Understanding Hydrometeorology using global models

• Alan K. BettsAtmospheric Research, VT

akbetts@aol.comRevised version of AMS-2004 Robert E. Horton Lecture

ftp://members.aol.com/bettspapers/HortonBAMS1848.pdf

•Thanks to NSF, NASA and NOAA for support•Special thanks to Anton Beljaars, Pedro Viterbo and Martin Miller at ECMWF for two decades of extensive collaboration -- and to John Ball for 15 years of patient data processing

Preamble

• Not a review talk• Title is meant to be a paradox• Simple models for understanding?

Hydrometeorology is too complex• Climate interactions of water

[phase changes and radiation interactions] are central to climate

• Let us confront the challenge

Climate is both global and local

• Need coupled earth system models • Need them locally to warn us of the first frost

[local diurnal cycle in September]• Improving our global models is central• Global models can be used as tools to understand

interacting processes• Contrast our model world, which we dimly

understand, and the real world, where we only understand fragments of a complex, living system.

What controls evapotranspiration?

• “Equilibrium evaporation”. Raupach (BLM, 2000, QJRMS 2001)

• Models for the growing daytime “dry BL”• Fascinating but simplified by ignoring some

key real-world physics, which control evaporation for climate equilibrium.

What is this ignored physics?• Cloud fields control cloud base, the surface net

radiation, and dominate the cooling rate of the CBL[It is not the dry BL solutions that are relevant]

• Climate problem is a 24-hr mean problem, with a superimposed diurnal cycle

[It is not just a growing daytime BL problem]

• First-order atmospheric constraints on evaporation. Global models with coupled cloud fields include these processes, so they can help us understand the coupling

Outlinea) Global scale feedbacks – seasonal forecasts

Idealized global soil moisture simulations and evaporation-precipitation feedback over continents

b) Land-surface coupling at daily timescale – 30 years of ERA40 river basin time-series

Coupling of soil moisture, cloud-base, cloud cover, radiation fields, sensible, latent heat; TCWV, precipitation and omega

a) Global scale feedbacks -Idealized soil moisture simulations and

evaporation-precipitation feedback

• Serendipity, and great flood on the Mississippi of July 1993

• Parallel ECMWF suite with a 4-layer soil model to better represent soil moisture memory

• Soil moisture sensitivity experiments for July, 1993

July 1993: wet-dry soil initialization

• Increase of monthly forecast precipitation: peaking at over 4 mm/day or >125 mm/month [Beljaars et al. 1996]

Seasonal forecasts with idealized soil moisture

• ERA40 model: 120-day forecasts at T-95 L60 from May 1, 1987 (DOY=121)

• Identical except a) Soil moisture initalized at 100% field

capacity for vegetated areasb) Soil moisture initalized at 25%

-- Soil Moisture Index 0 < SMI < 1 as PWP < SM < FC

P, E, P-E and SMI for Eastern US

• Reduction of SMI reduces precipitation, evaporation

• has little impact on P-E which averages to small values over summer

• Memory of soil moisture lasts all summer

Europe Amazon

Canada N. Asia

Monsoon India

Only in monsoon regions where P-E is large is memory of SMI reduced

Evaporation over land linked to precipitation: [away from monsoons]

• So what controls evaporation?• Not classic “equilibrium evaporation”• Recast equilibrium evaporation as as a

diurnally averaged problem, linked to cloud-base and cloud fields

[Betts, JHM 2000; Betts et al., JGR 2004, in press]

Surface energy balance, and ML “equilibrium”

• 3 Americas regions• 5-day means:

of wet and dry simulations

• Latent heat λE against SMI: weak relation: sensitive to Rnet

• Sensible heat H against SMI: tight relation

• linked to dependence of depth to cloud-base on SMI

Sensible heat flux: H

• H against PLCL : linear with slope related to cooling processes in ML• H is constrained by ML cooling, constrained by cloud-base• Net long-wave has similar behavior: coupled to PLCL

Amazon basin in more detail

• H , 8E quasi-linear with PLCL: 2-m Q and T quasi-linear with PLCL

• Over wetter soils, E increases; T decreases and Q increases in ML• New coupled state has lower LCL, with cooler, moister ML; reduced H

and larger E

Conclusions-1

• Climate and climate change over land depends critically on getting evaporation-precipitation feed-back right

• ERA-40 model has large E, P feedback over continents [Is it right?]

