Formation

1
Status of ARPEGE / ALADIN physics F.Bouyssel Mété
o-France
12-13/12/2005
Mesoscale physics and diagnostics tools
2
Presentation

• Operational physics
• Known problems
• Current developments

3
Operational physics
4
ARPEGE (global) and ALADIN (LAM) models

• Numerical weather prediction

ARPEGE Global spectral model with
stretched resolution 23 to 130 km, 41 levels (1
hPa), 982s, 4 runs/day, 4D-Var (T107, T149) 6h
assimilation cycle
Others operational configurations ARPEGE
Tropics and ALADIN models

• Climat simulations

Same dynamics H, 2SLSI
ARPEGE-Climat 250 km ALADIN-Climat 50
and 20 km
5
Physics in ARPEGE/ALADIN PNT-CLIMAT
SURFACE ISBA (Noilhan and Planton 89, Mahfouf
and Noilhan 96, Giard and Bazile, 2000) 2/4
layers force restore scheme (T,W,Wg), bare ground
and vegetation fraction, 1 layer snow scheme with
pronostic albedo RADIATION (2 schemes) 1)
ACRANEB (Geleyn and Hollingsworth 79, Ritter and
Geleyn 92) every time step 2) FMR15 (Fouquart
and Bonnel 80, Morcrette 93) every 1h or
3h Tegen aerosols, UGAMP ozone
climatology Randow maximum overlap
assumption MOUNTAIN DRAG linear and  form 
drag , anisotropy, resonance,  lift 
effect important modifications (Geleyn et al.,
Catry et al. submitted)
6
Physics in ARPEGE/ALADIN PNT-CLIMAT
TURBULENT VERTICAL DIFFUSION 2 schemes 1) 1st
order K diffusion scheme (Louis et al. 81)
modifs in stable case interactive mixing
lengths, shallow convection (Geleyn 87),
anti-fibrillation (Bénard et al. 00) 2) TKE
diagnostic (Mellor/ Yamada 2.0)  moist  
( F0/F1/F2  de Bougeault) LARGE SCALE
PRECIPITATION AND CLOUDS 2 diagnostic
schemes 1) revised Kessler type (evaporation,
melting, freezing), cloud condensates diagnosed
with RHc and buoyancy, cloudiness with XuRandall
(96) formulation 2) Ricard-Royer statistical
scheme SUBGRID PRECIPITATION AND CLOUDS Mass
flux scheme with a Kuo-type closure (Bougeault
85), vertical transport of horizontal momentum,
vertically varying detrainment and entrainments
rates, downdrafts (Ducrocq and Bougeault 95), etc.
7
Known problems
8
Known problems
Diagnostic schemes for precipitation and clouds
(no microphyics) lack of more physical
relations between clouds and precipitations Ove
restimation of small rainfall events ( lt 3
mm/day) particularly in the Tropics but also in
mid-latitudes (more sophisticated microphysics,
triggering of convection, ...)
9
Known problems
Overestimation of precipitation over mountains
(envelop orography, diagnostic RR schemes,
moisture convergence)
ALADIN
ARPEGE
RR error using raingauges (JAS 2005)
10
Known problems
Underestimation of low level clouds (turbulence
with cloud top entrainment, microphysics, shallow
convection)
TOP SW MODEL CERES (DJF)
11
Known problems

• Diurnal cycle of convection
• Problem of transition between shallow and deep
  convection
• Lack of moist convection in case of no synoptic
  forcing (moisture convergence only)
• Top of convection at neutral level

  (Guichard et al., 2004)
12
Known problems
Lack of sensitivity to humidity environment
Q2
25, 50, 70 or 90 relative humidity
MESO-NH
ARPEGE
  (Piriou, 2005)
13
Current developments
14
Current developments

• All problems presented before are common to all
  ARPEGE/ALADIN - PNT or Climat models
• Convergence of NWP and Climat ARPEGE/ALADIN
  physics towards more sophisticated physical
  parametrisarions for resolutions above 10km with
  a maximum of synergy with finer scale physics
  when possible!

