COSMO Status Report 2004 – 2005
Tiziana Paccagnella, ARPA-SIM
With the contribution of all the COSMO colleagues
For more dedailed information ….
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
COSMO Status Report 2004 – 2005
Tiziana Paccagnella, ARPA-SIM
1. Organizational Structure of COSMO
2. Lokal-Modell (LM) Overview
3. Operational Applications of the LM
4. COSMO scientific activities
5. Future developments
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Since 2004 the
Romanian Meteorological Service
is associated partner of COSMO
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Organizational Structure of COSMO
(3)
Rotach (MeteoSwiss)
Adrian
Papageorgiu
Raspanti
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Activities Planning and Monitoring
 STC meetings (2-3 times per year)
& SPM
 SAC (WP Coordinators + SPM)
& Chairman of STC:
~ May: monitoring and review of the activities
~ Sept. (beginning): final check of past activities and preliminary
planning
Several inter & intra working groups workshops
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
COSMO Status Report 2003 – 2004
Tiziana Paccagnella, ARPA-SMR
1. Organizational Structure of COSMO
2. Lokal-Modell (LM) Overview
3. Operational Applications of the LM
4. Activities of the COSMO WGs
5. Future developments
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Lokal-Modell (LM): Dynamics and Numerics
Dynamics
Basic equations:
Prognostic variables:
Diagnostic variables:
Coordinates:
Non-hydrostatic, fully compressible primitive equations;
no scale approximations;
advection form;
subtraction of a stratified dry base state at rest.
Horizontal and vertical Cartesian wind components, temperature, pressure
perturbation, specific humidity, cloud water content.
Options for additional prognostic variables: cloud ice, turbulent kinetic energy, rain,
snow and graupel content.
Total air density, precipitation fluxes of rain and snow.
Rotated geographical coordinates (λ,φ) and a generalized terrain
c
V
o
o
e
r
r
d
t
i
i
c
n
a
a
l
t
c
ς
e
o
o
-
f
o
l
l
o
w
i
n
g
.
r
d
i
n
a
t
e
s
y
s
t
e
m
o
p
t
i
o
n
s
:
 Hybrid reference pressure based σ
 Hybrid version of the Gal-Chen coordinate
 Hybrid version of the SLEVE coordinate (Schaer et al. 2002)
-
t
y
p
e
c
o
o
r
d
i
n
a
t
e
(
d
e
f
a
u
l
t
)
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Lokal-Modell (LM): Dynamics and Numerics
Numerics
Grid structure:
Arakawa C (horizontal)
Lorenz vertical staggering
Spatial
discretization:
Second order horizontal and vertical differencing
Time integration:
Numerical
smoothing:
Lateral
Boundaries:
Leapfrog (horizontally explicit, vertically implicit) time-split integration
including extension proposed by Skamarock and Klemp 1992. Additional
options for:
- a two time-level Runge-Kutta split-explicit scheme (Wicker and
Skamarock, 1998)
- a three time level 3-D semi-implicit scheme (Thomas et al., 2000)
- a two time level 3rd-order Runge-Kutta scheme (regular or TVD)
with various options for high-order spatial discretization ( Foerster
and Doms, 2004)
4th order linear horizontal diffusion with option for a monotonic version
including an orographic limiter (Doms, 2001);
Rayleigh-damping in upper layers;
3-d divergence damping and off-centering in split steps.
1-way nesting using the lateral boundary formulation according to Davies
(1976). Options for:
- boundary data defined on lateral frames only;
- periodic boundary conditions
Driving models: GME, IFS/ECMWF or LM.
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Lokal-Modell (LM): Physics
Physics
Grid-scale Clouds and
Precipitation:
Subgrid-scale Clouds:
