climate - accueil lmdcrlmd/these/seminaire_civilengineering_27april2… · climate changes...

Using water stable isotopi measurements tobetter evaluate the atmospheri and land surfa e omponents of limate modelsCamille RisiCIRES, Boulderwith ontribution of:Analysis: S Bony, D Noone, C CastetTES/SCIAMACHY: J Worden, J Lee, D Brown, C FrankenbergMIPAS:ACE: G Stiller, M Kiefer, B Funke, K Walker, P Bernath,FTIR: M S hneider, D Wun h, P Wennberg, V Sherlo k, N Deuts her, D Gri�thin-situ/MIBA: R Uemura, J. Ogée, T. Baria , L. Wingate, N. Raz-YaseefSWING2: C SturmHydrologi S ien es and Water Resour es Engineering SeminarSeries, 27 April 2011

Un ertainties in limate proje tions

CO2increase

temperatureincrease

climatefeedbacks

IPCC AR4

multi−model mean

surf

ace

glob

al w

arm

ing

(°C

)

Introdu tion 2/31


CO2increase

temperatureincrease

climatefeedbacks

Key uncertaintiesin climate models:

IPCC AR4

atmospheric processes

− atmospheric convection− clouds

− boundary layer− relative humidity

multi−model mean

surf

ace

glob

al w

arm

ing

(°C

)

Introdu tion 2/31


CO2increase

IPCC AR4

precipitation changeby 2100 (%)

<2/3 agree on sign

temperatureincrease

regionalprecipitation

changesclimatefeedbacks


IPCC AR4




multi−model mean

surf

ace

glob

al w

arm

ing

(°C

)

Introdu tion 2/31


CO2increase

land usechange

IPCC AR4

Boé et Terray2008


<2/3 agree on sign

evapo−transpiration changeby 2100 (mm/d)

<70% agree on sign

temperatureincrease


changes

land surfacehydrological

changesclimatefeedbacks land−atmosphere

feedbacks


IPCC AR4




multi−model mean

surf

ace

glob

al w

arm

ing

(°C

)

Introdu tion 2/31


CO2increase

land usechange

IPCC AR4

Boé et Terray2008


<2/3 agree on sign


<70% agree on sign

temperatureincrease


changes



feedbacks


IPCC AR4

atmospheric processes land surface processes− sensitivity of surface fluxes

to soil moisture− atmospheric convection− clouds

− soil/groundwater storage− spatial heterogeneities


multi−model mean

surf

ace

glob

al w

arm

ing

(°C

)

Introdu tion 2/31

Water isotopi omposition◮ H162 O, HDO, H182 O, H172 O, fra tionation

HDO

H16

2O

Introdu tion 3/31

Water isotopi omposition◮ H162 O, HDO, H182 O, H172 O, fra tionation◮ re ords phase hanges

Continental recyclingEvaporation

ConvectionCondensation

Precipitation

HDO

H16

2O

Introdu tion 3/31




Precipitation

HDO

H16

2O

isotopiccomposition

variablesmeteorological

and variability

present−dayclimate processes

physicalfuture climate

Introdu tion 3/31




Precipitation

HDO

H16

2O

isotopiccomposition

processevaluation


and variability

present−dayclimate

combiningdata

processesphysical

future climate

Introdu tion 3/31




Precipitation

HDO

H16

2O

isotopiccomposition

processevaluation

strengthenprojections credibility


and variability

present−dayclimate

combiningdata

processesphysical

future climate

Introdu tion 3/31

General strategy

controlling isotopicunderstand processes

compositionIntrodu tion 4/31

General strategy


composition

design observable,process−oriented

isotopic diagnostics

Introdu tion 4/31

General strategy


composition

climate modelsapply to

detect bias in models

propose improvements

understand reasons



Introdu tion 4/31

General strategy


composition

consequences onfuture changes

quantify, prioritizeuncertainties

climate modelsapply to

detect bias in models

propose improvements

understand reasons



Introdu tion 4/31

Outline

CO2increase

land usechange

IPCC AR4

Boé et Terray2008


<2/3 agree on sign


<70% agree on sign

temperatureincrease


changes



feedbacks


IPCC AR4





multi−model mean

surf

ace

glob

al w

arm

ing

(°C

)

Introdu tion 5/31

Outline

3)

2)

1)

CO2increase

land usechange

IPCC AR4

Boé et Terray2008


<2/3 agree on sign


<70% agree on sign

temperatureincrease


changes



feedbacks


IPCC AR4





multi−model mean

surf

ace

glob

al w

arm

ing

(°C

)

Introdu tion 5/31

1) Pro esses ontrolling relative humidity◮ tropi al/subtropi al free tropospheri relative humidity (RH)impa ts:

◮ water vapor feedba k (Soden et al 2008)◮ loud feedba ks (Sherwood et al 2010)

1) Pro esses ontrolling humidity 6/31



◮ but:◮ signi� ant dispersion in limate models (Sherwood et al 2010)◮ moist bias in the mid/upper troposphere (John and Soden 2005)




