PowerPoint Presentation

Report as inappropriate Your report has been received

Shared by: HC120831021944

Tags

Stats

views:: 1
posted:: 8/30/2012
language:: Unknown
pages:: 54

Public Domain

Document Sample

9th GIFS TIGGE WG

Geneva
31 Aug – 2 Sept 2011
TIGGE LAM Report
Tiziana Paccagnella
with contributions from
J. Kacker, J. Chen, C. Saulo
and other colleagues

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
Content

1. TIGGE LAM overall activities
a. New Structure of the Panel
b. The Plan
c. Archiving

2. Link with FDPs and RDPs & Link with the WWRP Mesoscale WG
a. Frost FDP RDP
b. Hymex

3. Activities in Europe
a. EurEPS Project (SRNWP)
b. LAM-EPS meeting & LAMEPS BC project
c. EPS DA meeting > Bologna
d. High resolution precipitation analyses (HIPRECA)
e. Link with SRNWP Interoperability Programme

4. Activities in NA by Josh Hacker
5. Activities in Asia by Jing Chen
6. Activities in SA by Celeste Saulo

7. TIGGE LAM @ ….

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
TIGGE LAM overall activities

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
Restructuring of the Panel

The Panel has been (and is being) restructured trying to make it
more efficient at Regional level.

New members have been asked to join the Panel to engage
people personally involved in LAM EPS and then in the
position to make plans and to take commitments.

The Panel has been divided in regional sub-groups:
N. America < J. Hacker
Asia < J. Chen
Europe < T. Paccagnella
S. America < C. Saulo
Africa < S. Landman
Australia/New Zealand

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
TIGGE LAM Plan

The TIGGE LAM plan
almost ready.

Some more work done
especially on the scientific
issues.

Tiziana and Josh will
revise it together during
the SRNWP meeting in
Tallin next month.

Plan on the web in
November

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
Archiving

Archiving of high priority parameters on regular lat/lon
grids.
 CMA is starting the archiving of TIGGE LAM HP
parameters from CMA Regional system.
 The archiving of European products at ECMWF has
been slowed down due to other priorities at ECMWF.

Archiving of TIGGE LAM products on their native grids
 LAM EPS are going to run at the convective-permitting
scale. Products on the native grids should be made
available and tools to manage these data should be
developed.
 As regards Europe, this task has been included into the
GEO-WOW FP7 < D. Richardson presentation

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
Link with FDPs and RDPs & Link with the
WWRP Mesoscale WG

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
Link with FDPs and RDPs
LAM EPS and Hymex

TTM1a
High resolution ensemble hydrometeorological
modelling for quantification of uncertainties
Leaders:
A. Montani V. Homar
ARPA-SIMC, Italy Universitat de les Illes Balears, Spain
amontani@arpa.emr.it victor.homar@uib.es

ARPA-SIMC (Italy)
ISAC CNR (Italy)
Météo-France (France)
UIB (Spain)
NRL (US) More about Hymex by L. Descamps
… others to join?…

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
TTM1a – main objectives

• LAM Ensemble Prediction Systems (EPS) with parameterized convection and typical horizontal
resolution of about 10 km, employed for short and medium-range forecasts (up to 5 days).
• Convection-Permitting Ensemble Prediction Systems (CPEPS) with explicit convection and
horizontal resolution of a few kilometres, employed for short-range predictions (up to 48 hours).
• Hydrological ensemble predictions.
• Calibration methods.
• Verification methods.

Interlinks with Data Assimilation.
On the one hand, data assimilation methods taking into account the flow-dependent instabilities need to
identify the growing perturbations at convective scales. On the other hand, the use of an ensemble system
for data assimilation purposes poses some constraints on the way the ensemble should be built.

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
Link with FDPs and RDPs
FROST-2014
Forecast and Research
in the Olympic Sochi Testbed

Tiziana !!!

