Climate modeling and uncertainties in climate projections. P. Braconnot IPSL/Laboratoire des Sciences du climat et de l environnement, France

Climate modeling and uncertainties in

climate projections

P. Braconnot

IPSL/Laboratoire des Sciences du climat et de l’environnement, France

The Intergovernmental Panel on

Climate Change (IPCC)

 The leading international body for the assessment of climatechange

 Established in 1988 by the United Nations Environment Programme

(UNEP) and the World Meteorological Organization (WMO)

 3 groups : The physical scientific basis (WG1) ; Climate Change impacts, 

adaptation and vulnerability (WG2) ; Mitigation of climate change (WG3)

 Provide the world with a clear scientific view on the current state of

knowledge in climate change and its potential environmental and socio-economic impacts.

4 reports : 1990, 1995, 2001, 2007 ‐‐‐ + AR5 in 2014 Special reports (ex/ report on extremes)

 Is a scientific body. It reviews and assesses the most recent scientific,

technical and socio-economic information produced worldwide relevant

to the understanding of climate change. It does not conduct any

WGI AR5 Timetable for 2013

WG I: The Physical Science Basis  23‐26 September 2013

WG II: Impacts, Adaptation and Vulnerability  25‐29 March 2014 WG III: Mitigation of Climate Change  7‐11 and 13 April 2014

WGI Contribution to the IPCC Fifth Assessment Report

Climate Change 2013: The Physical Science Basis

Summary for Policy Makers Technical Summary

WG1 AR5

• Chapter 1: Introduction

• Chapter 2: Observations: Atmosphere and Surface

• Chapter 3: Observations: Ocean

• Chapter 4: Observations: Cryosphere

• Chapter 5: Information from Paleoclimate Archives

 Chapter 6: Carbon and Other Biogeochemical Cycles

 Chapter 7: Clouds and Aerosol

• Chapter 8: Anthropogenic and Natural Radiative Forcing

• Chapter 9: Evaluation of Climate Models

• Chapter 10: Detection and Attribution of Climate Change: from Global to

Regional

 Chapter 11: Near-term Climate Change: Projections and Predictability

• Chapter 12: Long-term Climate Change: Projections, Commitments and

Irreversibility

 Chapter 13 : sea level

o Chapter 14: Climate Phenomena and their Relevance for Future Regional

Pressure for the modeling groups

Ensembles

Complexité

3 axes  for model evolution

 Various space and time scales (from global to local, from a few years to 

decades or centuries)

 Need to tell about possible evolutions of the caracteristics of the 

meteorology, climate variability et.. (hours to decades) 

 Need to tell how the coupling between climate and biochemical cycles 

affect both climate or the biochemical cycles (greenhouse gases, 

aerosols, land use, carbone cycle … )

• Coodinated experiments “MIPs”  • Provision of model results to a 

large community 

• Linkages between model quality 

and uncertainties in future 

climate projections?  • Progress since the AR4

Internal variabiliy

Volcanic activity

Solar activity

Anthropogenic factors

What drives climate variations and changes ?

Température moyenne Activité volcanique Activité solaire   Facteurs anthropiques ENSO        T (°C) T (°C) T (°C) T (°C) T (°C)

The role of climate modeling

• Different objectives for climate simulations :

–

Understand

mechanisms

of

climate

at

different

time

scales

–

Identify

how

climate

varies

depending

of

external

perturbations

–

Test

hypotheses

(ex/

response

to

different

anthropogenic

perturbation,

analyses

of

the

role

of

a

phenomenon

or

of

a

particular

feedback)

• Need to better assess uncertainties

–

Credibility

of

model

used

–

Complex

interactions

between

climate

and

environement

A 3D representation of the atmosphere, ocean, sea‐ice, land surface 

and of their coupling through momentum, heat and water fluxes. 

A  representation of the coupling with the biochemical cycles in the 

atmosphere ocean and land. (ex carbon cycle) 

From a film presenting climate modeling Copyright CEA

Climate or Earth System models

(General circulation model GCM )

Emission scenarios Ex: CO2 Ex : Annual Mean global temperature

Imagine 2100 : ex IPCC (2007)

From hypotheses on consequences of 

human activity on atmospheric 

composition, land surface characteristics

Compute climate caracteristics  : mean 

climate, seasonality, interannual 

Simulated climate (2100 – present)

temperature (°C) : 

Number of days with heat wave (A2)

augmen

te

But also

Frozen days (A2)

diminue

Etc ….

CMIP5 : an international project

Research/ societal needs

Climate simulation  in semi‐operational modes for modeling groups 

Constraints on models, computers,  storage, …. 