◮ but:◮ signi� ant dispersion in limate models (Sherwood et al 2010)◮ moist bias in the mid/upper troposphere (John and Soden 2005)

=⇒ need pro ess-based evaluation of RH in limate models1) Pro esses ontrolling humidity 6/31

Sensitivity tests to RH pro esses

800

600

200

100

0°

400

1000

P (hPa)

30°N

condensate detrainementin clouds

large−scale

rain evaporation

boundary layermixing

vertical mixing

circulation

dehydration

lateral and

Couhert et al 2010Folkins and Martin 2005Pierrehumbert 1998Sherwood et al 1996Wright et al 2009



800

600

200

100

0°

400

1000

P (hPa)

30°N

condensate detrainement

LMDZ−iso (Risi et al 2010a):

in clouds

large−scale

rain evaporation


vertical mixing

circulation

dehydration

lateral and

ontrol: AR4 version (19 levels)

AIRS datarelative humidity (%)

pression(hPa)

3020 40 50 60 70 801000

100300200400500600700800900Couhert et al 2010Folkins and Martin 2005

Pierrehumbert 1998Sherwood et al 1996Wright et al 2009



800

600

200

100

0°

400

1000

P (hPa)

30°N



in clouds

large−scale

rain evaporation


vertical mixing

circulation

dehydration

3 reasonsfor a

moist bias

lateral and

ontrol: AR4 version (19 levels)


pression(hPa)

3020 40 50 60 70 801000





800

600

200

100

0°

400

1000

P (hPa)

30°N



in clouds

large−scale

rain evaporation


vertical mixing

circulation

dehydration

3 reasonsfor a

moist bias

lateral and

ontrol: AR4 version (19 levels)di�usive verti al adve tion


pression(hPa)

3020 40 50 60 70 801000





800

600

200

100

0°

400

1000

P (hPa)

30°N



in clouds

large−scale

rain evaporation


vertical mixing

circulation

dehydration

3 reasonsfor a

moist bias

lateral and

ontrol: AR4 version (19 levels)di�usive verti al adve tionσq/10


pression(hPa)

3020 40 50 60 70 801000





800

600

200

100

0°

400

1000

P (hPa)

30°N


LMDZ−iso (Risi et al 2010):

in clouds

large−scale

rain evaporation


vertical mixing

circulation

dehydration

3 reasonsfor a

moist bias

lateral and


pression(hPa)

3020 40 50 60 70 801000

100300200400500600700800900

ontrol: AR4 version (19 levels)di�usive verti al adve tionσq/10

ǫp/2

Couhert et al 2010Folkins and Martin 2005Pierrehumbert 1998Sherwood et al 1996Wright et al 2009


Isotopi measurements(Worden et al 2007)TES

satellites

total column:(Frankenberg et al 2009)

SCIAMACHY

(Steinwagner et al 2010)(Nassar et al 2007)

MIPASACE−FTS

200

100

0°1000

30°N

P (hPa)

400

600

800




SCIAMACHY

satellites

ground−basedFTIR

total column(5 TCCON/NDACC sites:

2 US, 2 Australia,1 New−Zealand)

Schneider et al 2009)

profile(Canaries islands:


MIPASACE−FTS

200

100

0°1000

30°N

P (hPa)

400

600

800




SCIAMACHY

satellites

ground−basedFTIR




profile(Canaries islands:in situ

(GNIP−vapor,Angert et al 2007, Uemura et al 2007)


MIPASACE−FTS

200

100

0°1000

30°N

P (hPa)

400

600

800




SCIAMACHY

satellites

ground−basedFTIR




profile(Canaries islands:in situ

(GNIP−vapor,Angert et al 2007, Uemura et al 2007)


MIPASACE−FTS

200

100

0°1000

30°N

P (hPa)

400

600

800

◮ model-data omparison: ollo ation; simulations nudged byECMWF; averaging kernels; spatial/temporal variations1) Pro esses ontrolling humidity 8/31

Zonal anual mean

data

−180

−160

−140

−120

−100

−80

Eq 30N60S 60N30S

−150

−200

−250

−100

Eq 30N60S 60N30S

−60

Eq 30N60S 60N30S

−250

−200

−150

Eq 30N60S 60N30S−800

−300

−200

−400

−500−600

−700

Eq 30N60S 60N30S−600

−500

−400

−300

LMDZ "ǫp/2"LMDZ "σq /10"LMDZ "di�usive adve tion"LMDZ ontrolδD

(h)δD

(h)

200hPa, MIPAS

300hPa, ACEδD

(h) SCIAMACHY andground-based FTIRTotal olumn,

δD

(h) Surfa e, in-situ

600 hPa, TESδD

(h)1) Pro esses ontrolling humidity 9/31

Zonal Seasonal variations (JJA-DJF)

data

Eq 30N60S 60N30S

0

50

100

−100

−50

Eq 30N60S 60N30S

100

50

0

Eq 30N60S 60N30S−20

0

20

40

60

80

Eq 30N60S 60N30S

−100

0

100

200

Eq 30N60S 60N30S

0

100

−100

−50

50

150

−150

LMDZ "ǫp/2"LMDZ "σq /10"LMDZ "di�usive adve tion"LMDZ ontrol200hPa, MIPAS

∆δD

(h)

∆δD

(h)300hPa, ACEand ground-based FTIR

∆δD

(h)ground-based FTIRSCIAMACHY andTotal olumn

∆δD

(h)

∆δD

(h)

600 hPa, TESand ground-based FTIR

Surfa e, in-situ


What auses the moist biases in GCMs?