Moist Convection:
Cloud water condensation /evaporation by saturation adjustment.
Cloud Ice scheme HYDCI (Doms,2002).
Further options:
- prognostic treatment of rain and snow (Gassman,2002; Baldauf
and Schulz, 2004, for the leapfrog integration scheme)
- a scheme including graupel content as prognostic variable
- the previous HYDOR scheme: precipitation formation by a bulk
parameterization including water vapour, cloud water, rain and
snow (rain and snow treated diagnostically by assuming column
equilibrium)
- a warm rain scheme following Kessler
Subgrib-scale cloudiness based on relative humidity and height. A
corresponding cloud water content is also interpreted.
Mass-flux convection scheme (Tiedtke) with closure based on
moisture convergence.
Further options:
- a modified closure based on CAPE within the Tiedtke scheme
- The Kain-Fritsch convection scheme
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Lokal-Modell (LM): Physics
Physics
Vertical Diffusion:
Surface Layer:
Soil Processes:
Diagnostic K-closure at hierarchy level 2 by default. Optional:
- a new level 2.5 scheme with prognostic treatment of turbulent kinetic
energy; effects of subgrid-scale condensation and evaporation are
included and the impact from subgrid-scale thermal circulations is
taken into account.
Constant flux layer parameterization based on the Louis (1979) scheme
(default).
Further options:
- a new surface scheme including a laminar-turbulent roughness sublayer
Two-layer soil model) including snow and interception storage; climate
values are prescribed as lower boundary conditions; Penman-Monteith plant
transpitration. Optional:
- a new multi-layer soil model including melting and freezing
(Schrodin and Heise, 2001
δ
-
Radiation:
l
o
t
n
w
o
g
w
s
a
t
v
r
e
e
a
m
f
l
r
u
x
a
e
d
s
i
;
a
t
f
u
i
o
l
n
l
s
c
l
c
o
h
u
e
d
m
-
e
r
a
a
d
f
i
a
t
e
t
i
r
o
R
n
i
f
t
e
t
e
e
r
d
a
b
n
a
d
c
G
e
l
e
y
n
(
1
9
9
2
)
f
o
r
s
h
o
r
t
a
n
d
k
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Lokal-Modell (LM)
Initial Conditions:
Interpolated from GME, IFS/ECMWF or LM.
Nudging analysis scheme (see Table 2).
Diabatic or adiabatic digital filtering initialization (DFI) scheme (Lynch et
al., 1997).
Boundary Conditions:
One way nesting by Davies boundary relaxation scheme;
Physiographical data Sets:
Mean orography derived from the GTOPO30 data set(30"x30") from
USGS.
Prevailing soil type from the DSM data set (5'x5')of FAO.
Land fraction, vegetation cover, root depth and leaf area index from the
CORINE data set.
Roughness length derived from the GTOPO30 and CORINE data sets.
Code:
Standard Fortran-90 constructs.
Parallelization by horizontal domain decomposition.
Use of the MPI library for message passing on distributed memory
machines.
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Lokal-Modell (LM): data assimilation
Data Assimilation for LM
Method:
Implementation:
Realization:
Balance
Nudging coefficient
Analyzed variables
Spatial analysis
Nudging towards observations
Continuous cycle
Identical analysis increments used during 6 advection
time steps
1. hydrostatic temperature increments (up to 400 hPa)
balancing
'near-surface' pressure analysis increments
2. geostrophic wind increments balancing 'near-surface'
pressure analysis increments
3. upper-air pressure increments balancing total
analysis increments hydrostatically
6.10-4 s-1 for all analyzed variables except pressure
1.2 10-3 s-1 for “near-surface” pressure
horizontal wind vector, potential temperature, relative
humidity, 'near-surface' pressure (i.e. at the lowest
model level)
Data are analyzed vertically first, and then spread
