Consistency and robustness of selected emergent constraints for equilibrium climate sensitivityAxel Lauer and Veronika Eyring

Deutsches Zentrum für Luft- und Raumfahrt e.V. (DLR), Oberpfaffenhofen, Germany

CFMIP Meeting on Cloud Processes, Circulation and Climate Sensitivity25-28 September 2017, Tokyo, Japan

Equilibrium climate sensitivity (ECS)

Charneyreport

IPCC FAR

IPCC SAR

IPCC TAR

IPCC AR4

IPCC AR5

DefinitionProjectedequilibriumchangeinglobalannualmeansurfacetemperaturefollowingadoublingoftheatmosphericCO2concentration.0.0

1.0

2.0

3.0

4.0

5.0

6.0

1979 1990 1996 2001 2007 2013

ECS(K)

Emergent constraints

Emergentconstraints arearelationshipacrossanensembleofmodels,betweensomeaspectofEarthsystemsensitivityand anobservabletrendorvariationinthecurrentclimate

• Anecessarypropertyofemergentconstraintsisaphysicalbasisfortherelation.

• Offerthepotentialtoreduceuncertaintyinclimatefeedbacksandprojections.

• Canhelpguidingmodeldevelopmentontoprocessescrucialtothemagnitudeandspreadoffutureclimatechange andtoguidefutureobservations.

Selected published emergent constraints for equilibrium climate sensitivity

SelectedpublishedemergentconstraintsforECS

1) Precipitation – SouthernITCZindex(Tian,2015)

2) Humidity – Tropicalmid-tropospherichumidityasymmetryindex(Tian,2015)

3) Mixing – Lowertroposphericmixingindex(LTMI)(Sherwoodetal.,2014)

4) Clouds – Covarianceofdeseasonalized tropicallowcloud(TLC)shortwavereflectionwithSST(Brient andSchneider,2016)

Testconsistencyamongdifferentproposedemergentconstraintsandtheirsensitivityto:

• Modelensemble(CMIP3,CMIP5)• Observationaldataset(s)

ESMValTool version 1.1.0

http://www.esmvaltool.org/Eyringetal.,Geosci.ModelDev.,2016

• Communitydiagnosticsandperformancemetricstool fortheevaluationofEarthSystemModels(ESMs)

• Standardizedmodelevaluation canbeperformedagainstobservations,againstothermodelsortocomparedifferentversionsofthesamemodel

• manydiagnosticsandperformancemetricscoveringdifferentaspectsoftheEarthSystem(dynamics,radiation,clouds,carboncycle,chemistry,aerosol,sea-ice,etc.)andtheirinteractions

• Well-establishedanalysisbasedonpeer-reviewedliterature

• Currently≈70developersfrom28institutionsand>100users

GitHub:https://github.com/ESMValGroup/ESMValTool

Southern ITCZ index (Tian, 2015)

SoutheasternPacific30°S-0°,150°W-100°W

• Climatologicalannualmeanprecipitationbias

• Model– observation(mmday-1)• AveragedoversoutheasternPacific

Observational data

1) GlobalPrecipitationClimatologyProject(GPCP,1980-2005)2) CPCMergedAnalysisofPrecipitation(CMAP,1980-2005)3) TropicalRainfallMeasuringMission(TRMM,1998-2013)

Southern ITCZ index (Tian, 2015)

4.0 K

TRMM, CMIP5

Southern ITCZ index (Tian, 2015)TRMM GPCP CMAP

CM

IP3

CM

IP5

4.0 K 3.8 K 3.7 K

4.1 K 3.8 K 3.7 K

Southern ITCZ index (Tian, 2015)

4.1 K

3.7 K

Tropical mid-tropospheric humidity asymmetry index (Tian, 2015)

Observationallybaseddata

1) AtmosphericInfraredSounder(AIRS,2003-2010)

2) ERA-InterimReanalysis(1980-2005)

3) NCEP/NCARReanalysis1(1980-2005)

• Climatologicalannualmeanmid-troposphericspecifichumidityrelativebias• (model– observation)/observation(in%)• SHtropicalPacific(120°E-80°W,30°S-0°)– NHtropicalPacific(120°E-80°W,0°-

20°N)

NH tropical PacificSH tropical Pacific

Tropical Pacific

3.8 K

3.3 K

Tropical mid-tropospheric humidity asymmetry index (Tian, 2015)

Lower tropospheric mixing index (LTMI) (Sherwood et al., 2014)

Observationallybaseddata

1) ERA-InterimReanalysis(1980-2005)2) NCEP/NCARReanalysis1(1980-2005)

• Averagedovertropicaloceanregionofmeanascent(upper25%)• Small-scalemixingS=(DR700-850/100%- DT700-850/9K)/2• Large-scalecomponentofmixing=ratioofshallowtodeepoverturning

D=áDH(D)H(-w1)ñ /á-w2H(-w2)ñ• LowertroposphericmixingindexLTMI=S+D

Tropical region of mean ascent (ERA-Interim)

Lower tropospheric mixing index (LTMI) (Sherwood et al., 2014)

3.6 K

2.9 K

NCEP

ERA-Interim

Covariance of tropical low cloud (TLC) shortwave reflection with SST (Brient and Schneider, 2016)

Observationallybaseddata

1) Relativehumidity:ERA-InterimReanalysis,NCEP/NCARReanalysis1

2) TOAradiativefluxes:CERES,SRB,ISCCP-FD

3) SST:HadISST

• MeanoverTLCregions:25%oftropicaloceanarea(30°S-30°N)withlowest500hPa relativehumidity

• TLCreflection(atTOA)=- áScñ /áIñ (in%)• Regressioncoefficients:deseasonalized variationsofTLCshortwave

reflectionandseasurfacetemperature(in%perK)

Annual mean TLC regions (ERA-Interim)

Covariance of tropical low cloud (TLC) shortwave reflection with SST (Brient and Schneider, 2016)

3.7 K

3.3 K

ERAISCCP

HadISST

NCEPSRBHadISST

CMIP5

CMIP3

Southern Hemisphere Hadley cell extent (Lipat et al., 2017)

~3.0 K

ERANCEP

CMIP5

CMIP3

Conclusions and outlook• Allfourofthetestedemergentconstraintsare

sensitivetothemodelensembleand/ortheobservationaldatasets used

• EstimatedECSrangeintheoriginalfourstudies:3.8– 4.0K(2.4– 4.3K)

• Robustnessestimatesobtainedhere(additionalobservationsandensembles):2.9K– 4.1K

• Applyingcompilationofcurrentemergentconstraintstudies(forECS)donotallowtonarrowrangeofECSsignificantly

Outlook• Furtherselectionbasedonphysicalmechanisms andfundamentalprocesses• Focusonindividualfeedbacks toconstrainECS/TCR/TCRE(e.g.carboncycle)• Accountingforobservationaluncertainties• Investigationofadditionalpublishedemergentconstraints• Developmentofnewemergentconstraints

Emergentconstraint Originalrange Thisstudy

ITCZ bias 4.1 K 2.7– 4.1K

Humidityasymmetryindex

3.8K 3.3– 3.8K

LTMI 4.0K(3.0– ) 2.9– 3.6 K

TLCreflection 4.0K(2.4– 4.3K) 3.3– 3.7K