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
FROST

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
FROST-2014
Forecast and Research in the Olympic Sochi Testbed
WG1: Observations and nowcasting (including Verification)
WG Leader: Arcadi Koldaev / Valery Lukyanov / Yuriy Melnichuk (radars)
Members: Arcadi Koldaev, Yury Melnichuk, Yong Wang, Paul Joe, George Isaac, Roy
Rasmussen, Dmitri Moisseev, Peter Romanov, Anatoly Muravev, IRAM
representatives, Gdaly Rivin, Evgeny Vasilev, Pertti Nurmi?
WG2: NWP, ensembles and assimilation (including Verification)
WG Leader: Michail Tsyrulnikov / , Andrea Montani
Members: Stephane Belair, Andrea Montani, Detlev Majewski, Gdaly Rivin, Yong
Wang, Roy Rasmussen, Tiziana Paccagnella, Donghi Wang, Anatoly Muravev, Michail
Tolstykh, Inna Rozinkina, Pertti Nurmi?
WG3: IT including graphical tools, formats, archiving and
telecommunication
WG Leader: Dmitry Kiktev
Members: Alex Kolker, Vladimir Krupchatnikov, Sergei Loubov, Gennady Novikov,
Alexander Smirnov
WG4: Products, training, end user assessment and social impacts
WG Leader: Valery Lukyanov / Evgeny Vasilev
Members: Vladimir Oganesian, George Isaac, Anatoly Muravev, Dmitry Moiseev,
Pertti Nurmi, Galina Zaimskikh, Inna Rozinkina, Pertti Nurmi.

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
FDP (Forecast Demonstration Project) ensemble
activities
SOCHMEL7: limited-area ensemble system based on
COSMO
(“relocation” of COSMO-LEPS)

Horizontal resoluzion: 7 km.
Vertical resolution: 40 model levels.
Forecast range: 72 hours.
Starting time: 00UTC, 12UTC.
Ensemble size: 10 members.
Boundary conditions: selected ECMWF EPS members.
Initial conditions: interpolated from ECMWF EPS members.

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
Link with WWRP MWFR WG

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
Activities in Europe

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
From slides by C.
SRNWP-EurEPS-2013-2017
Marsigli
Cooperation on European Limited-area Ensemble
Prediction Systems

 EurEPS - A proposal by the SRNWP Expert Team on
Predictability and EPS
 In the Roadmap for the Forecasting Capability Area of
EUMETNET the creation of a Eur-EPS Programme is
envisaged
 A major cooperative effort is required to develop a capability
for convection-permitting ensembles in order to address
prediction of severe or high-impact weather in a probabilistic
framework
 EurEPS
Phase I is proposed to be carried out in 2013, in order to identify
properly all the needed technical facility, the requirements for
Research and Development to design properly this innovative
kind of systems and the potential framework for running the
Phase II as a demonstration project
Phase II would be executed over a 4-year time frame (2014–2017)
as a demonstration project
tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
SRNWP LAM-EPS meeting, Bologna, 22-23 February 2011
Some Outcomes

 The EUMETNET Forecasting Roadmap has been presented by A. Horanyi:
Short-Range EPS is one of the 4 priority areas for cooperation which have
been identified

 Requirements to ECMWF in terms of EPS boundary conditions to drive
LAM-EPSs have been discussed.

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
LAMEPS BC project

• Possibilities being explored:
• 2 extra EPS runs at 06 and 18 UTC
• Higher resolution for the first 2 days
• Lower resolution up to 6 days
• T1279 up to 48 h, 21 members
• T639 up to 144 h, 51 members
• M. Leutbecher ECMWF performs some
test cases and data sets are available
• LAM-EPS interested partners will
perform tests of their systems driven by
these BCs within 6 months (until January
2012).
• A meeting will be organised at ECMWF
in Spring 2012, to evaluate the possible
tpaccagnella@arpa.emr.it establishment of such LAMEPS2 BC 2011
9th GIFS TIGGE WG Geneva 31 Aug – Sept
SRNWP LAM-EPS meeting, Bologna, 22-23 February 2011
Some Outcomes from the minutes

 It was decided to make an inventory of the status and plans of LAM-EPS
activities within SRNWP, based on the review provided by this meeting,
highlighting the scientific issues.
 This inventory would provide the basis to write a common paper, to be
submitted to BAMS, where the status of the European contribution to
TIGGE-LAM will be presented. Tiziana will contact the editor.