−60

−40

−20

0

20

40

60

20 25 30 35 40 45 50 55

Ex essive ondensatedetrainementInsu� ientin-situ ondensationverti al adve tionEx essively di�usiveControl 400 hPa, 15◦N-30◦N mean

relative humidity (%)JJA-DJF∆δD(h)

Sensitivity tests:with LMDZ: ACE/AIRSdata



−60

−40

−20

0

20

40

60

20 25 30 35 40 45 50 55

Ex essive ondensatedetrainementInsu� ientin-situ ondensationverti al adve tionEx essively di�usiveControl 400 hPa, 15◦N-30◦N mean



◮ robustness? additional tests, theoreti al understanding1) Pro esses ontrolling humidity 11/31


−60

−40

−20

0

20

40

60

20 25 30 35 40 45 50 55

Ex essive ondensatedetrainementInsu� ientin-situ ondensationverti al adve tionEx essively di�usiveControl

ECHAMMIROCHadAM GSMGISSSWING2 models:CAM2

400 hPa, 15◦N-30◦N mean



◮ robustness? additional tests, theoreti al understanding◮ frequent reason for moist bias=ex essively di�usive adve tion1) Pro esses ontrolling humidity 11/31

What impa t on humidity proje tions?

−2.5

−2

−1.5

−1

−0.5

0

0.5

1

1.5

25 30 35 40 45 50 55 60 65 70 75

AIRS

∆

RH(%/K)

present-day RH (%/K)

tropi al average, 200hPa, 2xCO2 ontrol simulationLMDZ testsdi�usive adve tionσq /10ǫp/2CMIP3 models


What impa t on humidity proje tions?

−2.5

−2

−1.5

−1

−0.5

0

0.5

1

1.5

25 30 35 40 45 50 55 60 65 70 75

AIRS

∆

RH(%/K)

present-day RH (%/K)

tropi al average, 200hPa, 2xCO2 ontrol simulationLMDZ testsdi�usive adve tionσq /10ǫp/2CMIP3 models

◮ How a moist bias a�e t humidity hange proje tions dependson the reason for the bias1) Pro esses ontrolling humidity 12/31

What impa t on limate sensitivity?◮ radiative kernel de omposition (Soden et al 2008)

controldiffusive advection insufficient condensationexcessive détrainement

feed

back

par

amet

er (

W/m

2/K

)

−4

−3

−2

−1

0

1

2

3

surfa

ce albedo

water vapor

lapse ra

te

Planck

clouds

total

Global average



decrease in

humidityupper−tropospheric


feed

back

par

amet

er (

W/m

2/K

)

−4

−3

−2

−1

0

1

2

3

surfa

ce albedo

water vapor

lapse ra

te

Planck

clouds

total

Global average



decrease in

humidityupper−tropospheric

decrease infrontal clouds

decreasein low−levelsubtropicalmarine clouds


feed

back

par

amet

er (

W/m

2/K

)

−4

−3

−2

−1

0

1

2

3

surfa

ce albedo

water vapor

lapse ra

te

Planck

clouds

total

Global average


Summary on relative humidity◮ Water vapor isotope measurements as observationaldiagnosti s to understand the reasons for a moist bias in limate models


Summary on relative humidity◮ Water vapor isotope measurements as observationaldiagnosti s to understand the reasons for a moist bias in limate models◮ Ex essive verti al di�usion during water vapor transport is awidespread ause of moist bias in limate models


Summary on relative humidity◮ Water vapor isotope measurements as observationaldiagnosti s to understand the reasons for a moist bias in limate models◮ Ex essive verti al di�usion during water vapor transport is awidespread ause of moist bias in limate models◮ Understanding this reason is all the more important ashumidity hange proje tions depends on the reason for themoist bias


Summary on relative humidity◮ Water vapor isotope measurements as observationaldiagnosti s to understand the reasons for a moist bias in limate models◮ Ex essive verti al di�usion during water vapor transport is awidespread ause of moist bias in limate models◮ Understanding this reason is all the more important ashumidity hange proje tions depends on the reason for themoist bias◮ Consequen es on limate hange ompli ated by dynami aland loud feedba ks1) Pro esses ontrolling humidity 14/31

2) Conve tive pro esses◮ mi rophysi al pro esses? (Emanuel and Pierrehumbert 1996)

reevaporationdowndraftsunsaturated

detrainement onve tive onve tiveas ent subsiden eradiative100hPa

2) Conve tive pro esses 15/31

Pro esses along squall lines◮ rain sampled every 5 mins in Niamey during AMMA ampaign