laterally along horizontal surfaces.Vertical weighting:
approximately Gaussian in log(p); horizontal
weighting: isotropic as function of distance.
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
1. Organizational Structure of COSMO
2. Lokal-Modell (LM) Overview
3. Operational Applications of the LM
4. Activities of the COSMO WGs
5. Future developments
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Operational Application of LM
DWD Germany
LM
7 km
LME (Europe)
7 km
MeteoSwiss
aLMo
7 km
LAMI
7 km
EURO LM
7 km
Italy
HNMS
Lokal Model-HNMS 7 km
IMGW
Lokal Model-IMGW 14km
NMA
Lokal Model-NMA 14 km
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Table 2: COSMO major specifications
LM
configuration
s
ARPA-SMR
DWD
HNMS
IMGW
MeteoSwiss
UGM
LEPS
Grid Points
234 x 272
325 x 325
95 x 113
193 x 161
385 x 325
465 x 385
306 x 258
Hor. Res.
0.0625o
(7 km)
0.0625o
(7 km)
0.125o
(14 km)
0.125o
(14 km)
0.0625o
(7 km)
0.0625o
(7 km)
0.09o
(10 km)
N° Layers
35
35
35
35
45
35
32
Time Step
40 sec
40 sec
80 sec
80 sec
40 sec
40 sec
60 sec
Frcst Range
72 hrs
48 hrs
48 hrs
72 hrs
72 hrs
60 hrs
120 hrs
Start Time
00, 12 UTC
00, 12, 18 UTC
00 UTC
00, 12 UTC
00, 12 UTC
12 UTC
12 UTC
Lateral BCs
GME /1 hr
GME /1 hr
GME /1 hr
GME /1 hr
IFS/3 hrs
IFS/3 hrs
IFS-EPS/ 6 hrs
1 hr
1 hr
1 hr
3 hr
3 hr
6 hr
LBC Upd.
Freq.
Initial State
Nudging
Nudging
GME
GME
Nudging
IFS
IFS-EPS
Initialis.
None
None
None
None
None
None
None
External
Analysis
None
SST, Snow
Depth, Soil
Moisture
Analysis
None
None
Snow Depth
from DWD
None
None
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Table 2: COSMO major specifications
LM
configuration
s
ARPA-SMR
DWD
HNMS
IMGW
MeteoSwiss
UGM
LEPS
Dynamics
3 TL; splitexplicit
3 TL; splitexplicit
3 TL; splitexplicit
3 TL; splitexplicit
3 TL; splitexplicit
3 TL; splitexplicit
3 TL; splitexplicit
Physics
TKE; new
surface
layer
scheme
TKE; new
surface layer
scheme
old Turbulence
scheme
TKE; new
surface layer
scheme
old
Turbulence
scheme
TKE; new
surface
layer
scheme
TKE; new
surface layer
scheme
Hardware
IBM SP
pwr4
IBM SP pwr3
CONVEX
SGI Origin 3800
NEC SX-5
IBM SP
pwr4
IBM SP pwr4
No. of
Processors
used
32
(of 512)
165
(of 1920)
14
(of 16)
88
(of 100)
14
(of 16)
112
(of 1024)
84
(of 1024)
Skip ►
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
LM
at DWD
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
LME
at DWD
 Model Configuration
Grid spacing: 0.0625° (~ 7 km)
665 x 657 grid points per layer
40 vertical layers; lowest at 10m
Timestep: 40 sec
Daily runs at 00, 12, 18 UTC, +78h
 Boundary Conditions
Interpolated GME forecasts with
 Data Assimilation
Nudging analysis scheme
Variational soil moisture analysis
SST analysis at 00 UTC
Snow depth analysis every 6 hrs
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Lokal Modell
aLMo at
MeteoSwiss
3.9
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
LAMI
Italy
3.9
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
EURO LM
Italy
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Lokal Modell
at HNMS
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Lokal Modell
at IMGW
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Lokal Modell
at NMA
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
LMK (LM Kürzestfrist)
According to COSMO strategy, activities will be maily addressed to
LM application at the meso-γ (Cloud Resolving Scale)
LMK, started as a DWD project, is now included in the COSMO activities
LMK Goals
 Development of a model-based NWP system for very short
range forecasts (2-18 h) of severe weather events on the
meso- scale, especially those related to