Data management (service) 

Format, quality etc.. , 

Scientific projects

Analyses, understanding, 

evaluation, estimation of 

uncertainties

Simulations at regional scale or 

climate downscaling

Link with other communities : 

Impact studies,  mitigation, 

adaptation

Calendar imposed by IPCC (availability of model results, publications in 

New objectives for CMIP5/ CMIP3

considered in IPCC(2007)



Need to better understand the interactions

between climate and carbon cycle



Assessment of regional information



Near term projection for the next 20-40 years :

toward decadal prediction



Model evaluation and analyses + understanding

CMIP 5 long term experiments

Paleoclimate Modeling

Intercomparison Project (PMIP) 

Why paleo now in CMIP5

(6ky BP, 21 kyBP, Last millennium )

• Out of sample test of our understanding of climate changes

• Estimation of climate sensitivity and possibilities to assess how it is represented in models

• Quantification and evaluation of various climate feedbacks (ocean, vegetation, ice…)

• Evidence for large changes in the hydrological cycle

Ex: 6000 yrs BP the 

climate was more 

“RCP” (representative 

concentration Pathway) 

indicate the intensity of 

the radiative

perturbation imposed 

to the climate system 

via:

‐ Anthropogenic 

emission of greenhouse 

gases and aerosols. 

‐ Land use

New

references

to

investigate

socio

‐

economic

scnenarios

and

human

induced

climate

change

:

-64 -60 -56 -52 Temperat ure DC (°C) 800x103 600 400 200 0 -200 Age (ky BP) 160 120 80 40 0 _{Sea level (m)} 60 40 20 0 -20 NH June insol (W /m²) 360 320 280 240 200 C O 2 ( p pmv ) 800 600 400 CH 4 (pp bv)

Temps (milliers d’années)

actuel

*

*

Niveau des mers

Ensoleillement ppbv ppmv °C m 65°N Été (W/m²)

EPICA, 2004; Loulergue et al, 2008; Jouzel et al, 2007; Luethi et al, 2008

(°C)

How

big is the

perturbation

?:

LGM

/

RCP8.5

Last Glacial Maximum to present Present to 2100 with RCP8.5 CNRM-CERFACS IPSL

Hawskins and Sutton, BAMS, 2009 scenario model internal internal scenario model

Sources of uncertainties : noise vs trends

Global approach  For a particular region (British isles) 

WCRP/WGCM and IGBP/AIMES http://c4mip.lsce.ipsl.fr/ Understand  carbone‐climate coupling Estimate and evaluate the magnitude of  climate‐carbone feedback  Evaluate  how well model reproduce climate  and carbone cycle on the historical period. 

IPCC AR4, WG1 Fig. 10.20

2.6   –4 .1   °C 2.4   –5 .6   °C 830   ppm 730   – 1000   ppm Modèles climat‐carbone C4MIP

Modèles climats AR4

Modèles climat‐carbone C4MIP

Modèles climats AR4

Large incertitude sur le  cycle du carbone

Rétroaction positive Plus large réchauffement

Uncertainties on socio-economic scenarios

Uncertainties from climate models

Accord signe < 66% _{Accord signe > 90%}

Unconsidered feedbacks :

Melting of the ice sheet and coupling with ocean circulation

Permafrost and methane

Complexity :

Climate-carbone feedback

Coupled

uncoupled

Do not forget regional aspects

Ex aerosols

Ex land use Szopa et al. 2012

Résultats multi modèles

How do we communicate that

• Spread of projections in CMIP5 AOGCMs comparable to CMIP3, and first generation. ESMs produce comparable first order results to AOGCMs

• CMIP5 offers the opportunity :

– to study climate change with many additional capabilities (carbon and 

chemistry, short‐term climate change, comparison paleo/future, 

forcings and feedbacks diagnostics, high‐resolution, high‐frequency

outputs, etc)

– to better understand the spread and better assess the robustness of 

model results ; great value of idealized CMIP5 experiments.

• Decadal prediction : challenging...

• RCPs may not sample the range of plausible pathways regarding aerosols and land-use.

• Model biases :

– some quantities show improvement (e.g. rate of sea ice loss in Arctic, 

variability for some phenomenon : ex ENSO)

– many others have not significantly improved (e.g. double ITCZ, Arctic 

Traduction for other communities and

socio-economic actors

Climate modeling to : Understand

Test effect of socio-economical choices

Link meteorology, climate and

environment (heat waves, droughts, storms) Inform Climate services Decision  Adaptation Mitigation  

Characterize and estimate uncertainties

Need to deliver a message  that is : 

Credible

Understandable Actionable 

Need to understand user  needs to provide the right 

level of information

Each case is specific and 

requires a specific treatment 

of uncertainties 

Déandréis et al. Submitted

European project IS‐ENES : analysis of 17 use cases to explain the 

workflow from a climate model output to the entry of an impact 

Conclusion

• IPCC does not make science but triggers it and imposes a calendar. • CMIP5 simulations provide an fantastic amount of data that allow to

analyze lots of aspects of climate change.

