GCSS all 52005new ECMWF PBL schemeKhler

1
The new ECMWF PBL scheme and beyondunifying dry
BLs, stratocumulus and shallow convectionMartin
Köhler Roel Neggers

• approach moist combined Mass-flux/K-diffusion
  (MK)
• stratocumulus single M-flux,
  bottom and top K
• shallow convection dry/cloudy M-flux, bottom K
• thanks to Pier Siebesma, Joao Teixeira, Adrian
  Tompkins, Anton Beljaars

2
current PBL parameterization frameworks
moist mass-flux convection

• ECMWF (2004)
• (also NCAR, NCEP, Arpege, DWD, )
• Teixeira Cheinet (2003)
• diffusion only
• Siebesma, Neggers, Cheinet (1998-2004)
• mass flux only

dry diffusion PBL
cloud base
(figures Cheinet, 2004)
3
unification an MK approach

• key ingredients
• moist conserved variables
• combined Mass-flux/K-diffusion solver
• cloud variability
• transition between stratocumulus and shallow
  convection

4
stratocumulus LES DYCOMS II
simulation by Bjorn Stevens (dx35m, dz5m)
vertical velocity cloud
5
stratocumulus aircraft observations DYCOMS II
17.5
?vC
17.0
10
qvg/kg
9
2
wm/s
0
-2
x km
30 60
courtesy of Bjorn Stevens
6
shallow convection LES BOMEX
w m/s
(dxdydz40m)
vertical velocity cloud water
7
shallow convection aircraft observations
Florida SCMS
8
two box M/K decomposition(Siebesma and Cuijpers,
1995)
a
M
M-flux
env. flux
sub-core flux
9
BOMEX LES decomposition
M-flux
sub-core flux
total flux
env. flux
M-flux covers 80 of flux
Siebesma Cuijpers, 1995
10
M-flux and updraft properties initialization
PDF
(Holtslag Moeng, 1991)
a5
wsfc
0

• parcel excess
  (assumes correlation of w, T, q)
• M-flux
• K reduction due to M-term

11
parameterization choices
Mass-flux
K-diffusion

• updaft model
• entrainment , t500s, c0.55
• detrainment 310-4 m-1 in cloud
• parcel determines PBL depth (wup 0)
• mass flux

• diffusion
• surface and cloud top driven diffusion
• cloud top entrainment

cloud variability
12
Results Low cloud cover (new-old)
T511 time10d n140 2001 2004
old CY28R4
new PBL
13
Results EPIC column extracted from 3D forecasts
Liquid Water Path g/m2
WV Mixing Ratio g/kg
300
45
250
200
Model Levels
Pressure hPa
150
50
100
50
55
0
60
6
10
0
2
4
8
16
17
18
19
20
21
22
Local Time October 2001 days
14
unification with shallow convection

• key extension
• multiple updrafts

experimental
next
done
15
A statistical mass flux framework for organized
updrafts
PDF of w, qt ,ql
Top of updrafts that are explicitly modelled
Dry updraft
K diffusion
Moist updraft
16
BOMEX SCM vs. LES
updraft w m/s
updraft e 1/m
LES dry updraft LES moist updraft SCM dry
updraft SCM moist updraft SCM dry updraft SCM
moist updraft
17
moist updraft qt excess
dry updraft qt excess
g/kg
g/kg
18
mass flux m/s
diffusion coef. m2/s
wqt m/s
K
19
cloud fraction
Mean state properties
CY28R4
Added vertical transport by the dry updraft is
necessary to maintain humidity in the top half of
the subcloud layer
MK PBL
normalized qt /qt0
CY28R4
Skewness can be used to create a decreasing cloud
fraction with height
20
Concluding Thoughts

• moist combined Mass-flux/K-diffusion scheme
  successful in dry BL and
  stratocumulus
• now being extended for shallow convection
• stratocumulus
• K-diffusion appropriate to produce well-mixed
  layer
• but M-flux adds potential of counter gradient
  fluxes
• downdrafts initiated by CTEI might further
  improve fluxes
• shallow convection
• M-flux appropriate for moist convection (80)
• 2 intrinsic discretizations
• horizontal cloudy updraft vs. environment
• vertical below vs. above cloud base
• natural framework moist M1 and dry M2 K

21
future

• microphysics
• momentum flux
• downdrafts

22
Liquid water potential temperature excess Some
text
23
Mean state properties
24
stratocumulus to trade cumulus transitioncriteria

• cloud thickness
• cloud top entrainment instability
• static stability ?700hPa- ?sfc lt 20K
• buoyancy flux integral ratio N/P gt 0.1

EPIC, Oct 2001
Chile
25
stratocumulus to trade cumulus transitionimpact
on low cloud cover
(newPBL old model)
static stability ?700hPa- ?sfclt 20K
buoyancy flux integral ratio gt 0.1
26
improved continental winter stratus (Hungary Dec.
2004)
METEOSAT visibile 10 Dec 2004
LCC impact of newPBL 8-16 Dec 2004
27
Results seasonal forecasting
T2m newPBL-old model jja1987-2001
NINO4 mean absolute error
new MK PBL
68 cases 3 members
T2m old model-ERA40 jja1987-2001
old model
200 years of sims!