deep moist convection
(super- and multi-cell thunderstorms, squall-lines, MCCs,
rainbands,...)
 interactions with fine-scale topography
(severe downslope winds, Föhn-storms, flash floodings, fog, ...)
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
LMK (LM Kürzestfrist) - Development of a very short
range forecast model
M. Baldauf, A. Seifert, J. Förstner, T. Reinhardt, C.-J. Lenz, P. Prohl,
K. Stephan, S. Klink, C. Schraff
LMK-Configuration
• grid length: dx = 2.8 km
• direct simulation of the coarser parts of
deep convection (!?)
• interactions with fine scale topography
• about 50 vertical layers,
lowest layer in 22 m (new: 10 m) above ground
• center of the domain 10° E, 50° N
• 421 x 461 grid points
• boundary values from LM (LME)
(x = 7 km)
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
COSMO-LEPS
at ECMWF
More details later…
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
COSMO Status Report 2003 – 2004
Tiziana Paccagnella, ARPA-SMR
1. Organizational Structure of COSMO
2. Lokal-Modell (LM) Overview
3. Operational Applications of the LM
4. Some activities of the COSMO
WGs
5. Future developments
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Activities of the past year
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Some Highlights on activities related to Assimilation
from the Overview by Christoph Schraff
at the COSMO General Meeting 2005
Latent Heat Nudging
main problems related to prognostic precipitation solved to a large degree
Multi-Sensor Humidity Analysis (incl. GPS-obs)
GPS tomography; operational production of hourly humidity profiles over Switzerland
Use of 1dVar Satellite Retrievals both for MSG and ATOVS data
Good example of efficient cooperation among Germany, Italy and Poland
Assimilation of Scatterometer Wind
 First tests with LM planned for 12/05
Evaluation / Monitoring / Tuning of Nudging
Soil Moisture Initialisation
ELDAS project finished- Soil moisture initialisation for multi-layer soil model implemented in
LME with good results.
Snow Cover Analysis
DWD developed a version to incorporate snow mask into current snow analysis scheme
(snow mask corrects result of snow analysis). Cloud mask product and its impact needs to be
evaluated for longer periods.
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Some outocomes from the COSMO 1-day workshop to
discuss the COSMO strategy about Data Assimilation”
from the Overview by Christoph Schraff
at the COSMO General Meeting 2005
Long-term vision : how will / shall the NWP system look like in 2015 ?
 PDFs: deliver not only deterministic forecasts, but a representation of
the PDF (ensemble members with probabilities), particularly for the
convective scale
 resolution: global: 10 km , fine-mesh: x = 1 km
1-day workshop to discuss
 use of indirect observations at high frequency even more important
 high-frequency update (DA + FC)
the COSMO strategy
 emphasis on ensemble techniques (FC + DA)
about Data Assimilation
 due to special conditions in convective scale (non-Gaussian pdf, balance
flow-dependent and not well known, high non-linearity), DA split up into:
 generalised DA for global + regional scale modelling ( variational DA)
 separate DA for convective scale
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Activities of the past year
Advanced numerics seminar within the
LM-User Seminar 6 to 8 March 2006
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Highlights on some activities related to Numerics
from the overview by Juergen Steppeler
at the COSMO General Meeting 2005
LM-Z
Expected advantages:
 The atmosphere at rest can be represented in Zcoordinates, but not in terrain following coordinates
 Stratiform clouds and low stratus are predicted
better in LM-Z
 Mountain and valley winds are better with LM-z
 Precipitation amplitudes should be better with LMZ, in particular maxima and minima near
mountains
j - 1/2
j + 1/2
i - 1/2
i + 1/2
j + 1/2
i, j
i - 1/2
i + 1/2
Present status:
 Realistic cases possible (Euler Version)
 SL version now 3-d
 Low stratus cases seem to have a problem after last modification of
physics interface
 New QPF cases were run
 The QPF is improved significantly in all 4 available cases (1 Poland, 3
Greece)
j - 1/2
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Highlights on some on activities related to Numerics
from the overview by Juergen Steppeler
at the COSMO General Meeting 2005
Runge Kutta in LM: present status
 RK exists in LM and is considered to be
suitable for operational use
 It offers approximation order 3 which should
be sufficient for practical purposes
 The scheme is a modification (shortcut) of the
WRF scheme. Too little work concerning a
deeper evaluation has been done (such as
the investigation of conservation properties,
physics interface).
Skip ►
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Highlights on some activities related to Numerics
from the overview by Juergen Steppeler
at the COSMO General Meeting 2005
Implicit Time Integration
 Implicit integration is the basis for increasing
efficiency by enlarging the timestep
 Hydrostatic models reach savings of about a
factor of 2
 Non-Hydrostatic so far did not realise a
similar saving
 By working on the implicit solution procedures
LM could reach an efficiency gain of a factor
of 2
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Highlights on some activities related to Numerics
from the overview by Juergen Steppeler
at the COSMO General Meeting 2005
Implicit Conclusions
 A direct si- method was proposed
 The method is based on a generalised
Fourier Transform
 The generalised FT is potentially as efficient
as the FT (fast FT)
 1-d and 2-d tests have been performed
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Activities of the past year
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Highlights on some activities related to Physics
from the overview by Marco Arpagaus
at the COSMO General Meeting 2005
Surface transfer scheme and turbulence scheme