• The change in surface energy budget over dry and wet soils is consistent with a shift of the mean sub-cloud layer equilibrium

b) ERA40 river basin budgets

• Basin averages: hourly archive• Daily averages:1972-2002 [11000 days]

• Madeira : AmazonArkansas-Red : MississippiAthabasca : Mackenzie

• [ERA40 biases:see Betts et al. 2003a,b]

ERA40 for Madeira River basin compared with LBA Rondonia pasture site: 1999

• Large seasonal change of diurnal amplitude• ERA-40 basin ranges smaller than at pasture site

1960 1965 1970 1975 1980 1985 1990 1995 2000

-1

0

1

2

3

4

5

6

7

8

Year

P, E

, R (

mm

day

-1)

ERA-40 Amazon P12

P36

E12

E36

R12

CSM increment

∆CSM

ERA40 Annual means1957-2001

P: precipitationE: evaporationR: runoffCSM: Column soil water

1960 1965 1970 1975 1980 1985 1990 1995 2000

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

6.5

7.0

7.5

8.0

38

39

40

41

42

43

44

45

46

47

48

49ERA-40 Amazon

Year

P,

Bia

s +

5.7

(m

m d

ay-1

) TC

WV

analysis (kg m-2)

P: 0-12h FX Bias + 5.7 TCWV analysis

a

PP

PPP

P

PP P

PP PP

PP

PP

PP

PP

PPP

P

PPP P

PPP

PP

PPP

P P PP

PP P

RRRRRR RR R

RR RRR

R

RRRR

RR

RRR

RRR

R RRRR R

RR

RR

R R RR

RR R

BB

BBBB

BB B

BB B

BB

B

BB

BB

BB

BBB

BB

BBB

BBB

BBBBBB BB

BBB

B

38 39 40 41 42 43 44 45 46 47 48 49

-2

-1

0

1

2

3

4

5

6

7

8

TCWV analysis (kg m-2)

P, R

(m

m d

ay-1

)

P : 0-12h FX R : 0-12h FX B : Bias of P

Regression lines P= -12.6+.411*TCWV R= -12.1+.327*TCWV Bias=0.364*(TCWV-44.7)

b

Annual means

TCWV: Total column watervapor

P: precipitationP-bias from observations

Regression on TCWV

P: precipitationR: runoffP-bias

CC

CCC

CCC

C

CC

CC

CC

C

C

CCCC

CCC

CC

CCC

CC

CC

CCCCCCC

C

C

CC

CCCCCC

CCC

CC

CCC

CC

C

CCCC

CCC

CC

CCC

CCCC

CCCC

CCC

C

C

CC

1960 1965 1970 1975 1980 1985 1990 1995 2000

0

1

2

3

4

5

6

7

8

Year

P, R

(m

m d

ay-1

)

C P: corrected P (Marengo 2004) P (Dai et al.2004)

C R: corrected R (Marengo 2004)