• Paleoclimate offers opportunities to better assess model performances, climate sensitivity and climate variability

• Human activity affect Earth’s energetic and lead to a long term

climate warming, but the its magnitude depends on the magnitude of the radiative perturbation.

• Climate projections offers lots of possibilities to imagine the future climate and how climate change can affect the environment and society. There is room for new jobs as translators between science and applications.

• Need to make progress in the discussion of uncertainties. I should not be a barrier to action or adaptation. Lots of results are coherent and sufficient to guide efficient policies. This requires being able to formulate the right questions

It is difficult to communicate

• Confusion (+ communication noise on fundamental science bases by climate skeptics)

– Forecast/projection

– complexity/ chao

– Tendency / variability

– Short term/ long term

– Use of data (model initial state, « forcing », parameters, evaluation) 

• Model tuning (particular to climate / meteo)

– Energetic equilibrium

– Initial state / response to a radiative perturbation 

• Evaluation

– Need a basket of metrics, test cases, methods from mean climate to 

variability and extremes (in general for specialists… ) 

– Dynamical and physical content/ results and simulated climate 

• Multidisciplinary aspects

– Physics / biochemical cycles 

Paleoclimate in CMIP5

Paleoclimate Modeling

Intercomparison Project (PMIP3) http://pmip.lsce.ipsl.fr/

– Understand mechanisms of past climate change

– Evaluate roles of feedbacks from the different climate subsystems (atmosphere,

ocean, land-surface, sea-ice …) – Evaluate the ability of climate

models to simulate a climate different from that of today

Taylor et al, BAMS 2012 Clivar Newsletter, 2012 Braconnot et al, 2012

Anomalie du cumul de précipitations : écart entre le scénario et la période de référence [mm], Scénario d'évolution socio-économique intermédiaire (A1B)

Moyenne annuelle

Expériences/Modèles Référence (années 1970) Horizon moyen (années 2055) Horizon lointain (années 2085)

SCRATCHOB CERFACS-France CNRM ~ "'""" ~ 1 ... . ..-SCRATCHOB ..,.,_ CERFACS - GIEC CMIP3 Allemagne ... ~

/~

CMIP3 France IPSL

..

Perturbation anthropique? = combien de W/m

en plus ou en

moins

Estimation depuis le début de l’ère industrielle (1860)

Is it possible that these small perurbations

affect the recent increase in temperatures?

 Can we detect in the data a signal that

is different from the natural

fluctuations arising from interactions between the different elements of the climat system?

• requires observations, model simulations , statistical methods

Simulations Natural & Anthropogenic forcings

Simulations

Natural forcings observations

Simulations Natural & Anthropogenic forcingsSimulations Natural & Anthropogenic forcings

Simulations

Natural forcingsSimulations observations Natural forcings observations

 If yes, can we attribute these fluctuations to an external cause from natural

(volcanism, solar …) or antropogenic (greenhous gazes, aerosols, land use …)?

Amplitude / variations passées ?

 Ex: dernier maximum glaciaire (-21000 ans), calottes de glaces HN

et taux de CO2 plus faibles (200 ppm) : refroidissement 4-5°C pour

≈ -3W/m2 _{par rapport à l’actuel}

Fig. 6.5 (IPCC 2007)

Niveau de service Idées Organisation 

générale des services 

climatiques Partage et diffusion  de retours  d’expériences Mise en place  d’un cadre  réglementaire Fourniture  d’informations  climatiques

Création d’un portail 

unique de diffusion  des données Partage de  méthodologies Elargissement à  de nouveaux  fournisseurs de  données Diffusion  d’études de  cas et  d’exemples  d’études Efforts de  communicati on sur les  incertitudes FAQ,  Hotline,  Web  forum

Etudes au cas par cas

Identification claire 

des différents acteurs 

et du rôle de la  science Mise en œuvre et  suivi de l’adaptation Contrôle et validation  des mesures  envisagées Formation,  communication et  sensibilisation Diffusion de messages  synthétiques de 

qualité sur le climat

Création de  formations  académiques et  professionnelles Développemen t d’une culture  commune Workshops,  colloques,  conférences