 After the meeting, the proposal to write and submit a Community Paper to
BAMS has been extended to all the LAM EPS community and has been
accepted
 The content and the structure of the paper should be defined within October

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
EPS DA meeting >
Bologna

Q1. Model error Q15 In building analysis ensembles, some centres generate pairs of
positive and negative perturbations or implement some other
Q2 Covariance filtering centering of analysis perturbations (with the intention to
Q3 Localisation improve the quality of the ensemble mean forecast). Other
centres seem to avoid such centering of initial perturbations
Q4 Enforcing large scale from host model (with the intention to simplify the scheme and to let the
Q5 Non-gaussianity and non-linearity perturbations explore more phase-space directions). Is there any
recent evidence in favour of either approach?
Q6 The ensemble system should mimic the DA system
Q16 How to compare different DA methods on the
Q7 Choice of data assimilation method (proposed by DA parallel convectionpermitting scale in a scientifically acceptable way
session) (duration, domain, resolution, physics of the models, etc.)?
Q8 Is it possible a “good” selection of a few relevant Q17 What diagnostic methods can the data assimilation community
global/driving EPS members? offer to assess the realism of stochastic parameterizations for
Q9 Can we really get rid of balance constraint? representing model uncertainty? (e.g.through the correctness of
the background error covariance matrix?)
Q10 scalability
A question (Q18) has been then posed by T. Paccagnella: is it
Q11 Are there any things we should not perturb (such as the possible to initialise a CP model with an analysis obtained by
orography)? computing the ensemble mean of an EnKF? Are there balance
Q12 Experiments in toy models suggest that the use of flow- problems?
dependent covariances gives much improved analyses. To date Q19 What are the key factors controlling the structure and
the benefits of using ensemble forecasts in NWP analyses have magnitude of analysis errors and background errors? How
been much smaller. What is the cause of this discrepancy? essential is a representation of model uncertainty for obtaining
Some have suggested model error - is there a way to prove if reliable flow-dependent estimates of background error
this is the cause? covariances?
Q13 Is it possible to formulate a weak-constraint EnKF, similar to Q20 Given the increasing complexity of forecast models (e.g.,
weak constraint 4D-Var? nonhydrostatic formulation, micro physics, ...), what are the
Q14 (Q7 of DA list) Is an ensemble based on singular vectors any consequences of this process for data assimilation algorithms?
use for DA? What about the one based on error-breeding? More specific, what is the future of 4d-var with its crucial role of
tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
linear physics?
HIPRECA: HIgh resolution PREcipitation Analyses for verification as a
complement to TIGGE and TIGGE LAM products.

The possibility to implement an European archive of high resolution precipitation analyses based on high
density network data is being explored. Contacts are established with ECMWF and with other European
groups and initiatives with the same objectives.

Basic set-up: Proposal/Desiderata Possible upgrades
◦24 hour cumulated precipitation at the ◦Shorter accumulation periods
highest possible resolution ( a guess: 0.1 °) ◦More parameters
◦Access to the analyses through registration
and by submission of a short work
description.
◦Data providers will be automatically
informed about the registration and they will
have one week (?) to raise objections. After
one week the user is registered.
◦Analysis methodology: the methodology by ◦More analyses methodologies applied to the
Anna Ghelli (and Carlos Santos) same dataset
◦Availability of a standard verification
package

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
HIPRECA: HIgh resolution PREcipitation Analyses for verification as
a complement to TIGGE and TIGGE LAM products.

