What Good are Climate Models

1
What Good are Climate Models?
2
(No Transcript)
3
(No Transcript)
4
(No Transcript)
5
Models are not perfect...
... but are they useful?
6
What are the models good for?

• Conceptual frameworks
• Radiative forcing concept
• Quantification of theories of climate change
• Orbital variations, lake drainage events, GHGs,
  volcanoes, solar etc.
• Hypothesis generation
• Thermohaline shutdowns, ENSO teleconnections
• Hindcasting
• 20th Century climate, LIA, LGM
• Prediction
• Pinatubo, 21st Century

7
Solar
Aerosols
Greenhouse Gases
Land Use
Total Forcing 1.6 1.0 W/m2
8
Mid-Holocene (6kyr BP)?
More insolation in NH summer, less tropical
insolation Can this explain the 'Green' Sahara
apparent Northward shift of ITCZ/Asian Monsoon?
9
Mid-Holocene (6kyr BP)?
Fixed SST
Change in precipitation increases as complexity
of ocean component does... Also as vegetation
feedbacks are included
Qflux
Fully coupled
10
Response to Volcanic eruptions...
11
The 8.2 kyr event...
12
Multi-proxy comparisons...
Legrande et al (2006)?
d18Oice
Methane
10Be
Dust
d18Olake
d18Ocalcite
Lake
Ocean
Ice Core Data
Use ensemble mean to drive CH4 wetland emissions,
dust, 10Be (aerosol) concentrations at Summit
(Includes snow acc. decrease (15-20). Methane
change assuming no change in N-S gradient, but
with feedback on its own lifetime
13
(No Transcript)
14
20th Century climate hindcasts
Matches to observed data imply consequences that
can be looked for in the real world...
15
(No Transcript)
16
Observations 1993-2003 0.6 W/m2
Levitus et al, 2009
Model
Hansen et al, 2005
17
Response to ozone depletion and GHGs
Shindell and Schmidt (2004)?
18
Annular modes respond to GHG, O3,
volcanoes... IPCC AR4 models
Ozone forcing
No ozone
Miller et al (2007)?
19
Even when they're bad, they're good
Relationship between surface and atmospheric
temperature anomalies is robust to amplitude to
tropical variability.
GISS-ER
GISS-EH
Increasing surface variability
Santer et al 2005
20
3 Scenarios A - exponential growth,
B - business as usual 'most plausible'
C - no further GHG growth
after 2000
SAT Trends 1984-2008
OBS 0.220.07/0.190.05 (different
indices)? Scen. B
0.250.05
21
Good for forcing estimates... ...and
temperature response... ...but
underestimated sea level rise.
Rahmstorf et al (2007)?
22
(No Transcript)
23
(No Transcript)
24
(No Transcript)
25
Getting your projected ducks in a row
Does the result make theoretical sense?
Does the result appear in multiple models?
Is there some support for the result in the
observations?
Do the predicted and observed magnitudes match?
Then it might be robust...
26
2005, 1992 surface melt
27
(No Transcript)
28
To conclude...

• Models actually do a reasonable job
• Many questions have been sharpened (if not
  answered completely)
• Quantitative explanations of many aspects of
  past climate variability
• Model feedbacks/interactions increasing
• Ocean atmosphere aerosols
  chemistry biology vegetation
• More interesting validation is now possible
• i.e. direct modelling of paleo-proxies, satellite
  irradiances...

29
What can you (mostly) trust?

• Model results that are robust across models,
  with clear physical mechanisms and observational
  support
• Global warming, water vapour feedbacks, impacts
  of NADW slowdown etc.
• Big picture changes (continental scale and up)?
• NOT single grid point changes
• NOT small net effects from very complex
  balances between competing process
• NOT impacts on sub-gridscale features
• hurricanes, ocean convection, slope flow

30
Prediction is difficult, particularly of the
future.
-Niels Bohr Grid models can't save us now!

-Day after Tomorrow
http//www.realclimate.org
Available in all good bookstores!