Modélisation climatique globaleModélisation climatique globaleet observations des nuages et du rayonnement :et observations des nuages et du rayonnement :

Quelles interactions ?Quelles interactions ?

Sandrine Bony, LMD/IPSL, Paris

● Quels enjeux dans les prochaines années ?

● Processus nuageux & études climatiques

● Implications pour les observations

Modélisation climatique globale et observations des nuages et du rayonnement :Modélisation climatique globale et observations des nuages et du rayonnement :Quelles interactions ?Quelles interactions ?

Sandrine Bony, LMD/IPSL, Paris

● Quels enjeux dans les prochaines années ?

● Processus nuageux & études climatiques

● Implications pour les observations

A changing paradigm for climate change modelingsince the IPCC AR4

● Further global warming inevitable overthe next few decades→ from “alarm” to “action” ● Need to inform decisions about climateadaptation and mitigation→ focus on two different time scales (near-term, long-term)→ importance of regional climate changes and extreme events

● The need to improve the reliability of climatemodels and to assess the robustness of climateprojections has never been so high

Projections ofanthropogenic climate change :

●Promotes a standard set of model simulations in order to :➢ evaluate how realistic the models are in simulating the recent past➢ provide projections of future climate change on two time scales➢ understand some of the factors responsible for model differences

●Two timescales and two sets of science problems

●Will be assessed by the IPCC AR5

Near-Term :(next 3-4 decades)

→ decadal climate predictability

→ regional climate changes (high resol)

→ climate extremes

→ air quality changes (aerosols, chemistry)

Long-Term :(past to 2100 & beyond)

→ evaluation of climate models(recent past, A-Train, paleo)

→ climate sensitivity andphysical feedbacks (e.g. clouds)

→ biogeochemical feedbacks(e.g. carbon)

→ ice sheets and sea level

CMIP5 : a framework for climate change modeling over the next 5+ years

Clouds& Precipitation

Earth's energy balance& Hydrological Cycle

GeneralCirculation

Processus nuageux :Une composante clé de la modélisation du climat

Rôle critique de ces interactions dans :● Sensibilité climatique (rétroactions)● Réponse de la précipitation en changement climatique● Evénements extrêmes● Modes de variabilité (MJO, ENSO..)● Structure de l'ITCZ● Biais du climat moyen (affecte la prévisibilité)

Climate change cloud feedbacks & climate sensitivity

:

(Dufresne & Bony 2008)

multi-modelmean T :

inter-modeldifferences : cloud

feedbacks

CRF SW

SST

Low-sensitivityOAGCMs

(Bony & Dufresne 2005 )

High-sensitivityOAGCMs

boundary-layerclouds

IPCC AR4, Bony et al. (2006)'s review :1. Inter-model differences in cloud radiative feedbacks constitute the primary source of uncertainty in climate sensitivity estimates.2. The response of marine boundary-layer clouds is the primary contributorto inter-model differences in global cloud feedbacks

Tropospheric cloud-radiative effects and the large-scale atmospheric circulation

Cloud-radiative effects strengthenthe Hadley-Walker circulation,make the ITCZ more narrow,

and affect its large-scale structure

Precipitation with/without ACRF

1st order effectin the simulation

of climate

Rôle des processus nuageux dans la variabilité tropicale (e.g. intra-saisonnière)

(Zurovac-Jevtic, Bony & Emanuel, 2006)

Simulations 2Daqua-planète

Importance des interactionsnuages - rayonnement et

vapeur d'eau – convectiondans l'organisation à grande échelle

et la variabilité de l'atmosphère tropicale

Cloud Feedback Model Inter-comparison Project Phase-2Cloud Feedback Model Inter-comparison Project Phase-2CFMIP-2 (http://www.cfmip.net)CFMIP-2 (http://www.cfmip.net)

Understanding Evaluation

GCM process& sensitivity studies

CRMs/LES/SCMsvia GCSS

A-Train/ISCCP & simulators

Assessment ofcloud-climate

feedbacks

+ Upcoming european project (FP7) : EUCLIPSE(EU CLoud Intercomparison, Process Study & Evaluation project)

Bridging models and observationsto better evaluate cloud and moist processes in models

At the large-scale :

CFMIP Satellite Simulator (COSP)to facilitate the comparison of model

outputs with satellite observations(CALIPSO, CloudSat, PARASOL, ISCCP, MISR)

↓Used in some CMIP5 experiments

COSP oriented satellite data

Link available from www.cfmip.net

Bridging models and observationsto better evaluate cloud and moist processes in models

At the large-scale :

CFMIP Satellite Simulator (COSP)to facilitate the comparison of model

outputs with satellite observations(CALIPSO, CloudSat, PARASOL, ISCCP, MISR)

↓Used in some CMIP5 experiments

At the process scale :

Detailed model outputs at selected locationswhere field experiments or

instrumented sites are available(e.g. ARM, VOCALS, AMMA)

↓Included in CMIP5 outputs

CFMIP/CMIP5 model outputs at selected locations(118 locations, high-frequency, detailed cloud diagnostics)

● ARM, CEOP, CloudNet instrumented sites● GPCI / Tropical West & South East Pacific / AMMA transects● Field experiments / GCSS case studies● Locations of large inter-model spread of cloud feedbacks (CMIP3)

GPCIAMMA

VOCALS

Oklahoma

Barrow

TOGA-COARE

ASTEX

GATE

SHEBA

SIRTAChibolton

Tibet

RICO

Darwin

Simulation of water stable isotopes (Oxygen 18, Deuterium)in LMDZ GCM

Observations LMDZiso

On-going model evaluation using isotopic data from :in -situ observations (GNIP) + satellite measurements (TES, SCIAMACHY, etc) + ice cores

Water isotopes: great tools to evaluate the model representation of :●past climate changes (temperature at high latitudes, precip at low latitudes)●troposphere-stratosphere exchanges●convective processes (e.g. rain reevaporation, precip efficiency)●land-surface processes (e.g. partition between evaporation and transpiration)

(Thèse de C. Risi, LMD/IPSL)

Vertical distribution of D simulated by the LMDZ4-iso GCM

p

max = 0.999

p

max = 0.99

tropical mean D

-650%o

-500%o

Besoin d'observations de la compositionisotopique de l'atmosphère pour contraindre :

- dans la haute troposphère : le transport convectif de vapeur d'eau et la microphysique de la précipitation

- dans la moyenne troposphère et dans la couche limite : les processus convectifs (e.g. downdrafts),la réévaporation de la pluie et les échanges sol-atmosphère (e.g. evapotranspiration)

(C. Risi & S. Bony, LMD)

Conclusion

● L'étude du rôle des processus nuageux dans le climat, et l'évaluation de ces processus dans les modèles sont primordiales pour la modélisation du climat à grande échelle.

● Une révolution est en marche avec l'arrivée des nouvelles observations (e.g. A-Train).

● Dans l'avenir, il importe :

- de maintenir la surveillance sur des échelles décennales d'un grand nombre de variables climatiques (e.g. bilan radiatif, nuages, vapeur d'eau)

- de faciliter l'accès des modélisateurs aux différents jeux de données : (grillage, format, programmes de lecture, documentation, assessments, etc)

- d'interagir autour des besoins spécifiques des modélisateurs (e.g. simulateurs d'observations, besoin d'observations particulières)