Transcript Presentation - Copernicus.org

Using climate model experiments
to explore difference between
degrees of global warming
Lessons from a study of African precipitation
Rachel James and Richard Washington
Climate Research Lab, Oxford University Centre for the Environment
[email protected]
Changes in seasonal temperature
and precipitation at 1°C, 1.5°C,
2°C, 3°C, 4°C, 5°C, 6°C ΔTg
2009  present
Data from CMIP3, CMIP5, 2 PPEs,
RCMs  almost 400 model runs
What do the models show?
To what extent do the model
experiments provide a
trustworthy evidence base?
Changes in seasonal temperature
and precipitation at 1°C, 1.5°C,
2°C, 3°C, 4°C, 5°C, 6°C ΔTg
2009  present
Data from CMIP3, CMIP5, 2 PPEs,
RCMs  almost 400 model runs
What do the models show?
To what extent do the model
experiments provide a
trustworthy evidence base?
Political rationale
• “dangerous climate change”
• 2˚C as a benchmark for danger
• “1.5˚C to stay alive”
Political rationale
• “dangerous climate change”
• 2˚C as a benchmark for danger
• “1.5˚C to stay alive”
Now: “The 2013-2015 Review”
Political rationale
• “dangerous climate change”
• 2˚C as a benchmark for danger
• “1.5˚C to stay alive”
Now: “The 2013-2015 Review”
Political rationale
• “dangerous climate change”
• 2˚C as a benchmark for danger
• “1.5˚C to stay alive”
Key question: What is the difference between 1.5°C and 2°C?
Political rationale
• “dangerous climate change”
• 2˚C as a benchmark for danger
• “1.5˚C to stay alive”
Key question: What is the difference between 1.5°C and 2°C?
And 2.5°C, 3°C, 3.5°C etc.
“Time-slices” e.g.2080-2100
What’s the difference between 1.5°C and 2°C?
And 2.5°C, 3°C, 3.5°C etc.
So how to provide more information about 1.5°C
relative to 2°C? And 2.5°C, 3°C, 3.5°C etc.
• Pattern scaling
• Mitigation scenarios
• Time of ΔTg intervals
So how to provide more information about 1.5°C
relative to 2°C? And 2.5°C, 3°C, 3.5°C etc.
• Pattern scaling
• Mitigation scenarios
• Time of ΔTg intervals
assumes linear change
So how to provide more information about 1.5°C
relative to 2°C? And 2.5°C, 3°C, 3.5°C etc.
• Pattern scaling
• Mitigation scenarios
• Time of ΔTg intervals
assumes linear change
not many available
(except RCP2.6)
So how to provide more information about 1.5°C
relative to 2°C? And 2.5°C, 3°C, 3.5°C etc.
• Pattern scaling
• Mitigation scenarios
• Time of ΔTg intervals
assumes linear change
not many available
(except RCP2.6)
Temperature Intervals ΔTg e.g. 2°C
1°C, 2°C, 3°C
Generates patterns of change for each model…..
e.g. CSEM1-BGC in RCP8.5
Precipitation anomalies
relative to 1985-99
Generates patterns of change for each model…..
e.g. FIO-ESM
Precipitation anomalies
relative to 1985-99
Generates patterns of change for each model…..
e.g. ACCESS 1.0
Precipitation anomalies
relative to 1985-99
Generates patterns of change for each model…..
e.g. BNU-ESM
Precipitation anomalies
relative to 1985-99
Generates patterns of change for each model…..
e.g. CanESM2
Precipitation anomalies
relative to 1985-99
Generates patterns of change for each model…..
e.g. CMCC-CESM
Precipitation anomalies
relative to 1985-99
Generates patterns of change for each model…..
e.g. GFDL-CM3
Precipitation anomalies
relative to 1985-99
Generates patterns of change for each model…..
e.g. bcc-csm1.1
Precipitation anomalies
relative to 1985-99
Generates patterns of change for each model…..
e.g. CCSM4
Precipitation anomalies
relative to 1985-99
Generates patterns of change for each model…..
e.g. CNRM-CM5
Precipitation anomalies
relative to 1985-99
Generates patterns of change for each model…..
e.g. CSIRO-Mk3.6.0
Precipitation anomalies
relative to 1985-99
Can we extract any information from so many
different model projections?
• Focus on consensus response
• Present range of futures
• Look for consistent messages
Can we extract any information from so many
different model projections?
• Focus on consensus response
• Present range of futures
• Look for consistent messages
CMIP3
Precipitation anomalies
relative to 1985-1999
19 models
SRES A1B
James and Washington (2013) Climatic Change
Can we extract any information from so many
different model projections?
• Focus on consensus response
• Present range of futures
• Look for consistent messages
Should we emphasise
consensus from an
“ensemble of
opportunity”?
Can we extract any information from so many
different model projections?
• Focus on consensus response
• Present range of futures
• Look for consistent messages
CMIP3
Precipitation change
relative to 1985-1999 (%)
19 models
SRES A1B (blue)
SRES A2 (purple)
James and Washington (2013) Climatic Change
Using PPEs:
examining
uncertainty
Precipitation change
relative to 1985-1999 (%)
CMIP3 A2 (purple)
CMIP5 RCP8.5 (blue)
HadCM3 PPE (red)
HadSM3 PPE (triangles)
James et al. (2014) Journal of Climate
local precipitation (%)
Using PPEs:
examining
uncertainty
Precipitation change
relative to 1985-1999 (%)
CMIP3 A2 (purple)
CMIP5 RCP8.5 (blue)
HadCM3 PPE (red)
HadSM3 PPE (triangles)
global temperature (°C)
James et al. (2014) Journal of Climate
local precipitation (%)
Using PPEs:
examining
uncertainty
Precipitation change
relative to 1985-1999 (%)
CMIP3 A2 (purple)
CMIP5 RCP8.5 (blue)
HadCM3 PPE (red)
HadSM3 PPE (triangles)
global temperature (°C)
Are these ranges useful for impacts assessment
and decision-making?
James et al. (2014) Journal of Climate
Can we extract any information from so many
different model projections?
• Focus on consensus response
• Present range of futures
• Look for consistent messages
Generates patterns of change for each model…..
e.g. FIO-ESM
Precipitation anomalies
relative to 1985-99
Generates patterns of change for each model…..
e.g. FIO-ESM
Main difference
between ΔTg
intervals is in the
magnitude and
spatial extent of
change
Precipitation anomalies
relative to 1985-99
e.g. HadGEM2-ES
Also true at higher
degrees of
warming
e.g. HadGEM2-ES
e.g. HadGEM2-ES
Message for policy?
No particular thresholds to be avoided on the basis of mean climate
e.g. HadGEM2-ES
Lack of thresholds/discontinuities?
e.g. HadGEM2-ES
Lack of thresholds/discontinuities?
Can models represent feedbacks associated with nonlinear responses?
Conclusions
• Impact of 1.5°C vs. 2°C is model dependent
• Can we extract any useful information?
• Range at each ΔTg ?
» Different for each ensemble
• Consistently progressive change with warming
» Can we trust this signal?
Thank you for your attention
Rachel James and Richard Washington
17 April 2015
[email protected]