−8−7.5

−7−6.5

0 50 100 150 200 250

15km

200km

time (minutes after the start of the rain)

11 August 2006(Risi et al 2010b QJRMS)

convective zone stratiform zone

convectiveascent

front−to−rear flowmeso−scale

ascent

meso−scale descent

rear−to−front flow

cold poolfrontgust

δ18O (h)2) Conve tive pro esses 16/31

Pro esses along squall lines◮ rain sampled every 5 mins in Niamey during AMMA ampaign◮ interpretation with 2D model of transport/mi rophysi s

−8−7.5

−7−6.5

0 50 100 150 200 250

15km

200km




convectiveascent


ascent



cold poolfrontgust



−8−7.5

−7−6.5

0 50 100 150 200 250

15km

200km




rainenrichmentthroughreevaporation

convectiveascent


ascent



cold poolfrontgust



−8−7.5

−7−6.5

0 50 100 150 200 250

15km

200km




subsidenceof dry anddepleted airrainenrichmentthroughreevaporation

convectiveascent


ascent



cold poolfrontgust


Conve tive/large-s ale �uxes ompensatingsubsiden e large-s aleas ent

reevaporationdetrainement onve tive

large-s ale ondensation

downdraftsunsaturated onve tive s heme large-s ale

onve tiveas ent100hPa


Conve tive/large-s ale �uxes

more di�usive verti al adve tionstronger ondensate detrainementless large-s ale ondensationless large-s ale pre ipitation ontrol: AR4

ompensatingsubsiden e large-s aleas entverti al di�usionreevaporationdetrainement onve tive

large-s ale ondensation

downdraftsunsaturated onve tive s heme large-s ale

onve tiveas ent

Sensitivity tests with LMDZ:

100hPa


Conve tive ontribution to water budget−50

−40

−30

−20

−10

0

10

20

−0.4 −0.2 0 0.2 0.4 0.6 0.8 1

Amazon600hPaTES data∆

δD

DJF-JJA(h)

∆PLS/∆Ptot DJF-JJA ontrolverti al adve tion more di�usivestronger ondensate detrainmentless large-s ale ondensationless large-s ale pre ipitation



−40

−30

−20

−10

0

10

20

−0.4 −0.2 0 0.2 0.4 0.6 0.8 1


δD

DJF-JJA(h)

∆PLS/∆Ptot DJF-JJAPconv PLS

enri hmentby ondensateevaporation by ondensatedepletionloss ontrolverti al adve tion more di�usivestronger ondensate detrainmentless large-s ale ondensationless large-s ale pre ipitation

100hPa

subsiden e verti al di�usionas ent, ondensation onve tivedetrainement

environmental large-s alelarge-s ale



−40

−30

−20

−10

0

10

20

−0.4 −0.2 0 0.2 0.4 0.6 0.8 1


δD

DJF-JJA(h)

∆PLS/∆Ptot DJF-JJAPconv PLS

enri hmentby ondensateevaporation by ondensatedepletionloss ontrolverti al adve tion more di�usivestronger ondensate detrainmentless large-s ale ondensationless large-s ale pre ipitation

100hPa

subsiden e verti al di�usionas ent, ondensation onve tivedetrainement

environmental large-s alelarge-s ale

◮ PLS/Ptot ill-de�ned quantity, but in�uen es loudiness,intra-seasonal variability, hemi al tra or transport2) Conve tive pro esses 18/31

Summary on onve tion1000

800

600

400

200100

P (hPa)

environmentalas entunsaturatedreevaporation

large-s aleas entin-situ ondensation

downdrafts onve tivesubsiden e onve tion vs large-s aleunsaturated downdraftsrain reevaporation



800

600

400

200100

P (hPa)




◮ Perspe tives:◮ New physi s of LMDZ for AR5 (Rio et al 2009)



800

600

400

200100

P (hPa)




◮ Perspe tives:◮ New physi s of LMDZ for AR5 (Rio et al 2009)◮ Design diagnosti s based on loud pro esses/isotopes link:

◮ High frequen y data: e.g. ground-based remote-sensing◮ A-train synergy: TES+CALIPSO/Cloudsat2) Conve tive pro esses 19/31


800

600

400

200100

P (hPa)




◮ Perspe tives:◮ New physi s of LMDZ for AR5 (Rio et al 2009)◮ Design diagnosti s based on loud pro esses/isotopes link:

◮ High frequen y data: e.g. ground-based remote-sensing◮ A-train synergy: TES+CALIPSO/Cloudsat

◮ impa t of misrepresentation of onve tive pro esses onpre ipitation hanges?2) Conve tive pro esses 19/31

3) Land atmosphere feedba ks

precipitation

atmospheric properties

soil moisture

surface fluxes− evaporation− transpiration− sensible heat flux

− moisture− boundary layer

3) Land-atmosphere feedba ks 20/31

3) Land atmosphere feedba ks◮ inter-model spread (Koster et al, Guo et al 2006)

uncertaintiesin atmospheric

component

uncertaintiesin land surface

component

precipitation


soil moisture




3) Land atmosphere feedba ks◮ inter-model spread (Koster et al, Guo et al 2006)

uncertaintiesin atmospheric

component

uncertaintiesin land surface

component

precipitation


soil moisture



continentalrecycling

partitionning


Evapo-transpiration partitioning◮ ORCHIDEE-iso land surfa e model (Risi et al in rev,a)