ERA-40 AmazonERA40 Annual meanscorrected

Mean is ± 10%, but

Little signal ininterannual variability

Data for P uncertain

1111

1

111

1

111 11 1

1

11

1

11

1

1

1 11

11

1

11 1

1

11

111

11

1

1

11

1

111 1

11

1

1

11 1 1

1111

11

1

1

11 1

11

1

111

11

1

1

11

11

11 1

1

11

1

111

11

11 11

11

1

1111 1

1111

11

111

1 11

1

11

11 1

1

11

111

111 1

11

1

1111

1 1

1111

1

11

111

1

11

1

1 1

111

11 1

11

1

1 1

11

1

111

11

112

222

2

2222

2

222

2

22

2

22

2

22 2

222 22

2

2

2 2

22

22

22 22

2

2

22

2

2 2 2

2

22

2

2

2

22

2

222

22

22

2

22

2

22 2233

333

33 3

33

3

33

33

33

33

33

333

33

3

3

33

3

33 3

3 3

33

33

3

333

33

3 3 33 33

3

333 3

3 3

33 3 33

333 3

33

333

3333

333 3

33 3 3

33

33

333

3 3 33 333

3

33

333

333 3

3

333

333

3 3

33 3

33

333

333

33

33

3

33

33

3 33

3

3

3 33

3

3

33

33

33

33

3

3 33

33

33

3 3

33

333

3

33

333

3

33

3333

33

3333

33

33 3

33

3

33

33

333

3

3

3

3

33

3

3

333

333

3

3

333 3

3

3 3 3333

33

33

3

333

33

33

3

3

33

33

3

3333

3

33

33

33

33

33 3

3

33

33

33

3

3

11 11

1

111

1

11

111 11

111

1 1

11

11

1

11

111 1 1

11

111

11

11

1 11

1 1

1 1 1

11

1

1

1 11

1

1

11 1

111

11

1 1

1

1 11

11

1

1111

1

1

1

1

11 1

1

111

1

1

11

111

11111

1

1

1

11

11

111

111 1

11

11

11 1

11

1

1

1 11

11 11

1

1

111 1

1 1 1 1

111

1

111 11

1

1 11

11

1

1111

1 111 1

11

11

11

11

1

12

22

222

222 2 22

22

22

2

22

2 22 2

2

22 2

2222

22 2

2

2

22 2

2

2222

22

22

22

2

22

222

2 22

22

2

22

2222

2 2 2

23

3 3

3

33

33

33

3

3

33

33

333

3 3

33

3

3 3

3

33

333

3

33

33

33

3

333

3

33 3

3

33

333

33 3

33 3

33

3

33

3

3

3 3 3

3 3

3 33

33

33

333 3

3

3 33

3

333

33 3

3 33

33

33

33

33 3

333

3

33

3

33

333

3

33

3

3333

3333

33 3

3

333

33

3

33 3 3

3333

33

333 33 3

3

33

3

333

33 3 3 3

333

33

333

3 3

33 3

3333

33

33

333

3333 3

3

333

33 3

33

33

33

33

3

3

3 33 3

3 3

33333

33 3

3 3

33

33

333

33

33

3333

333

3 33 33

33

33

3

3333

33 3

3

33

33

333 3

33 3

30 35 40 45 50 55

-4

-2

0

2

4

6

8

10

12

TCWV (kg m2)

P, B

ias

from

dat

a (m

m d

ay-1

)

Amazon (monthly)

1 P:1958-1972 2 P:1972-1978 3 P:1979-2001 Fit from Fig 3b

Bias from data

Monthly precipitation andP-bias from observations against TCWV

P against TCWV:[similar to annual]

Bias against dataERA40 iswet in dry seasondry in wet season[if P-data is correct]

1960 1965 1970 1975 1980 1985 1990 1995 2000

23.5

24.0

24.5

25.0

25.5

26.0

26.5

Year

T (

o C)

Dai et al. 2004 ERA40

AmazonAnnual mean

a

1960 1965 1970 1975 1980 1985 1990 1995 2000

23.5

24.0

24.5

25.0

25.5

26.0

7

6

5

4

3

Year

T,

Tbi

as+

24.7

(o C

) P: 0-12h F

X (m

m day

-1)

ERA40 Amazon

T

Tbias+24.7

P: 0-12hFX

b

5.1 5.2 5.3 5.4 5.5 5.6 5.7 5.8 5.9 6.0 6.1 6.2 6.3

24.0

24.2

24.4

24.6

24.8

25.0

25.2

T (T-24.67)=1.126*(P-5.73)

AmazonDai et al. 2004

P: data (mm day-1)

T: d

ata

(o C)

c

2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5

23.5

24.0

24.5

25.0

25.5

26.0

P: 0-12h FX (mm day-1)

T: E

RA

40 (

o C)

T: ERA40 (T-24.5)=0.644*(P-5.1)

