Land surface parameterization

busyicicleMechanics

Feb 22, 2014 (3 years and 6 months ago)

98 views

Land surface in climate models

Land surface parameterization
schemes in climate models

Bart van den Hurk

(KNMI/IMAU)

Land surface in climate models

The global energy budget

Trenberth, 2009

Land surface in climate models

The global hydrological cycle

Peixoto & Oort, 1992

residence time



in atmosphere: ~10 days



in ocean: ~3000 yrs



land: ~ 1
-
5 yrs


Land surface in climate models

The global carbon budget

IPCC, 2007

Land surface in climate models

General setup of General Circulation
Models (1)


What determines the evolution of the atmosphere?


Motion


Equation of motion


U, V = f (pressure gradient, friction)


Temperature


Conservation of energy:


T = f (thermodynamics, radiation)


Moisture content


Conservation of mass


q = f (evaporation, condensation)

Land surface in climate models

General setup of General Circulation
Models (2)


Basic equations are solved on a
grid


Computational constraints


Typically 100
-
500 km horizontal


1


1
°
: 65,000 surface points


Typically 20
-
50 vertical layers


1


5
million

grid points


Numerical constraints


Numerical stability limits time step of
integraton


10


60 minutes/time step


One year = 10
5

time steps, one century = 10
7

Land surface in climate models

Land treatment in GCMs

General setup of General Circulation
Models (3)


Land surface in climate models

General setup of General Circulation
Models (3)


Many processes are sub
-
grid, and need to be
parameterized


Fine scale processes (fluxes) expressed in terms of
resolved variables (mean state) using (semi
-
)
empirical, observation based equations


Example: turbulent sensible heat flux




a
s
H
p
UC
c
H






s


a

H = Sensible heat flux [W/m
2
]



= air density [kg/m
3
]

c
p

= specific heat [J/kg K]

U = wind speed [m/s]

C
H

= exchange coefficient [
-
]


s

-


a

= temperature gradient [K]


s


a

H

Land surface in climate models

Parameterizations in GCMs


Examples


Radiation


Condensation/cloud formation


Convection


Turbulent mixing


Land surface processes

Land surface in climate models

Landprocesses in atmospheric models


Energy
-
budget


Albedo


Surface


Albedo

Dark forest


9
-
12%

Grassland


15
-
20%

Bare soil


20
-
30%

Snow in forest

15
-
25%

Open snow


50
-
85%


Land surface in climate models

Landprocesses in atmospheric models


Energy
-
budget


Albedo


Evaporative fraction


Q*

H

LE

G

Surface



LE/Q*

Boreal forest



25%

Forest in temperate climate


65%

Dry vineyard



20%

Irrigated field in dry area


100%


Land surface in climate models

Landprocesses in atmospheric models


Energy
-
budget


Albedo


Evaporative fraction


Water budget


Runoff
-
fraction


P

LE

Infiltration

Direct runoff

Drainage

Land surface in climate models

Landprocesses in atmospheric models


Energy
-
budget


Albedo


Evaporative fraction


Water budget


Runoff
-
fraction


Soil water reservoir

Season

Shallow

rootzone

Deep

rootzone

Land surface in climate models

Landprocesses in atmospheric models


Energy
-
budget


Albedo


Evaporative fraction


Water budget


Runoff
-
fraction


Soil water reservoir


Carbon budget

CO
2

H
2
O

Land surface in climate models

Fluxnet data analysis


Fluxnet
: collection of ground stations worldwide
over various surface types

Teuling et al, 2010

Land surface in climate models

General form of land surface schemes


Energy balance equation


K

(1


a
) +
L




L


+

E

+
H

=
G



Water balance equation



W
/

t

=
P



E


R
s



D



Q*

H

LE

G

P

E

Infiltration

R
s

D

Land surface in climate models

General form of land surface schemes


Energy balance equation


K

(1


a
) +
L




L


+

E

+
H

=
G



Water balance equation



W
/

t

=
P



E


R
s



D



Coupled via the evaporation

Q*

H

LE

G

P

E

Infiltration

R
s

D

Land surface in climate models

Development history of land schemes


Late 1960’s: bucket scheme (Manabe, 1969) with
depth of the reservoir = 15cm

P

E

Direct runoff

E = (W/W
max
) E
pot

R = 0

(W<W
max
)

R = P


LE

(W

W
max
)

Land surface in climate models

Development history of land schemes


Mid 1970’s: explicit treatment of vegetation
(Penman
-
Monteith ‘big leaf’)









To be combined with submodel for soil
infiltration/runoff

P

E

Direct runoff







a
c
a
p
r
r
D
r
c
G
Q
LE
/
/
*








Land surface in climate models

First Soil
-
Vegetation
-
Atmosphere Scheme
(SVAT)


