SpectralDecompositionofCloudRadiativeEffectandCloudRadiativeFeedbacks

Xiuhong Chen,XiangleiHuangTheUniversityofMichigan

QingYueCaltech/JPL

TheBillRossow Symposium:Clouds,TheirProperties,andTheirClimateFeedbacksJune6,2017,NewYork

Acknowledgements:NASATerra/Aqua&CloudSat Programs

Outline• Whygofrombroadbandtospectral?

– Twoexamples• TraitsinthespectraldecompositionofCRE• Band-by-bandLWCRE:CESMvs.obs• Band-by-bandLWcloudfeedbacks:2xCO2CESMsimulations

• Band-by-bandLWshort-termcloudradiativefeedbacks(fluctuations):modelvs.obs in2003-2013

• DiscussionandConclusions

Whatspectraldimensioncanoffer?Revealcompensatingdifferencesthatcannotberevealedinbroadbanddiagnosticsalone.

Spectraldecompositionofbroadbandlapse-ratefeedback(Huangetal.,2014,GRL)

LWBroadband H2Obands(0-540cm-1,>1400cm-1) windowregion(800-980cm-1)

òò

D

DDD

-=

v sv

vv svv

dvTB

TOAFdvTBg

)(

)()(

clear-skygreen-houseefficiencyAMIPrunsforcedbyobservedSST

Obs fromcollocatedAIRSandCERES(Huangetal.,2008;Chenetal.,2013)

(GEOS5simulationprovidedbyL.Oreopoulos etal;CanAM4providedbyJ.Cole)

A trait of spectral (band-by-band) CRE

CRELW =σTs4 −[ fσTc

4 + (1− f )σTs4 ]= f σTs

4 −σTc4"# $%

CRE(Δv) = f [Fclr (Δv)−Fcld (Δv)] Fractional contribution

r(Δv) = CRE(Δv)CRELW

=Fclr (Δv)−Fcld (Δv)

σTs4 −σTc

4"# $%

t>>1Cloudfraction:f

Fclr(Dv)

Surface

Fcld(Dv)

Tc

Ts

r(Dv) changes with Tc

288K250K220K

1. Blackbody cloud2. Ignore atmospheric absorption

Band-to-Band ratio: sensitive to CTH but not cloud amountLW CRE: sensitive to both CTH and cloud amountOutcome: ratio-then-broadband approach (Huang et al., 2014, J Climate)

Derivationofspectralfluxes/CRE/feedbacks

• Observations– InvertfromAIRSradiancesfollowingthescenetypeclassificationofCERES(Huangetal.,2008;Chenetal,2013;Huangetal.,2014)

– Outcome:spectralfluxat10cm-1intervalovertheentireLWspectrum(09/2002topresent)

– Observation-basedcloudradiativekernel(Yue etal.,2016)• Models

– Simplecodemodificationtooutputband-by-bandfluxes– Spectralradiativekernels(Huangetal.,2014,GRL)toderivespectraldetailsofPlanck/Lapse-rate/WVfeedbacks

– Adjustmentmethodstogetspectralcloudradiativefeedbacks

10-yearmeanspectralCREoverthedifferentclimatezones

37.9 Wm-2

27.2 Wm-2

(Huangetal.,2014,JClimate)

Observedaveragesof2003-2015

CAM5forcedwithobservedSSTfrom2003to2015(totalrun2000-2015)

DifferencesofModel- Obs

Observation Band-by-bandCRE(RRTMG_LWbandwidths)

500-630cm-1

820-980cm-1

Observation:2003-2015 CAM5forcedbyobservedSST2003-2015

(Wm-2)

(Wm-2)

500-630cm-1 500-630cm-1

820-980cm-1 820-980cm-1

CESMfully-coupledrun30-yearmean

Band-by-bandLWcloudfeedbackintheNCARCESM(2xCO2 run)

BroadbandLWcloudfeedbackSlaboceanrun:0.25Wm-2/KFullycoupledrun:0.31Wm-2/K

Notscaledwiththeband-by-banddecompositionofCRE

10-250cm-1,0.005(globalval) 250-500cm-1, 0.044 500-630cm-1,0.095

700-820cm-1,0.102 820-980cm-1,0.017 980-1080cm-1,0.017

1080-1180cm-1,2e-4 1180-1390cm-1,0.020 1390-1480cm-1,0.004

Fullycoupledrun CloudfeedbackineachbandWm-2/K

Short-termfluctuationof2003-2015(Preliminary)

• CESMsimulation:usingDessler’s methodtoobtainanestimationofshort-termcloudfeedback

• Observation:applyingYue etal.(2016)toMODIS,AIRSandCERESdatatoobtainthesamequantity

CESM,0.61Wm-2/K Obs,-0.21Wm-2/K

Long-termvs.short-termcontrast

BroadbandLWcloudfeedbackSlaboceanrun:0.25Wm-2/KFullycoupledrun:0.31Wm-2/KForcedSSTrun:0.61Wm-2/KObservation:-0.21Wm-2/K

ConclusionandDiscussion

• Spectraldecompositionhelpsrevealingcompensatingbiases.– Forcloud,ithelpsuntanglethebiasesinCTHandcloudamount