Single column model developed

Extended documentation almost ready
Soil processes

Multi-layer soil model

Operational implementation of the multi-layer soil model

Usage of satellite derived (weekly updated) plant cover

Revision of external parameters for plants

Implementation of lake model into the LM

Intercomparison of soil models in the framework of soil moisture validation
Shallow convection on the meso-γ scale

some tests with LMK
Microphysics

Three-category ice scheme included
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Activities of the past year
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
The COSMO-LEPS suite @ ECMWF
since June 2004
d
d-1
d+1
d+2
middle EPS
00
d+3
d+4
d+5
4 variables
ZUVQ
3 levels
500 700 850 hPa
Cluster Analysis and RM identification
youngest EPS
12
2
time
steps
European
area
Complete
Linkage

COSMOLEPS
clustering
area
COSMOLEPS
Integratio
n Domain




suite running every day at
ECMWF managed by ARPASIM;
Δx ~ 10 km; 32 ML;
LM_3.15 since 18/4/2005;
fc length: 132h;
Computer time provided by
the COSMO partners which
are ECMWF member states.
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Operational set-up
Core products:
 10 perturbed LM runs (ICs and 6-hourly BCs from 10 EPS members) to
generate probabilistic output (start at 12UTC; t = 132h);
Additional products:
 1 deterministic run (ICs and 3-hourly BCs from the high-resolution
deterministic ECMWF forecast) to assess the relative merits between
deterministic and probabilistic approach (start at 12UTC; t = 132h);
 1 proxy run (ICs and 3-hourly BCs from ECMWF analyses) to “downscale”
ECMWF information (start at 00UTC; t = 36h).
From 1 July 2005, COSMO-LEPS forecasts are archived on MARS at
ECMWF (class=co).
Skip ►
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Archiving of C0SMO-LEPS products
From 1 July 2005, COSMO-LEPS forecasts are archived on
MARS at ECMWF (class=co).
Deterministic run (fc+0h to fc+132h every 3h).
Ensemble Prediction System:
10 perturbed forecasts (fc+0h to fc+132h every 3h):
PLEV (500, 700, 850 hPa): Z, RH, T.
SURF: albedo, LCC, MCC, TCC, SW radiation flux, CAPE, hzerocl, snowlmt,
mslp, T_2m, Td_2m, TMAX_2m, TMIN_2m, U_10m, V_10m, UVMAX_10m,
large-scale rain, convective rain, large-scale snow, TP.
Forecast probability (various intervals and thresholds):
SURF: CAPE, hzerocl, TMAX_2m, TMIN_2m, UVMAX_10m, TP, snowfall,
showalter index.
Clustering information (population, clustering variable used, …).
back archiving of past runs (from 5/11/2002 onwards) will start soon.
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
COSMO–LEPS verification
(SON 2004)
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Highlights on some activities related to Applications
from the overview by Pierre Eckert
at the COSMO General Meeting 2005
ENSEMBLE
COSMO-LEPS
The impact of moist singular vectors and ensemble size on predicted storm tracks
for the winter storms Lothar and Martin
Evaluation of the COSMO-LEPS forecasts for the floods in Switzerland in August
2005 and further recent results…
Use of Multi-Model Super-Ensemble Technique for complex orography weather forecast
Multimodel Ensemble and SuperEnsemble permit a strong improvement of all the
considered variables with respect to direct model output.
First experiments toward the new SREPS system.
Perturbation oh Physics. One case done (not completely)
Interpretation of the new high-resolution model LMK
Using information of a single model forecast by applying the Neighbourhood Method
Using information resulting from LMK forecasts that are started every 3 h (LAFEnsemble)
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Activities of the past year
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Summary of verification results
strong negative bias of 2m-temperature during winter
 strong overestimation of precipitation, especially during
winter
 strong overestimation af small wind gusts
 general overestimation of wind speed for stations over flat
terrain (especially during night)
 strong phase shift of maximum in the daily cycle of
precipitation during summer time (maximum occurs too early
in the forecasts)

COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Summary of verification results
 overestimation of high cloud covers
 overestimation of amplitude of dew point
(especially during winter)
 negative temperature bias in the boundary layer
and during winter in the whole model atmosphere,
positive temperature bias in the middle troposphere
during summer.
 MSLP negative bias, especially during summer for
GME driven LM’s
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Joint WG3-WG5 workshop (9.3.2005)
Conditional verification (excerpt from the minutes):
 WG3 provides a list of criteria to use for
conditional verification. [done]
 WG5 is invited to provide a tool (‘data finding
tool’) that allows to easily and efficiently find
days / grid-points which satisfy specific criteria.
 WG5 is kindly asked to supply a conditional
verification of the operational model suites on a
regular basis.
 Testing of the reference version should also
include conditional verifications.
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Activities of the past year
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
LM: Changes in the last year
from the overview by Ulrich Schaettler
at the COSMO General Meeting 2005
9 LM Version 3.13:





Changes to the new multi-layer soil
model
Introduction of the Graupel scheme
Introduction of a 3D turbulence scheme
(only for Runge-Kutta scheme)
Introduction of the Latent Heat Nudging
Introduction of a second option for the
Rayleigh damping
 LM Version 3.14:

03.12.04
9 LM Version 3.16:





22.07.05
Update of Runge-Kutta scheme
Modification to prognostic TKE scheme
Update of multi-layer soil model
New external parameters for evergreen and
deciduous forest
And of course: Bug corrections…
 . Next
Reference Version: LM 4.1
25.01.05
Mainly bug corrections and updates to the new
schemes
 LM Version 3.15
03.03.05

Aerosol bug fix

Technical changes for LME

Introduction of Namelist variable nunit_of_time
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
COSMO Status Report 2003 – 2004
Tiziana Paccagnella, ARPA-SMR
1. Organizational Structure of COSMO
2. Lokal-Modell (LM) Overview
3. Operational Applications of the LM
4. Activities of the COSMO WGs
5. Future developments
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
COSMO Longer-term perspectives (3-5
years)
1. Develop and improve ensemble prediction systems based on
LM (including cloud resolving scale)
2. Assimilation of ‘all’ available observations
 sfc / profiles / radar / satellites…
 high resolution
 also slowly-changing information (soil moisture,
vegetation state, etc)
3. LM performance is known, critically discussed and improved
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Planning of 2005-2006 activities (1)
Starting this year the work plan will be structured in:
Priority Projects:
•Proposed by SAC
(WG Coordinators and SPM)
•Discussed/completed/modi
fied…. with COSMO
scientists (September GCM)
Activities:
Not structured
as projects and
carried on if
extra-resorces
are available
•Approved by STC which is
also responsible of
allocating the required
human resources
(December)
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
The new structure of the work plan
To improve the efficiency of the COSMO Cooperation (trying to
minimize criticities of past years)
The COSMO work plan will be structured in:
Priority Projects
Activities
 Projects have been thought:

To better concentrate joint
COSMO efforts
 To optimize the use of
COSMO human resources
 To facilitate management
related to goals which are
Considered to be prioritary
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Selected Priority Projects
Support Activities:
includes many of the activities of WG6 Implementation and reference
version (Source codes managment, documentation, editing of newsletters,
tech. Reports, WEB site)
SIR: Sequential Importance Resampling filter
Aim:
Motivation:
Prerequisites:
Development of a prototype for a new data assimilation system
for the convective scale by means of a SIR filter
Method avoids problems related to strong non-linearity, unknown
flow-dependent balance, and non-Gaussian probability densities
present in the convective scale.
Approach provides the initial conditions for the convective-scale
ensemble forecasting.
Availability of a realistic larger scale ensemble with very short-range
spread (LM SREPS) for providing the lateral boundary conditions.
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Selected Priority Projects
Retrievas for nudging: 1dVar for satellite radiances
Aim: Operational assimilation of ATOVS and SEVIRI data for LM
Motivation: Use of satellite radiances especially over data-sparse areas over
sea has the potential to provide essential information to improve forecasts.
Note: Memory management for 1D-Var and flow of processing facilitates
inclusion of further satellite instruments in the future.
1.Further development of LM_Z
Aim: Have LM_Z ready for operation in 5 years.
2006 aim: Improve the physics interface; Extensive case study; demonstrate the
advantage of LM_Z; Considerations of numerical efficiency should be addressed
in the first year.
Motivation: LM_Z avoids the violation of an approximation condition. This has
the effect that without this coordinate artificial circulations are created near
mountains, with potentially unrealistic results in respect of mountain valley
circulations, low stratus and precipitation. Problems with steep orography are
expected to be more severe with cloud resolving resolution.
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Selected Priority Projects
1.Further development of the Runge Kutta method
Aim:
create an improved and properly checked Version of RK in the timescale of 3
years.
Motivation:
The RK method is implemented in LM and will be used for high resolution
applications with LMK and possibly replace the Klemp Wilhelmson (KW) scheme
later. It offers a substantial gain in accuracy at no additional cost. Aim:
2006 plans:
Continue validation of RK; 3rd order for vertical; Diagnosis of conservation
properties; Advection of scalar quantities in conservation form; Looking at
pressure bias; Investigation of convergence; Deep valleys; Different filter
options for orography;
Physics coupling scheme (WG3)
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Selected Priority Projects
1.UTCS: Towards Unified Turbulence-Shallow Convection Scheme
2.Tackle deficiencies in precipitation forecasts
3.Development of Short range ensemble
4.Advanced Interpretation of Lm output
5.Implementation of an Extended Common Verification Suite - Conditional
Verification.
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Some outocomes from the COSMO 1-day workshop to
discuss the COSMO strategy about Data Assimilation”
from the Overview by Christoph Schraff
at the COSMO General Meeting 2005
Consensus:
• Ensemble DA should play a major role
• Nudging at moment:
 robust and efficient
 requires retrievals for use of indirect observations, but no severe
drawbacks if we can make them available
 not foreseeable when / whether nudging will cease, but we keep on
reviewing the situation
 can use it for fallback in SIR
Strategy:
• Start development of SIR (for the longer-term, with option to include
nudging)
• further develop nudging, in particular retrieval techniques (for mid-term
+ fallback)
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Towards Unified Turbulence-Shallow Convection
Scheme (UTCS)
Implementation of Transport Equations for the Sub-Grid Scalar Variance,
Coupling with Convection, Turbulence and Cloud Diagnostic Schemes
Motivation
There are many deficiencies of LM that are related to inadequate representation of
boundary-layer turbulence and shallow convection. Among them are



