ACE Report3 Gettelman Climate2

8/7/2019 ACE Report3 Gettelman Climate2

1/24

KeyUncertain,esinClimateSimula,on:

CloudsandAerosols

A.Ge&elman,NCAR

NCAR is supported by the U.S.

National Science Foundation


2/24

Beijing, 4th Ring Road, July 27th 2010, 6p BST

Keyknownunknownsinpredic,ngclimate Feedbacks:theroleofclouds Forcings:theroleofaerosols Currentprogressandlimita,ons

Wherewecanmakeprogress


3/24

Beijing, 4th Ring Road, July 27th 2010, 6p BST


4/24

Whatwillfutureclimatebe?

Usemodelstosimulate:ourbestguessisuncertain

IPCC2007PolicymakersSummaryFigure5

Notes:

Our current trajectory

looks like A2 or A1FI !

Spread within each

scenario is large. Evenwhen we understand the

carbon cycle (carbon

scenario), uncertaintiesare large. Why?


5/24

Predic,ngfutureclimate

Currentstateoftheartmodels Approachingglobalweatherresolvingmodels(50-25km) Detailedsatellitesimulatorsforevalua,on

Reduceretrievaluncertainty

Includeoceanlandseaice

Wherewearegoing: GlobalcloudpermiQng(5-15km)orcloudresolving(1kmorless)

modelsforclimate(>10yrs)

Hybrid:high-resolu,onparameteriza,onsofclouddynamicsembeddedinGeneralCircula,onmodels

Willwantobserva,onsatthesescales Includeicesheetandcarboncyclemodels


6/24

Radia,veForcing



7/24

Radia,veForcingUncertainty

AnthropogenicforcingF=~3Wm-2AerosolEffect(AE) IfAElargethennetforcing(F)issmallforfixedDT(1oC) Aerosoleffectsalterobservedclimatesensi,vity(g)

(F=DT/g)

g=DT/FforfixedDTAEreducesFgislarger


8/24

Feedbacks

2xCO2=+4W/m2Forcing.notthewholestory!

Halfofexpectedwarmingisfromfeedbacks Example:Watervaporfeedback:+T+H

2

O

sinceH2Oisagreenhousegas+H2O+F

necessarytokeepearthhabitable ClimateFeedbacksdetermineclimate

sensi,vity(andDTifFisknown)


9/24

DifferentClimateFeedbacks

CAM4

CAM5

Soden, 2008 (also Colman, Bony)

H2O and T(z)

Clouds

The Water Vapor feedback is large, positive and has small spreadThe sign of cloud feedback is uncertainSpread in cloud feedbacks as large as the Water Vapor feedback (1.5Wm-2K-1)


10/24

CloudFeedbacksaffectSensi,vity

CannonicalStephenspictureorBonyModels with - cloud feedback are less sensitive than those with + cloud feedback

Stephens, 2005, J. Climate

Low cloud area decreases result in

less cooling (+ feedback) and

higher sensitivity


11/24

Shortwave(Solar)CloudFeedbacks

LW

CAM4: Low sensitivity (3oC) CAM5: High sensitivity (4oC)

Clouds are the major difference in climate sensitivity in these models !

Trace feedbacks to regions:Equatorward flanks of storm tracks over the oceansSignificant impacts of cloud processes in the ArcticTrace feedbacks to processes:

The response of shallow convective clouds to warming


12/24

Summary:Feedbacks

CloudFeedbacksarethebiggestuncertainty Needtoevaluateresponsesofcloudregimes

Needsta,s,cstorepresentprocesses. Fastphysics:dontneed50yearsforprogress Globalobserva,onstohelpparameterizeclouds

Resolu,ons:modelsgoingto5-25km S,llneedparameteriza,ons Needver,calinforma,ononcloudmicrophysics

especiallyradia,velyimportantquan,,es(par,clesizes)andliquid/icecontentsprecip

Mul,-parameterviewcri,cal:interac,onswithenvironment


13/24

Radia,veForcing:Aerosols



14/24

Aerosol-CloudInterac,ons

Directeffects:Cool IndirectEffects:Cool Indirecteffectsonprecipita,on

Constrained Known Uncertain Impact

Radiative Forcing

Precipitation impact


15/24

Aerosol-CloudInterac,ons(2)

Directeffects:aerosolsscaer&absorbradia,on

AerosolIndirectEffects(AIE):

AerosolsactasCloudCondensa,onNuclei(CCN)

SeaSaltSulfateDust

AerosolsmayalsobeIceNuclei(IN) DustSulfateSoot?

MoreCCN Moresmallerdrops&brighterclouds(AlbedoEffect) Smallerdropsmayseleslowerwithlongerlife,me&less

precipita,on(Life,meEffect)

Modelstypicallyshowlargerimpactsthanobserved


16/24

CloudNucleiv.AerosolAbsorp,on

AlbedoEffect:RosenfeldetalScience2008


17/24

WhyAIEMaer:Precipita,on

Aerosolsdelayprecipita,on:

butmaymakeitmoreintense(alsodependsonice)

Rosen

feldetal,Science

2008


18/24

Evalua,onofAIE

LocalEvalua,on(v.fieldobserva,ons) Regional/GlobalEvalua,on(v.Satellites) Difficul,es:

Correla,onisnotcausa,onCo-variancewithmeteorologicalstateSystemmaybeheavilybufferedwithcompe,ng

effects(e.g.:precipita,oneffects)


19/24

Currentobserva,onsprovide

NecessarybutnotSufficientTests

Aerosol:(e.g.AOD) Microphysics:

(DropNumber)

Co-Variability:(AODv.DropNum)

MODIS July AOD CAM5 July AOD

From: X. Liu, PNNL


20/24

But:Correla,onisnotCausa,on1.Ver,calStructure:CALIPSOLidarJuly2008:

W.AfricanCoast&S.Atlan,c:

AerosolridesOVERStratocumulus

2.Correla,onstoa3rdvariable:

E.g.:AOD&Cloudscorrelatedbutboth

correlatedwithhumidityandwindspeed

Nigeria South Atlantic

Aerosol LayerStratocumulus

Convective & Cirrus Clouds


21/24


22/24

Whereobserva,onsarenow

A-TrainSynergy:

MODIS(columnmaps)CloudSat(cloudver,calstructure)CALIPSO(aerosolver,cal

structure)AIRS/MLS(humidity)

TRMM(precip) Problems

Limitedver,calinforma,ononaerosolscloudsPrecip(TRMM)notlinkedtocloudmicrophysics


23/24

AerosolSummary

Aerosolsarebiggestuncertaintyinnarrowingcurrentforcing:helpswithunderstandingclimatesensi,vity!

Aerosolsmayperturbthehydrologiccycleandalterprecipita,onpaernsandintensity.

Wecannotobservetheseprocessessufficientlytoconstrainmodels

Cloudandaerosolmicrophysicalprocessesandprocessinterac,onsarethelargest

uncertainty


24/24

Overallsummary

Modeluncertain,esinclimatepredica,onarerelatedtocloudphysics(cloudfeedbacks)and

aerosolcloudinterac,ons(climateforcing)

Modelprocesses(especiallyaerosol-cloudinterac,ons)arepushingobserva,onallimits:

theycannotconstraineffectfurther

Needmul,-sensorplaormsforsimultaneousver,callyresolvedobserva,onsofclouds

aerosolsandprecipita,ontomakeprogress

Documents

ACE Report3 Gettelman Climate2