700 zi) ??CL?(zi
  zLCL)
• (?700 ?0) zi(??FT? ??CL?)
  ??FT?z700??CL?zLCL
• (a) Neglect term with zi as this term is
  generally small
• (b) Replace ??FT? and ??CL? with a single moist
  adiabat ?850?m(T0T700/2,
  850 hPa)
• (c) Assume surface RH of 80 to estimate zLCL
• Then define an estimated inversion strength (EIS)
  as
• EIS LTS ?850(z700
  zLCL)
• EIS solely a function of surface and 700 hPa
  temperatures for a reference surface pressure
  p01000 hPa

7
LTS/EIS and low cloud amount
Subtropical and tropical
Midlatitude
EIS is a far better predictor of low cloud amount
than LTS over a wider temperature range
8
Low cloud amount in an ensemble of 2xCO2-control
GCM simulations is poorly estimated using LTS
(for which a general increase is predicted)Much
better agreement with change in saturated
stability (related to EIS)
Change in LTS (K)
Williams et al. (2006)
9
Multiscale approach to the problem

• We use a suite of model simulations
• Climate model runs from CAM and GFDL (SST2K and
  2?CO2/SOM, Wyant et al. 2006)
• SP-CAM global run with SST2K (Wyant et al.
  JAMES) and 4?CO2/fixed SST change.
• CRM and LES runs using forcings derived from
  SP-CAM simulations for different stability
  percentiles (Blossey et al. JAMES)
• Uses only data from tropics (30oS-30oN)

10
Cloud amount vs LTS (SST2K)
Klein and Hartmann
SST2K
control
11
Cloud amount vs EIS (SST2K)
12
Cloud vs LTS (SST 2K, with CRM and LES runs)
13
Cloud vs EIS (SST 2K, with CRM and LES runs)
14
EIS and cloud changes
The three models studied here have significant
increases in EIS However, most climate models
show decreasing SWCF in the tropics in AR4
runs Why the discrepancy?
from Bony and Dufresne (2005)
15
SST2K vs 2xCO2/SOM
Somewhat weaker low cloud changes for 2xCO2 runs
16
SST2K vs 4xCO2/fixed SST (SP-CAM)
Completely different low cloud changes for 4xCO2
runs
17
MBL depth for control and perturbed runs
SST2K
MODIS Obs
NCO2
MBL depth decreases despite reduced subsidence
from CO2 FT warming ? MBL turbulence weakens
18
Conclusions, SST2K

• CAM3, AM2, and SP-CAM under SST2K show large LTS
  increases while low cloud cover changes increase
  more slowly than predicted by LTS
• These models under SST2K all show increases in
  EIS too. Cloud changes in CAM and SP-CAM increase
  somewhat more rapidly than predicted by EIS
• CRM driven by SP-CAM output consistent with
  SP-CAM
• LES driven by SP-CAM output not consistent with
  SP-CAM

19
Conclusions

• Cloud responses to changing CO2 very different
  from those due to SST changes, even in slab-ocean
  models. CO2 induces additional atmospheric
  radiative forcing at the top of the MBL in
  addition to warming the surface.
• CO2 perturbs the relationship between MBL depth
  and EIS whereas SST2K does not
• MBL depth changing for a given EIS consistent
  with cloud changes even at constant EIS
• Hypothesize that a single metric may be
  insufficient to capture the low cloud changes
  from radiative forcing by CO2 and from increased
  SST.

20
Doubling CO2 stabilizes lower troposphere
independent of SST changes

• Standard tropical profile, Fu-Liou RT model

21
A possible two-metric phase space
low cloud cover increasing
NCO2 with SST
free trop. rad. heating/LW down changes
NCO2 fixed SST
SST2K
EIS, stability-driven changes
22
The end
23
Why different sensitivity for SST2K vs NxCO2?
Models all show decreasing MBL depth (defined as
50 RH level) with increasing stability
24
Comments on previous

• CRM forced from SP-CAM has larger low cloud
  amounts, while LES gives much smaller cloud
  amounts than parent model.
• SWCF for CRM similar to that from parent model,
  while LES is still much weaker
• Climate changes in CRM similar to that in parent
  model

25
Comments on previous

• All models show decreasing MBL depth with
  increasing EIS
• SST2K runs tend to deepen MBL depth or stay same
  whereas NxCO2 runs tend to result in shallower
  boundary layers.
• The SP-CAM results particularly troublesome to
  interpret as 4xCO2 run shows huge decreases in
  MBL depth but much weaker cloud cover changes. So
  increased CO2 results in shallower MBL but with
  lower cloud cover.
• Could this be explained by methodology for
  compositing (i.e. using percentiles)?

26
SWCF vs LTS and EIS

• LTS EIS

27
SWCF vs LTS and EIS (with CRM and LES runs)
28
Comments on previous

• LTS better predictor of current climate model
  cloud cover than EIS
• Increases in LTS are large in all models (?1K).
  Corresponding increases in low cloud cover are
  roughly consistent with KH93 for CAM and SP-CAM,
  but clouds decrease in AM3 despite large LTS
  increases
• Increases in EIS are seen in all models but these
  are much weaker than LTS increases (order
  0.3-0.5K). CAM3 and SP-CAM cloud cover increases
  more strongly than expected from low cloud-EIS
  relationship of Wood and Bretherton.
• Similar conclusions drawn for SWCF vs LTS/EIS