3.2 3.4 3.6 3.8 4 4.2 4.4 4.6 4.8 5 5.2−0.2

0

0.2

0.4

0.6

0.8

1

1.2

1.4

controlstomatal resistance /5no drainage, only surface runoffsoil capacity /2less vegetation coverroot extraction depth /4

Le Bray (France)

observations

D

E/I (%)δ

18O

soil−

δ18O

p

(h)Rs

ET Pmoisturesoil I


Does it matter for limate hange?P

rese

nt−

day

more bare soil+stronger stomatalresistan eORCHIDEEo�-line simulationsat Le Bray: ontrol

δ18Os (h)-6-5-4-3 1hqsoil (mm)

240260220

5%0

10%0.40.81.2

ET (mm/d)E/ET (%)1020300



rese

nt−

day

clim

ate

chan

ge more bare soil+stronger stomatalresistan eORCHIDEEo�-line simulationsat Le Bray: ontrol


240260220

5%0

10%0.40.81.2

ET (mm/d)E/ET (%)1020300de reases by 50%0

-100-150-50 40% less

∆qsoil (mm/d)Pre ipitation

de rease3) Land-atmosphere feedba ks 22/31


rese

nt−

day

clim

ate

chan

ge

Temperaturein reases by 4K60% morein rease

∆ ET (mm/d)0.40.30.20.10more bare soil+stronger stomatalresistan e

ORCHIDEEo�-line simulationsat Le Bray: ontrol


240260220

5%0

10%0.40.81.2

ET (mm/d)E/ET (%)1020300de reases by 50%0

-100-150-50 40% less

∆qsoil (mm/d)Pre ipitation

de rease3) Land-atmosphere feedba ks 22/31

Impa t on response to pre ipitationFun tional relationships (Koster and Milly 1996)

0 50 100 150 200 250 0

1

2

3

4

ET

ET+R

control

ET(mm)

soil moisture (mm)3) Land-atmosphere feedba ks 23/31


0 50 100 150 200 250 0

1

2

3

4

E remains strongin winter despitelow soil moistureand limits runoff

controlmore bare soil, stronger stomatal resistance

ET

ET+R

ET(mm)



0 50 100 150 200 250 0

1

2

3

4



ET

ET+R

P

ET(mm)



0 50 100 150 200 250 0

1

2

3

4



P decreasesET

ET+R

P

ET(mm)



0 50 100 150 200 250 0

1

2

3

4



P decreasesET

ET+R

P

ET(mm)

soil moisture (mm)soil moisture de reases less strongly3) Land-atmosphere feedba ks 23/31

Summary on evapo-transpiration partitioning◮ Isotopi measurements an dete t misrepresentation in baresoil evaporation/transpiration partitioning, even if it has noimpa t on traditional observable variables


Summary on evapo-transpiration partitioning◮ Isotopi measurements an dete t misrepresentation in baresoil evaporation/transpiration partitioning, even if it has noimpa t on traditional observable variables◮ This partitioning impa ts the land surfa e hydrologi alresponse to limate hange, even if it doesn't impa tpresent-day hydrology


Summary on evapo-transpiration partitioning◮ Isotopi measurements an dete t misrepresentation in baresoil evaporation/transpiration partitioning, even if it has noimpa t on traditional observable variables◮ This partitioning impa ts the land surfa e hydrologi alresponse to limate hange, even if it doesn't impa tpresent-day hydrology◮ Perspe tives:

◮ To what extent does it ontribute to inter-model spread inhydrologi al proje tions?3) Land-atmosphere feedba ks 24/31

Summary on evapo-transpiration partitioning◮ Isotopi measurements an dete t misrepresentation in baresoil evaporation/transpiration partitioning, even if it has noimpa t on traditional observable variables◮ This partitioning impa ts the land surfa e hydrologi alresponse to limate hange, even if it doesn't impa tpresent-day hydrology◮ Perspe tives:

◮ To what extent does it ontribute to inter-model spread inhydrologi al proje tions?◮ To what extent does it feedba k on pre ipitation hanges?