• ForRRTMGbandwidths,820-980cm-1contributesmosttotheCRE.But700-820cm-1contributesmosttothecloudfeedback(2xCO2 run)

• Thelong-termvs.short-termcloudfeedbackhasdifferentspectraldecomposition– Implications

T(z)

qH2O(z)qO3(z)qCH4(z)…

Aerosols

Tskin,es(v)

Geophysical variables Broadband Radiation Budget

F = Fv dvΔv∫

ITOA (v;θ,φ)Broadband Radiative Feedbacks

λx = −δx FδX

δXδTs

Spectral Radiances

Sounding communityEnergy budget and feedbackscommunity

Spectral FluxFv = dφ ITOA (v;θ,φ)cosθ sinθ dθ0

π2∫0

2π∫

λxv = −δx FvδX

δXδTs

Spectral Radiative Feedbacks

Cloud,ISCCPeffort

ThankYou!References:1. Huangetal.,2008:SpectrallyresolvedfluxesderivedfromcollocatedAIRSandCERES

measurementsandtheirapplicationinmodelevaluation,PartI:clearskyoverthetropicaloceans, JGR-Atmospheres,113,D09110,doi:10.1029/2007JD009219.

2. Chenetal.,2013:Comparisonsofclear-skyoutgoingfar-IRfluxinferredfromsatelliteobservationsandcomputedfromthreemostrecentreanalysisproducts,JournalofClimate,26(2),478-494,doi:10.1175/JCLI-D-12-00212.1.

3. Huangetal.,2014:Aglobalclimatologyofoutgoinglongwave spectralcloudradiativeeffectandassociatedeffectivecloudproperties,JournalofClimate,27,7475-7492,doi:10.1175/JCLI-D-13-00663.1.

4. Huang,X.L.,X.H.Chen,B.J.Soden,X.Liu,2014:ThespectraldimensionoflongwavefeedbacksintheCMIP3andCMIP5experiments,GeophysicalResearchLetters,41,doi:10.1002/2014GL061938.

5. Yue,Q.,B.H.Kahn,E.J.Fetzer,M.Schreier,S.Wong,X.H.Chen,X.L.Huang,2016:Observation-basedLongwaveCloudRadiativeKernelsDerivedfromtheA-Train,JournalofClimate,29,2023-2040.

Monthly griddedspectralfluxandCREavailableviahttp://www-personal.umich.edu/~xianglei/datasets.html.Thespectralradiative kernelsavailableuponrequest.

Backupslides

10-250cm-1,0.003 250-500cm-1, 0.046 500-630cm-1,0.072

700-820cm-1,0.070 820-980cm-1,0.023 980-1080cm-1,0.012

1080-1180cm-1,0.002 1180-1390cm-1,0.020 1390-1480cm-1,0.004

Slaboceanrun,6-35yrCloudfeedbackWm-2/K

CERES2s radiometriccalibrationuncertainty:1%(i.e.~2.5Wm-2)

Surface Type

Daytime Nighttime

OLRAIRS_Huang-OLRCERES(Wm-2)

OLRAIRS_Huang-OLRCERES (Wm-2)

Forest 0.58±1.43 -0.42±1.41

Savannas -0.03±2.52 0.68±1.50

Grasslands 0.19±2.61 0.63±1.65

Dark Desert -0.71±2.85 0.36±1.74

Bright Desert 1.67±2.62 1.42±2.28

Ocean 1.09±1.55 0.90±1.26

Allcollocatedclear-skyobservationsin2004(80°S-80°N)

(Chen et al., J Climate, 2013)

StratifyingOLRAIRS_Huang-OLRCERES(Wm-2):cloudyobservationsoverthelands

∆Tsc

f

Overdeserts Overnon-desertlands

<15k 15K-40K >40K <15k 15K-40K >40K

0.001-0.52.44±3.79

(0.9%)

3.25±5.12

(1.2%)

1.49±7.61

(0.5%)

2.34±2.86

(0.8%)

3.62±4.48

(1.3%)

2.84±5.94

(1.0%)

0.5-0.752.79±4.16

(1.1%)

3.34±7.80

(1.3%)

1.39±12.75

(0.5%)

2.90±3.86

(1.1%)

4.24±7.25

(1.7%)

2.61±11.38

(1.0%)

0.75-0.9992.67±3.67

(1.1%)

1.45±6.47

(0.6%)

-1.17±10.97

(-0.5%)

2.81±3.56

(1.2%)

3.14±6.68

(1.4%)

0.47±11.45

(0.2%)

0.999-1.02.61±2.80

(1.2%)

3.15±4.00

(1.6%)

1.28±6.64

(0.7%)

2.86±2.83

(1.3%)

4.04±4.33

(2.0%)

2.48±7.16

(1.5%)

CERES2s radiometriccalibrationuncertainty:1%(i.e.~2.5Wm-2)

GlobalOLRAIRS_Huang-OLRCERES:annualmeansandyeartoyearchanges

ConstructionoftheSRK

Validation:comparisonswiththePRPresults

Validation:comparisonswiththePRPresults