Amazon ERA40

d

T control by precipitation

Coupling of soil moisture index, cloud-base height and Evaporative fraction

• Mean cloud-base height increases over drier soils and with larger surface Rnet

• Evaporative fraction increases with soil moisture, and decreases with Rnet

• 3 basins similar: with additional dependence on unstressed resistance

Madeira basin for July and November

• July: dry season• Nov: wet season

• Surface fluxes as function of cloud-base and cloud cover

LWnet dependencies

• Soil moisture index• Cloud-base• Total cloud cover• Diurnal range: Ts

• 2 months merge to single quasi-linear distribution

SWnet dependencies

• Tight coupling to LWnet

• Cloud-base• Total cloud cover• Sensible heat flux H

• Distinct distributions except for H

Sensible heat flux H• Diurnal range: Ts• Maximum Ts• Cloud-base• SWnet

• Distinct distributions except where coupled to SWnet

• Subcloud heating rates• 3K/day in July• 6K/day in November

Latent heat flux λE and H• Coupling of H to SMI

through PLCL stronger than coupling of λE

• λE has more variation with Rnet in rainy season

• H splits into 2 branches as function of Rnet[contrast SWnet]

LW coupling for other basins

• LWnet tightly coupled to cloud cover and cloud-base• Madeira has 50hPa lower cloud-base• Red-Arkansas has 0.25 lower cloud cover

0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2

0

20

40

60

80

100

120

140

160

0.5

0.6

0.7

0.8

0.9

1.0

1.1

1.2

1.3Madeira River1990-2001

Soil moisture index

PLC

L (hP

a)

EF

PLCL EF

a

0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 1.2

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

-70

-60

-50

-40

-30

-20

-10

Soil moisture index

LCC

LWnet (W

m-2)

LCC

LWnet

b

Surface-coupled physics

Daily Means (Madeira River)Soil moisture index against Soil moisture index againstPLCL: LCL of cloud-base LCC: Low cloud coverEF: Evaporative fraction LWnet : surface net longwave

-100 -80 -60 -40 -20 0 20 40 60

20

25

30

35

40

45

50

55

60

0.00.10.20.30.40.50.60.70.80.91.0

Ωmid

(hPa day-1)

TC

WV

(kg

m-2

)

TC

C

TCWV TCC

a

Madeira River1990-2001

Ascent Descent

20 25 30 35 40 45 50 55

0

5

10

15

20

25

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

TCWV (kg m2)

P (

mm

day

-1)

TC

C

P TCC 0.03*(TVWV-20.2)

b

Tropospheric-coupled physics

Daily meansmid: mid-tropospheric omega TCWV againstagainst TCWV and TCC Precipitation and TCC

[Linear regression]

-100 -50 0 50

0

5

10

15

20

Madeira River1990-2001

P

E

0.11*(Ωmid- 40)

Ωmid

(hPa day-1)

P, E

(m

m d

ay-1

)

a

Descent

Ascent

-5 0 5 10 15 20

0

5

10

15

20

b

TCWV*(Ωmid

-42)/420 (mm day-1)

P (

mm

day

-1)

Madeira River1990-2001

P 1 : 1 line

Coupling of ascent with precipitation

Daily meansP and E against “Moist circulation”mid: mid-tropospheric omega and precipitationNote P0 at mid 40hPa/day

Conclusions-2• Model data such as reanalyses can be used to

understand coupling of processes• Coupling of surface processes in ERA-40, though

complex, is comprehensible.• Soil moisture, cloud-base, cloud cover, the

radiation fields and evaporative fraction are coupled quite tightly [sub-seasonally]

• Mid-tropospheric omega, TCWV and TCC coupled Mid-tropospheric omega and precipitation closely linked on daily time-scale

Conclusions-3

• Evaporation is controlled somewhat indirectly by the controls on net radiation and sensible heat flux

• The long-wave flux control by cloud-base height and cloud cover is particularly tight across all basins

• The sensible heat flux is coupled to cloud-base height, cooling processes in the sub-cloud layer, as well as directly to the shortwave flux [in ERA-40]

Conclusions-4• Proposing a framework for analyzing model data for

land-surface feedbacks• Proposing analysis framework for comparing global

models and climate observations

• RH, cloud-base and cloud cover need to be measured with the radiation fields as climate variables

• Climate modeling with interchangeable plug-in modules is fraught with peril, as the feedbacks change

Comparisons with data

ERA-40 ‘point’ Harvard forest tower

SW-cloud coupling to PLCL-Total cloud cover: ERA40

-Transmitted fraction SW

-Transmitted fraction PAR

-compare LW coupling