ALARO Physics developments at LACE

meatcologneInternet and Web Development

Nov 3, 2013 (3 years and 10 months ago)

69 views

28th EWGLAM meeting, Zurich, October 2006

ALARO P
hysics

developments

at LACE



Neva Pris
tov


LACE Working group for physisc


28th EWGLAM meeting, Zurich, October 2006

ALARO physics package
-

introduction



continous transition from ARPEGE/ALADIN to
AROME (continuity + improvements)


to treat ‘grey
-
zone’ 3
-
7 km mesh size


economical computation, numerical efficiency


algorithmic flexibility


good basis for
further developments



28th EWGLAM meeting, Zurich, October 2006

Main developments topics


Radiation



Orographic forcing



Large scale precipitation



Prognostic turbulent scheme



Precipitating
convection

Luc Gerard
, J
-
M Piriou, D. Banciu,
I.Stiperski, R. Brožkova
,
J
-
F Geleyn

J
-
F Geleyn, G. Hello,
N. Pristov
, Y. Bouteloup,

M. Derkova
, J. Mašek, A.Trojakova
, R. Fournier

B. Carty, F. Bouyssel,
R. Brožkova
, J
-
F Geleyn,

M. Derkova, R. Mladek, J. Cedilnik, D. Drvar
, I. Beau

B. Carty, J
-
F Geleyn,
R. Brožkova,
J. Cedilnik, D. Drvar

F. Vana,

J. Cedilnik,
M
.
Tudor
,
J
-
F Geleyn
,
A.Simon

28th EWGLAM meeting, Zurich, October 2006

Radiation

Aim:


using the current delta
-
two stream approximation of
radiative transfer equation for solar and thermal bands


economical computation (a good quality cost ratio)


better consideration of clouds


New features:


new technique for thermal radiative fluxes computation on
the basis of Net Exchanged Rate (NER) formalism


gaseous transmition functions for computation of optical
depth closer to RRTM scheme


introduction of the complete aerosol model


updating of the cloud optical properties

28th EWGLAM meeting, Zurich, October 2006

Cloud optical properties


availability of prognostic cloud water calls for more
sophisticated treatment of clouds in radiative transfer


static treatment of saturation

effect

in old scheme
(
clouds
are
too opaque in some circumstances
)


new scheme:


coefficients k_abs, k_scat dependent on cloud water content


dynamic treatment of saturation, taking into account

cloud
depth and geometry


remain cheap,
only

two bands
:

solar and thermal

28th EWGLAM meeting, Zurich, October 2006

Cloud optical properties


strategy


Dependency of coefficients k_abs, k_scat on cloud water
content was

fitted using experimental sample of 7 liquid and 16
ice cloud types


Evaluation of
saturation effect in idealized framework:


multi
-
layer delta
-
two stream system


only clouds taken into account, gases and aerosols ignored


cloud geometry with random or maximum overlaps between
adjacent

layers


atmosphere illuminated from one side by direct flux (solar)

or
diffuse flux (thermal), reflected and transmitted fluxes evaluated


broadband fluxes diagnosed by summing up monochromatic
results computed for several hundred separate wavelengths

28th EWGLAM meeting, Zurich, October 2006

C
loud optical properties

red


liquid clouds, blue


ice clouds, black


mixed clouds

Saturation factor versus optical depth for homogeneous
clouds, solar band


k_absorbtion

k_scatter

28th EWGLAM meeting, Zurich, October 2006

C
loud optical properties

parameterized versus reference solar reflectance,
sample of non
-
homogeneous 3 layer clouds


old

new

28th EWGLAM meeting, Zurich, October 2006

C
loud optical properties

Solar absorbtion new (blue) versus old (red) scheme


Problematic peak


around model level 20


28th EWGLAM meeting, Zurich, October 2006


S
tatistical method for the weighting function

in NER


parameterized versus computed thermal flux


28th EWGLAM meeting, Zurich, October 2006

Pseudo
-
prognostic turbulent scheme pTKE



the turbulent memory of the previous timesteps
is kept


the advection and diffusion of TKE are added to
the current scheme


more general computation of mixing length


prognostic variable TKE

28th EWGLAM meeting, Zurich, October 2006

TKE vs pseudoTKE



=
advection


+
dif
f
usion


+ mechanical or shear production/destruction


+ buoyancy production/consumption


+ viscous dissipation

full



=
advection


+
dif
f
usion



+ Newtonian relaxation

towards stationary


solution

pseudo

28th EWGLAM meeting, Zurich, October 2006

Academic test with 1D model:

GABLS II experiment

28th EWGLAM meeting, Zurich, October 2006

Mixing length computation
-

current formulation



Empirical formula dependent on the height of the
PBL (Ayotte,

et al.

1996)



= f (z, h
PBL
)
-
exponential function

Increase of the mixing length

with the PBL height (h
PBL
)

Modifications:





1

,

2





=
f (h
PBL
)


28th EWGLAM meeting, Zurich, October 2006

Tests with the BL89 p
arameterization


Bougeault
-
Lacarr
ère (1989) : “non
-
local” mixing length
dependent on TKE and integral buoyancy (used e.g. in
ARPEGE and Mèso
-
NH)


Direct application of BL89 in the ps
e
udo
-
TKE scheme:
numerically stable but not consistent with the original K
-
theory part (too large mixing length


increase of
exchange coefficients and turbulent fluxes)


Consequences:


systematic increase of TKE and winds at surface,


too big diffusion in the upper troposphere, danger of
tropopause erosion

28th EWGLAM meeting, Zurich, October 2006

Merger with BL89 or TKE/N schemes


current mixing length (GC) taken as a first guess


additional information about TKE and buoyancy from the
BL89 scheme, modulation with k
-
parameter


more suitable for computation of K
-
coefficients


Merger with TKE/N scheme (only extensively tested, 1
-
D
evolution too noisy, not so good properties)

28th EWGLAM meeting, Zurich, October 2006

Academic test with 1D model:

GABLS II experiment



Vertical profile and time course of mixing length

Modified GC


GC

+BL89

PBL top

28th EWGLAM meeting, Zurich, October 2006

3
-
D case: strong jet stream

15 December 2005, Northern Germany

TKE and


crss
-
sectins

Modified GC


BL89

A

B

only small

TKE (no CAT)

B

A

impact in

jet
-
stream

area

higher TKE

in the PBL

28th EWGLAM meeting, Zurich, October 2006

3
-
D case: strong jet stream


GC+BL89: more feasible for PBL, TKE signals in jet
region (CAT) but not too violent


TKE/N: Too much TKE in upper troposphere

GC

+BL89

GC

+TKE/N

28th EWGLAM meeting, Zurich, October 2006

Large scale precipitation



a simple micro
-
physics scheme with 5 water
species included into precipitation scheme


cloud water, cloud ice,
rain
, s
now

-

new
prognostic variables


fluxes


statistical approach for sedimentation of
rain
and snow


28th EWGLAM meeting, Zurich, October 2006

ALARO
-
0 minus 3MT

better space distribution of precipitation

oper

alaro

6h precipitation

6h precipitation

28th EWGLAM meeting, Zurich, October 2006

ALARO
-
0 minus 3MT

Vertical profiles of horizontal averaged values


cloud water



cloud ice


28th EWGLAM meeting, Zurich, October 2006

Plans


Evaluation of parameterization developments


Technical and scientific validation


Case studies, verification



Training



Continuation of work


Convection (
3MT
)


More complex options