Date post: | 11-Jan-2016 |
Category: |
Documents |
Upload: | anthony-mccormick |
View: | 215 times |
Download: | 0 times |
Convection mini-workshop, Hamburg, July 14, 2004
Parameterization of convection
Andreas ChlondDepartment Climate Processes
*Thanks to F. Nober, P. Bechthold, A. Tompkins and the ECMWF
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
Outline
• Introduction Why do we have to parameterize? What is a parameterization?
• How to parameterize? Classical approach Super-parameterization Parameterization of intermediate complexity
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
Why do we have to parameterize?
• Climate und regional models predict coarse grained, large-scale variables
• Small scales processes are not resolved by large-scale models, because they are sub-grid scale
• The effect of the the sub-grid processes on the large-scale has to be presented statistically
• The procedure of expressing the effect of sub-grid processes is called a parameterization
Turbulence, convection, clouds, radiation etc.
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
What is a parameterization?•The statistical contribution of sub-grid processes
must, therefore, be expressed in terms of the large-scale parameters themselves. The mathematical procedure involved is generally called a parameterization.
SUBROUTINE vdiff (kidia,kfdia,klon,klp2,ktdia,klev,klev, & & paclcm,paphm1,papm1,pgeom1,pum1,pvm1,pxm1)
! Description:!-- Computation of the exchange coefficientsIMPLICIT NONE! Scalar arguments with intent(In):INTEGER, INTENT (IN) :: kfdia, kidia, klev, klevm1, klevp1,& & klp2, ktdia, ktrac
DO jl = kidia, kfdia zdu2 = MAX(zepdu2,pum1(jl,klev)**2+pvm1(jl,klev)**2) zqmitte = (pqm1(jl,klev)+zqs(jl)*zhsoil(jl))/2.zmult4 = zfux*zmult5 - 1. zcons = zcons12*paphm1(jl,klevp1)/(ptm1(jl,klev)* & & (1.+vtmpc1*pqm1(jl,klev)-pxm1(jl,klev))) END DO
From Nature
To its
Representation
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
What is convection doing?
•Convection is of crucial importance for the global energy and water balance (Convection transports heat, water vapor, momentum … and chemical constituents upwards …. Water vapor then condenses and falls out -> net convective heating/drying
•Convection generates and/or influences a number of phenomena important to forecasting (thunderstorms, heavy precipitation, hurricanes)
•An important parameter for the strength of convection is CAPE
•Convection affects the atmosphere through condensation / evaporation and eddy transports
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
Convection parameterization – How?
Conventional approach:
•Start from first principles and write down exact equations for the process under consideration
•Close equations (Closure problem)
Introduce additional (empirical) information
Calibrate constants (observations, PRMs)
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
Task of convection parametrisation(1) total Q1 and Q2
To calculate the collective effects of an ensemble of convective clouds in a model column as a function of grid-scale variablesRecall: these effects are represented by Q1-QR, Q2 and Q3
Hence: parametrization needs to describe CONVECTIVE CONTRIBUTIONS to Q1/Q2: condensation/evaporation and transport terms and their vertical distribution.
p
secLQQQ RC
)(11
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
Task of convection parametrisation (2):in practice this means:
Determine vertical distribution of heating, moistening and momentum changes
Cloud model
Determine the overall amount of the energy conversion, convective precipitation=heat release
Closure
Determine occurrence/localisation of convection
Trigger
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
Types of convection schemes
• Schemes based on moisture budgets–Kuo, 1965, 1974, J. Atmos. Sci.
• Adjustment schemes–moist convective adjustement, Manabe, 1965, Mon. Wea. Rev.
–penetrative adjustment scheme, Betts and Miller, 1986, Quart. J. Roy. Met. Soc., Betts-Miller-Janic
• Mass-flux schemes (bulk+spectral)–entraining plume - spectral model, Arakawa and Schubert, 1974, J. Atmos. Sci.
–Entraining/detraining plume - bulk model, e.g., Bougeault, 1985, Mon. Wea. Rev., Tiedtke, 1989, Mon. Wea. Rev., Gregory and Rowntree, 1990, Mon. Wea . Rev., Kain and Fritsch, 1990, J. Atmos. Sci., Donner , 1993, J. Atmos. Sci., Bechtold et al 2001, Quart. J. Roy. Met. Soc.
–episodic mixing, Emanuel, 1991, J. Atmos. Sci.