Deardorff (1978) combined


Penman
-
Monteith


Partial vegetation coverage, but
still one energy balance equation
(lumped surface types)


‘effective’ surface resistance
(interpolating between canopy
value for full vegetation, and
large value for bare ground)

Land surface in climate models

First Soil
-
Vegetation
-
Atmosphere Scheme
(SVAT)


Deardorff (1978) combined


Interception of snow and precipitation by leaves


(small) bucket equation







Prognostic equation for soil temperature and
moisture (‘force restore’)

dW
l
/dt = P


E


(W
l

< W
lmax
)

dW
l
/dt = 0


(W
l



W
lmax
)

W
lmax

= c LAI


(c = 0.2 mm)

E = E
pot

I = P


E


dW
l
/dt

dW
1
/dt = (C
1
/z
1
) I


C
2

(W


W
equ
)/


dW/dt = I/z
2

Land surface in climate models

Explicit multi
-
component SVATs


Separate treatment of vegetation and
understory/bare ground (Shuttleworth et al,
1988)


canopy resistance


evap. resistance for bare ground


Complex rewriting of PM, involving


separate net radiation for two
components


solution of T,q “within canopy” (at
network node)


separate aerodynamic coupling of two
components


Evaporation at bare ground affects canopy
transpiration and vice versa

Land surface in climate models

Tiled scheme


For instance ECMWF (2000)


Multiple fractions (“tiles”)


vegetation (transpiration)


bare ground (evaporation)


interception/skin reservoir
(pot. evaporation)


snow (sublimation)


Multi
-
layer soil


diffusion


gravity flow


Explicit root profile

Land surface in climate models

More on the canopy resistance


Active regulation of evaporation via
stomatal aperture



Two different approaches


Empirical (Jarvis
-
Stewart)


r
c

= (r
c,min
/LAI) f(K

) f(D) f(W) f(T)



(Semi)physiological, by modelling photosynthesis


A
n

=


f(W)

CO
2

/ r
c


A
n

= f(K

, CO
2
)


CO
2

= f(D)

Land surface in climate models

Summary of development


Soil hydrology


single bucket


two
-
layer force restore


multi
-
layer diffusion/gravity flow


Evaporation from surface


E = b E
pot


PM ‘big leaf’ (effective r
c
)


PM ‘multi
-
source’


Tiling


Canopy resistance


constant


empirical dependence on environment


photosynthesis
-
based

Land surface in climate models

Some other developments


Replace lat/lon grid by sub
-
catchment as spatial
unit (Koster et al, 1996)






Explicit parameterization of surface runoff (Dumenil
& Todini, 1992)

Infiltration curve

(dep on W and

orograpy)

Surface runoff

Land surface in climate models

Carbon allocation


Carbon allocation


distribution over leaf, stems, roots


decay and cycling through soil

GPP = Gross Primary
Production

NPP = Net Primary
Production

AR = Autotrophic
Respiration

HR = Heterotrophic
Respiration

C = Combustion




GPP

120

AR

60

HR

55

NPP

60

C

4

Land surface in climate models

Other biochemical processes


Nitrogen cycle







Land use change

http://www.visionlearning.com/library/module_viewer.php?mid=98

Land surface in climate models

International comparison/
evaluation experiments


Project for Intercomparison of Land
-
surface
Parameterization Schemes (
PILPS
; Pitman et al)


Observed atmospheric forcing


Comparison between partitioning of energy and
water


Single site (e.g. Cabauw)


2D catchments (e.g. Sweden)

Land surface in climate models

International comparison/ evaluation
experiments


Global Soil Wetness Project (
GSWP
)


Global 2D


Forcing from satellite, in situ and meteorological
(re)analysis data


Latest version: 10 yrs (GSWP2) + 3 yrs spin
-
up

Land surface in climate models

International comparison/ evaluation
experiments


Atmospheric Model Intercomparison Project
(
AMIP
)


Comparison of land surface processes in multiple
GCM’s


Forcing is not similar for all models

abs soil moisture content

soil moisture anomaly

Land surface in climate models

Orders of magnitude


Estimate the energy balance of a given surface type


What surface?


What annual cycle?


How much net radiation?


What is the Bowen ratio (H/LE)?


How much soil heat storage?


Is this the complete energy balance?


The same for the water balance


How much precipitation?


How much evaporation?


How much runoff?


How deep is the annual cycle of soil storage?


And the snow reservoir?

Land surface in climate models

More information


www.knmi.nl/~hurkvd


hurkvd@knmi.nl