Isotopi signature of evaporative origin

5 10 15 20 25 30 4050 60 70 80 90

120W 60W 0 60E 120E

60N

30N

Eq

30S

60S

Water tagging:ET P

evapo-transpiration (rcon)% vapor from ontinental3) Land-atmosphere feedba ks 25/31

Isotopi signature of evaporative origin

−30 −24 −18 −12 −6 0

120

−30

0

60

90

30

5 10 15 20 25 30 4050 60 70 80 90

120W 60W 0 60E 120E

60N

30N

Eq

30S

60S

Water tagging:

d-ex ess(h)

δ18

O (h)

monthly, all tropi al land pointsPDF of vapor ompositionbare soilevaporation

transpirationtotal vaporo eani vapor

ET P

evapo-transpiration (rcon)% vapor from ontinental3) Land-atmosphere feedba ks 25/31

Water isotopes and ontinental re y lingKoster et al 2006

Land-atmosphere feedba ks"hot spots"

de rease in pre ip varian e when soil moisture is pres ribed


Water isotopes and ontinental re y ling

120E60E060W120W

30S

Eq

30N

60N

60S−0.8

−0.6

−0.4

0.2−0.2

0.4

0.6

0.8

Koster et al 2006 orrelation δ18O - rcon, intra-seasonal s ale, annual meanLand-atmosphere feedba ks"hot spots"

de rease in pre ip varian e when soil moisture is pres ribed


Diagnosing land-atmosphere feedba ksET րsoil moisture րmoisture onvergen e ր P ր

P ր

strong pre ipitation omposite minus seasonal average:


Diagnosing land-atmosphere feedba ks

30S

Eq

30N

60N

60W120W 0 60E 120E

−10 −5 −2 2 5 10

ET րsoil moisture րmoisture onvergen e ր P ր

P ր

strong pre ipitation omposite minus seasonal average:∆rcon (%) JJA


Isotopi signature of feedba ks

60W120W 0 60E 120E

−10 −5 −2 2 5 10

30S

Eq

30N

60N

8 1240−4

−4 0 4 8 12 16

04

−4

8

12

−8−12Am

azon

, DJF

Wes

tern

US

, JJA

−4

0

4

8

12

16

positivefeedba kland-atmosphere ontrol bylarge-s ale onvergen e

∆rcon (%) JJA∆

δD

v(h) Strong pre ipitation ompositeminus seasonal average:

∆rcon (%)

∆rcon (%)∆

δD

v

(h)


Could we dis riminate bewteen di�erentsimulated feedba ks?Eq

30N

60N

30S

−10 −5 −2 2 5 10

Eq

30N

60N

30S

60W120W 0 60E 120E

60W120W 0 60E 120E

control

less bare soil

∆rcon (%) JJA, ontrol3) Land-atmosphere feedba ks 29/31

Could we dis riminate bewteen di�erentsimulated feedba ks?Eq

30N

60N

30S

−10 −5 −2 2 5 10

Eq

30N

60N

30S

60W120W 0 60E 120E

60W120W 0 60E 120E

−3 −2 −1 0 1 2 3 4 5−5

−4

−3

−2

−1

0

−8−6−4−2 0 2 4 6

−8 −6 −4 −2 0 2 4 6

control

less bare soil

Am

azon

DJF

Wes

tern

US

, JJA

ontrolless bare soil∆rcon (%) JJA, ontrol

∆rcon (%)

∆δD

v

(h)∆rcon (%)

∆δD

v

(h)

positiveland-atmospherefeedba klarge-s ale onvergen e ontrol by3) Land-atmosphere feedba ks 29/31

Summary on land-atmosphere feedba ks◮ Water vapor isotopes re ord land-atmosphere feedba ks atintra-seasonal s ale and ould help dis riminate betweendi�erent simulations


Summary on land-atmosphere feedba ks◮ Water vapor isotopes re ord land-atmosphere feedba ks atintra-seasonal s ale and ould help dis riminate betweendi�erent simulations◮ Perspe tives:

◮ Use data: in-situ, satellite (GOSAT)3) Land-atmosphere feedba ks 30/31


◮ Use data: in-situ, satellite (GOSAT)◮ Robustness of isotopi diagnosti s?

◮ model inter- omparisons: ORCHIDEE, isoLSM, soon CLMand ORCHIDEE-multi-layer3) Land-atmosphere feedba ks 30/31


◮ Use data: in-situ, satellite (GOSAT)◮ Robustness of isotopi diagnosti s?

◮ model inter- omparisons: ORCHIDEE, isoLSM, soon CLMand ORCHIDEE-multi-layer◮ Relevan e for hydrologi al proje tions?

◮ Do some pro esses determine behavior at intra-seasonals ales, and in ontext of- global warming- land use hange (deforestation, irrigation)3) Land-atmosphere feedba ks 30/31

Con lusion200

400

800

600

100P (hPa)

unsaturateddowndraftsreevaporationsurfa ewaterbudget onve tionvs large-s ale ontinentalre y lingadve tiondi�usive

Con lusion 31/31

Con lusion200

400

800

600

100P (hPa)


◮ Ultimate goal: isotopi diagnosti s to evaluate models andtheir proje tions:Con lusion 31/31

Con lusion200

400

800

600

100P (hPa)


◮ Ultimate goal: isotopi diagnosti s to evaluate models andtheir proje tions:◮ new isotopi data

Con lusion 31/31

Con lusion200

400

800

600

100P (hPa)


◮ Ultimate goal: isotopi diagnosti s to evaluate models andtheir proje tions:◮ new isotopi data◮ new model-data omparison methodologies

Con lusion 31/31

Con lusion200

400

800

600

100P (hPa)