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
The “Kuo” scheme
Closure: Convective activity is linked to large-scale moisture convergence
dzt
qbP
ls
0
)1(
Main problem: here convection is assumed to consume water and not energy
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
Adjustment schemes
e.g. Betts and Miller, 1986, QJRMS:
When atmosphere is unstable to parcel lifted from PBL and there is a deep moist layer - adjust state back to reference profile over some time-scale, i.e.,
t
q ref
conv
.TT
t
T ref
conv
.
Tref is constructed from moist adiabat from cloud base but no universal reference profiles for q exist. However, scheme is robust and produces “smooth” fields.
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
The bulk mass-flux approach
p
secLQ C
)(1
Aim: Look for a simple expression of the eddy transport term
Condensation term Eddy transport term
?
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
The bulk mass-flux approach:Cloud – Environment decomposition
Total Area: A
Cumulus area: a
Fractional coverage with cumulus elements:
A
a
Define area average:
ec 1
ec
The bulk mass-flux approach
Simplifications :
•Neglect subplume correlations
•The small area approximation:
ec ;1)1(1
ecc *Define convective mass-flux:
**
cc
c wg
M
Then: c
cgM
Then:
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
The bulk mass-flux approach
With the above we can rewrite:
p
ssMgecLQ
cc
C
)(
)(1
p
qqMLgecLQ
cc
)(
)(2
To predict the influence of convection on the large-scale with this approach we now need to describe the convective mass-flux, the values of the thermodynamic (and momentum) variables inside the convective elements and the condensation/evaporation term. This requires, as usual, a cloud model and a closure to determine the absolute (scaled) value of the mass-flux.
A bulk mass flux scheme:What needs to be considered
Entrainment/Detrainment
Downdraughts
Link to cloud parameterization
Cloud base mass flux - Closure
Type of convection shallow/deep
Where does convection occur
Generation and fallout of precipitation
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
Outline
• Introduction Why do we have to parameterize? What is a parameterization?
• How to parameterize? Classical approach Super-parameterization Parameterization of intermediate complexity
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
Convection parameterization – How?
•Run a 2D CSRM as a “super-parameterization” in a GCM
•This idea was first suggested by W. Grabowski of NCAR
Super-Parameterization:
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
Grabowski’s approach
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
What do we get?
•Explicit deep convection
•Explicit fractional cloudiness
•Explicit cloud overlap and possible 3d cloud effects
•Convectively generated gravity waves
But
•A GCM using a super-parameterization is three orders of magnitude more expensive than a GCM that uses conventional parameterizations. On the other hand super-parameterizations provide a way to utilize more processors for a given GCM resolution
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
Outline
• Introduction Why do we have to parameterize? What is a parameterization?
• How to parameterize? Classical approach Super-parameterization Parameterization of intermediate complexity
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
• An ensemble of potentially different clouds has to be described by ONE AVERAGED cloud.
• Dynamical and microphysical details are not represented
• Due to the mass-flux approach the information about cloud cover and vertical velocity is not available.
Problems with mass-flux schemes
Convection parameterization – How?
Parameterization of intermediate complexity: The ensemble approach
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
LES vs. SCM: Diurnal variation of cloud cover (ARM case)
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
Cumulus-Prey-Predator Model (1)
Philosophy (adapted from population dynamics):
• Different possible clouds correspond to different
species. These species are in competition for an external food supply.
• In case of clouds this food is given by CAPE (Convective Available Potential Energy)
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
1
2
K_22
K_12 K_21
F_1
F_2
K_11
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
Cumulus-Prey- Predator Model (2)
Mathematical: Search for the cloud ensemble (consistingof the possible clouds) that is most effective in consumingCAPE.
The solution of the Lotka – Volterra – Equation does it!
The interaction between the different cloud types and the clouds and the non-convective processes is determined by the energetic of the system.
Cloud – Field – Model
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
ECHAM GCMInput data: vertical profiles of temperature and humidity
Cloud – Field - Model
Cloud Model
Calculation of interactioncoefficients
Cloud Spectrum
Output data: heating ratestransport, precipitation
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
Comparison: Cloud – Field – Model, LES and ECHAM5 (1)
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
Comparison: Cloud – Field – Model, LES and ECHAM5 (1)
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
We should discuss and must decide …
•Classical approach
Simple, but to take conventional approach much beyond were we are now, it seems likely that we will have to make the parameterizations very complicated
•Super-Parameterization
More expensive, but SPs can use thousands of processors with good computational efficiency.
•Parameterization of intermediate complexity
Good compromise
Convection mini-workshop, Hamburg, July 14, 2004
Andreas Chlond: Parameterization of convection
Heating Rates ECHAM (zonal mean, DJF) (K/day)Convection Stratiform Clouds
Vertical Diffusion Radiation