prediction of stratiform clouds,
triggering and diurnal cycle of deep convection,
prediction of partial cloud cover and cloudiness-radiation interaction,
to mention a few. These deficiencies manifest themselves in large forecast errors,
first of all, of precipitation rate and timing and of 2m temperature. The proposed
project is aimed at addressing the above issues in a unified turbulence-shallow
convection framework.
Goals

To account for turbulence and shallow non-precipitating convection in a unified
framework (hopefully, for LMK and similar models with mesh-size of 2-3 km this is a
‘final’ solution, although convection-turbulence modelling with such a grid size is still
terra incognita).
 To achieve a better coupling between boundary layer turbulence, convection and
radiation (for LM and similar models with the mesh-size 7-10 km; may require
adaptation of the convection scheme).
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Tackle deficiencies in precipitation forecasts
Motivation
There are various indications, both from verification (WG5) and from
customers (e.g., forecasters) in various COSMO countries, that the LM has
serious deficiencies in forecasting precipitation. Some of these problems are
longstanding (and not necessarily unique to the LM), some others are fairly
recent.
Aim
This project aims at collecting, highlighting and investigating some of these
deficiencies, and putting forward possible improvements. The sensitivity of
precipitation systems under various situations and in different countries will
be investigated based on a standardized set of model changes (physics,
numerics, data assimilation). As such, it will need resources from members
of many working groups, most notably WGs 1, 2, 3, and 5.
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Development of short range ensemble (SREPS)
Motivation
Short
range ensembles have been developed in several countries like
the USA, Canada, Japan, the UK, Norway, France…. Improvements,
especially in situations of high impact weather have been noticed.
COSMO already has experience in ensemble forecasting, but mainly as
a downscaling of the ECMWF ensemble in the (early) medium range.
Perturbations have for the moment been included though the lateral
boundaries, they are not present in the initial state. Perturbation of the
physics has “only” been introduced by running two different convection
schemes.
Link with the SIR project
Aim
Develop of a COSMO SREPS system at around 7-10 km h.res.
Strong cooperation with INM :
LM will be added to the INM-SREPS regional models
INM will provide LM runs to provide boundary values to COSMO
SREPS
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Advanced interpretation of LM outputs
Motivation
 Models usually show various types of systematic errors. These errors can be
removed locally by the use of statistical methods like MOS, Kalman filters, non
linear training. These errors are expected to decrease in the future due to the
improvement of the models. The diagnostics of high impact weather is
however often left to the appreciation of the forecasters. Objective guidance in
decision making could be helpful to them.
 The foreseen increase in resolution of the models will lead to a proliferation of
grid points and probably also to an increase of the noise in the forecasts. The
effects of the double penalty will increase for events not predicted exactly at the
right place at the right time.
Tasks
 Explore more grid point statistics, also on other parameters than precipitation
 New classification and regression algorithms for detection of high impact weather.
These include the use of complex predictors like instability indices and non linear
classification algorithms like neural networks.
 Generation of weather types (thunderstorm, drizzle, fog,…)
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Implementation of an Extended Common Verification Suite Conditional Verification
 Outcome
 Deliver
a SW package, a “library”, acting as tool
to perform verification of LMs using specific
criteria that have to be satisfied in carrying out
the scores, in the more flexible way
 The
package will have to be a part of the
common COSMO software as the Common
Conditional Verification suite
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Projects will be presented in their final form
before the 1st of November to be approved by
the STC during the December meeting.
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia
Thank You!
COSMO Report – Tiziana Paccagnella - SRNWP/EWGLAM meeting 3-5 October 2005 Ljubljana Slovenia