understanding hydrometeorology using global …...climate is both global and local • need coupled...
TRANSCRIPT
Understanding Hydrometeorology using global models
• Alan K. BettsAtmospheric Research, VT
[email protected] 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