15
Original scheme Lopez, 2001
16
Cloud statistical scheme with triangular PDF
Smith, 90 Dependency of s with Dx and
h Separation of qc in ql, qi function of
temperature Kessler type autoconversion
(threshold and rate funtion of T) Specification
of hydrometeor distributions for coll., evap.,
subl. Concentration N(D) N0exp(-lD)
(Marshall-Palmer) Mass M(D)aDb Fall velocity
V(D)gDd Precipitation sedimentation
semi-lagrangien conservative schema Constant
fall speed for snow and rain (0.9 and 5 m/s)
17
Case study 09/12/2004
18
Austrian alpin case (9 december 2000)
Oper
New
Observations (from ZAMG)
  (Haiden and Wittman)
19
oper
Lopez
Impact of advection
Lopez no advection of qr
Lopez no advection of ql, qi and qr
20
Operational at ECMWF and used in MESO-NH and
AROME Rayonnement thermique RRTM (Mlawer et
al., 97)(Morcrette et al.,98) Rayonnement solaire
(Fouquart and Bonnel, 80)(Morcrette,93) Maximum
randow overlap assumption Cloud optical
properties liquid water cloud SW / Fouquart
(87), LW / Smith-Shi (92) ice water cloud SW /
Ebert-Curry (92), LW / Smith-Shi (92) Coud
effective radius liquid water cloud Martin
et al. (94) ice water cloud Sun et al. (01)
21
- Pronostic microphysics RRTM/SW6 radiation
vertical diffusion of cloud conservative
variables (qt, sl) - Remove biais of RS more
EARS data and SATOB data - New climatological
files Improvment in ARPEGE and ALADIN forecasts
on cloud, precipitation and wind
fields BUT Degradation of geopotentiel scores
after 60h forecast over Europe in assimilation
mode but not in forecast one! Problem of
microphysics tunings linked with stretched
grid? Degradation of linear tangent
approximation? Problem with obs modifications
linked with bias?
22
DJF - AMIP (forced) Mean Sea Level Pressure
(hPa)
OPER
NEW
(Déqué)
23

• Sea and ocean
• prescribed SST, Charnock formula
• will soon include a better bulk formulation
• reflexion to implement a 1D oceanic mixing layer

Lakes prescribed temperature, Charnock formula
Vegetation and soil ISBA (Interface Soil
Biosphere Atmosphere)
Town TEB (Town Energy Balance)
24
Soil options Force restore, 2 layers , temp,
water, ice Force restore, 3 layers , temp,
water, ice Diffusion, N layers ,
temp, water, ice Vegetation options Noilhan
and Planton 89 (Jarvis) AGS (photsynthesis and
CO2 exchanges) AGS and iteractive
vegetation Hydrology options no subgrid
process subgrid runoff, subgrid drainage Snow
options Douville 95 (1 layer, varying albedo,
varying density ) Boone and Etchevers 2000
(3 layers, albedo, density, liquid water in
snow pack)

Tiling 1 to 12 patches
25
TKE-l Cuxart et al (2000)
Moist variables
Redelsperger and Sommeria( 81), Bougeault (82)
and Bechtold (93)
26
Diffusion
with
Mixing length
Bougeault and Lacarrère (89)
now
In Hirlam
with
27
2 works

• Modifications around current scheme using CAPE
  for CVNP and CVP (Gueremy)
• Proposition of equations  MT-CCF  for
  developing convective parametrizations
• Move the parametrization effort from detrainment
  towards microphysics.
• Continuity CV dry ? CVNP ? CVP
• Easier validation with CRM
• (Gerard, Piriou, Geleyn, Stiperski)

28

ARPEGE oper
CRM MNH
ARPEGE new
(Piriou, 05)
29

ARPEGE oper
CRM MNH
ARPEGE new
(Piriou, 05)
30
Low clouds (1000-680hPa) mean JJA EXP - ISCCP
ARPEGE climat reference
ARPEGE climat TKE pron. (CBR) with top PBL
entraînement Microphys pron. CVPPCVP (CAPE)
(Marquet Gueremy)
31

• Common physics for gt10km (PNT and Climat)
• Lot of common points with finer scale physics
  (surface, radiation, turbulence, convection ?).
  Potentially important for coupling.
• Need common phys/dyn interface and diagnostics
  for large collaboration (cross validation, etc
  ...)