◮ Ultimate goal: isotopi diagnosti s to evaluate models andtheir proje tions:◮ new isotopi data◮ new model-data omparison methodologies◮ model inter- omparisonsCon lusion 31/31

Con lusion200

400

800

600

100P (hPa)


◮ Ultimate goal: isotopi diagnosti s to evaluate models andtheir proje tions:◮ new isotopi data◮ new model-data omparison methodologies◮ model inter- omparisons◮ pro ess/feedba ks studies omparing models behavior forpresent limate and for proje tionsCon lusion 31/31

Suplementary material

Supplementary material 32/31

Evaluation against SCIAMACHY120W180W 060W 60E 120E 180E

30S

Eq

30N

30S

Eq

30N

120W180W 060W 60E 120E 180E

δD (h) total olumn-200 -180 -170 -160 -150 -140 -130 -120 -110 -100

-80 -60 -40 -20 -10 10 20 40 60 80∆δD (h) total olumn JJA-DJF

LMDZ ontrolSCIAMACHY data

Risi et al in rev,bSupplementary material 33/31

Evaluation against TES30S

Eq

30N

30S

Eq

30N

120W180W 060W 60E 120E 180E120W180W 060W 60E 120E 180E

−230 −220 −210 −200 −190 −180 −170 −160 −150

120W180W 060W 60E 120E 180E

30S

Eq

30N

120W180W 060W 60E 120E 180E

−80 −50 −30 −20 −10 10 20 30 50 80

TES data LMDZ (-31h)

LMDZTES data δD (h) 600hPa annual mean

∆δD (h) 600hPa JJA-DJF Risi et al in rev,bSupplementary material 34/31

What auses the moist bias?

−60

−40

−20

0

20

40

60

20 25 30 35 40 45 50 55

Ex essive ondensatedetrainementInsu� ientin-situ ondensationverti al adve tionEx essively di�usiveControl

ECHAMMIROCHadAM GSMGISSSWING2 models:CAM2verti al resolution

400 hPa, 15◦N-30◦N mean




Consequen es on proje tions 150

200

250

300

350 0 50 100 150 200 250 25 30 35 40 45 50 55 60 65 70 75

−2.5

−2

−1.5

−1

−0.5

0

0.5

1

1.5

AIRS Relative humidity (%)present-day RH (%/K)

∆

RH(%/K)

Pressure(hPa)

tropi al average, 200hPa, 2xCO2

σq /10ǫp/2

ontrol simulationdi�usive adve tion EverythingLMDZ tests pre ipitates:presentSupplementary material 36/31


200

250

300

350

(Harmann and Larson 2002)

fixed anviltemperature

0 50 100 150 200 250 25 30 35 40 45 50 55 60 65 70 75−2.5

−2

−1.5

−1

−0.5

0

0.5

1

1.5


∆

RH(%/K)

Pressure(hPa)


σq /10ǫp/2 present2xCO2 ontrol simulationdi�usive adve tion EverythingLMDZ tests pre ipitates:



200

250

300

350



0 50 100 150 200 250 25 30 35 40 45 50 55 60 65 70 75−2.5

−2

−1.5

−1

−0.5

0

0.5

1

1.5


∆

RH(%/K)

Pressure(hPa)


σq /10ǫp/2 present2xCO2 ontrol simulationdi�usive adve tion EverythingLMDZ tests pre ipitates: Additionalupward transportpresentSupplementary material 36/31


200

250

300

350



0 50 100 150 200 250

less warmanomaly

anomaly warm

25 30 35 40 45 50 55 60 65 70 75−2.5

−2

−1.5

−1

−0.5

0

0.5

1

1.5


∆

RH(%/K)

Pressure(hPa)


σq /10ǫp/2 present2xCO2 2xCO2present ontrol simulationdi�usive adve tion EverythingLMDZ tests pre ipitates: Additionalupward transport



200

250

300

350



0 50 100 150 200 250

less warmanomaly

anomaly warm

25 30 35 40 45 50 55 60 65 70 75−2.5

−2

−1.5

−1

−0.5

0

0.5

1

1.5


∆

RH(%/K)

Pressure(hPa)


σq /10ǫp/2 present2xCO2 2xCO2present ontrol simulationdi�usive adve tion EverythingLMDZ tests pre ipitates: Additionalupward transportCMIP3 modelsSupplementary material 36/31

Upper troposphere detrainment120W180W 060W 60E 120E 180E

30S

Eq

30N

60N

60S

-700 -640 -600 -560 -520 -480 -440 -400 -360-320δD (h)

MIPAS data at 200hPa, annual



30S

Eq

30N

60N

60S

120W180W 060W 60E 120E 180E

30S

Eq

30N

60N

60S-700 -640 -600 -560 -520 -480 -440 -400 -360-320

LMDZ ontrol

δD (h)




30S

Eq

30N

60N

60S

120W180W 060W 60E 120E 180E

30S

Eq

30N

60N

60S

50

0 0.01 0.02 0.03

0.011

0.012

0.013

0.014

0.015

0.04

−5

0

5

10

15

20

−10

-700 -640 -600 -560 -520 -480 -440 -400 -360-320

LMDZ ontrol

ontrolverti al adve tion more di�usivestronger ondensate detrainmentless large-s ale ondensationless large-s ale pre ipitation

Di�eren e 15◦S-15◦N minus

δD (h)


∆δD

(h)30◦S-30◦N at 200hPa

∆q

(g/kg)

moistening by detrainement(g/kg/day)

ACE: 30hMIPAS: 50h



30S

Eq

30N

60N

60S

120W180W 060W 60E 120E 180E

30S

Eq

30N

60N

60S

50

0 0.01 0.02 0.03

0.011

0.012

0.013

0.014

0.015

0.04

−5

0

5

10

15

20

−10

-700 -640 -600 -560 -520 -480 -440 -400 -360-320

LMDZ ontrol CAMMIROCHadAMGISSECHAMGSM

ontrolverti al adve tion more di�usivestronger ondensate detrainmentless large-s ale ondensationless large-s ale pre ipitation

Di�eren e 15◦S-15◦N minus

δD (h)


∆δD

(h)30◦S-30◦N at 200hPa

∆q

(g/kg)

moistening by detrainement(g/kg/day)

ACE: 30hMIPAS: 50h


Soil water isotopes

30S30N60N60SEq

-4 -2 0 2 4 6 8120W 60W 0 60E 120E

120W 60W 0 60E 120E30S30N60N60SEq

LMDZ-ORCHIDEE

δ18Osoil − δ18Oprecip (h)

GNIP+USNIP+MIBA


Estimating evapotranspiration partitioning60N

30N

Eq

30S

60S

1 3 5 7 10 15 20 30 50 70 90 95

120W 60W 0 60E 120E

δ18Osoil − δ18Op

δ18OvRH, T

E/P (%)

isotopi "measurements"estimated from simulated


Estimating evapotranspiration partitioning60N

30N

Eq

30S

60S

60N

30N

Eq

30S

60S

1 3 5 7 10 15 20 30 50 70 90 95

120W 60W 0 60E 120E

δ18Osoil − δ18Op

δ18OvRH, T

120W 60W 0 60E 120Esimulated by LMDZ-ORCHIDEE

E/P (%)

isotopi "measurements"estimated from simulated

r=0.91Supplementary material 39/31

Estimating ontinental re y ling 0

0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 1

y=x

Risi et al 2010d

1 − rcon

rcon

June-July 2006 in Niamey, daily

dδvoce

dxknown x estimat

edre y ling

simulated re y lingd ( r on1− r on)

/dx =

dδv/dx − dδvo e/dxδp − δvSupplementary material 40/31


0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 1

y=x

Risi et al 2010ddδvoce

dxdependslinearly on pre ipitation

1 − rcon

rcon


dδvoce

dxknown x estimat

edre y ling


/dx =

dδv/dx − dδvo e/dxδp − δvSupplementary material 40/31


0.2

0.4

0.6

0.8

1

0 0.2 0.4 0.6 0.8 1

y=x

Risi et al 2010ddδvoce

dxdependslinearly on pre ipitation

1 − rcon

rcon


dδvoce

dxknown x estimat

edre y ling


/dx =

dδv/dx − dδvo e/dxδp − δv

◮ Main limitation in using vapor isotopi measurements for ontinental re y ling: understanding atmospheri ontrolsSupplementary material 40/31

Monitoring land-atmosphere feedba ksrelated to land use hange or global warming90W 30W60W

-80 -50 -20-5 5 20 50 80 3020105-5-10-20-30

Amazon, JJA

40S20SEq20N

40S20SEq20N

90W 30W60WDeforestation

4xCO2∆rcon (%)pre ipitation hange (%)Supplementary material 41/31


-80 -50 -20-5 5 20 50 80 3020105-5-10-20-30

-40 -30 -20 -10 0-21-10

∆rcon (%)

Amazon, JJA

40S20SEq20N

40S20SEq20N


4xCO2∆rcon (%)pre ipitation hange (%)

∆δ

18O

v

(h)



-80 -50 -20-5 5 20 50 80 3020105-5-10-20-30 0 10 20 30 40

-40 -30 -20 -10 0-21-10

∆rcon (%)40S20SEq20N

90W 30W60W90W 30W60W 40S20SEq20N

135462

Amazon, JJA

40S20SEq20N

40S20SEq20N


4xCO2∆rcon (%)pre ipitation hange (%)

∆δ

18O

v

(h)∆

δ18O

v

(h)

∆rcon (%)Supplementary material 41/31

Impa t on response to temperature

1

2

3

ET

P

ET+R

200 250 150 100


ET(mm)

soil moisture (mm)Supplementary material 42/31


1

2

3

ET

P

ET+R

E is more sensitivethan T totemperature

200 250 150 100


+4K

ET(mm)



1

2

3

ET

incr

ease

sm

ore

stro

ngly

ET

P

ET+R

E is more sensitivethan T totemperature

200 250 150 100


+4K

ET(mm)


climate - accueil lmdcrlmd/these/seminaire_civilengineering_27april2… · climate changes...

Documents