Criticity > lack of resources

Cooperation with other similar initiatives in Europe.
A meeting has been organized by Anna Ghelli at ECMWF on the 2nd Dec 2010. This meeting was
aimed to coordinate this activity with the EUMETGRID/EUMETNET project leaded by Ole Tveito
from Met.No. Participants: Anna Ghelli (ECMWF), Ole Tveito (Met.No), Dan Hollis (MetOffice),
Tiziana Paccagnella (ARPA-SIMC TIGGE LAM). This initiative should also link with other ongoing
projects like EURO4M and ARCIS (Northern Italy initiative)

Possibility to set-up a coordinated activity (to be hopefully funded by submitting a project proposal to
EU)

Ongoing contacts with Ole Tveito.

This issue will be raised next week at the MWFR WG meeting in Berlin.

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
Courtesy R. North SRNWP Interoperability Programme
Coordinator: Rachel North – Met Office

Continuation of the SRNWP Programme under discussion
 Possible deliverables:
 Maintenance Plan for the Adaptors
 Extension of the existent SW tools to provide Lateral Boundary Conditions
from Global and LAM to LAM
 Soil aspects and Interoperability of Surface fields

The ET EPS and TIGGE LAM have been requested by Rachel North to
express their opinions about the different possible options for phase two
Next steps (during Autumn):
 Final agreement among the Programme Partners about the content of the
continuation proposal
 Submission of the proposal to STAC/PFAC and the the EUMETNET Assembly

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
Activities in NA by Josh Hacker

Joshua Hacker

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
Joshua Hacker

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
Activities in Asia by Jing Chen

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
Asia Regional EPS Activities
Jing Chen

• China, Japan and Korea will held workshop on Numerical
Weather Prediction at Seoul from 1 Sep. 2011 to 2 Sep
2011

• To share experience and knowledge on promotion of
ensemble prediction system (EPS) products under
uncertainty

• Discussing Regional cooperation for developing
8/30/11
Comparison of GEPS and REPS for China Precip.
Precipitation Probability (>0.1mm,36-60h)
in China average for 2010092612-2010102512

1.0
0.9 GEPS REPS

Observed Relative Frequency
Brier Score 0.8
Reliability diagram
0.7
0.6

0.5

36-60hr 0.4
36-60hr
0.3
0.2
0.1
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Forecast Probability

Area of ROC, Total Precipitation >(13.0mm) in China
average for 2010092612-2010102512
0.9
Locations of 2510 rain gauge stations
Area of ROC for heavy rain
GEPS REPS
0.8
Area of ROC

0.7

0.6

0.5
6 12 18 24 30 36 42 48 54 60
Forecast range [hours]
CMA Future plans
• The Numerical Prediction Center (NPC) ,CMA was
established in April. 2010.

• NPC will develop The Global/Regional Assimilation and
Prediction System (GRAPES) as the next generation
medium-range operational model of CMA.

• NPC will develop GRAPES-based GEPS and REPS also.

• 2015 : real running of GRAPES-REPS with 0.15 degree
8/30/11
Development of Mesoscale Ensemble Prediction
System (MEPS) at JMA
Objective: provide information on uncertainty in operational numerical
prediction from the mesoscale model (MSM).
(MSM: operated on a domain covering Japan and its surrounding region with 5km
grid spacing.)

Current Status:
Methodologies have been implemented to create perturbed Initial conditions.
Singular Vector Method, Ensemble Data assimilation (EnKF, En3DVar)
Consideration is underway on
resolution and ensemble size:
more emphasis on probability information (=> low resolution, large ensemble
size) / on multiple scenarios (=> high resolution, small ensemble size)
design of perturbed boundary condition, model-physics perturbation etc.
Forecast Example from the MEPS (3-h precipitation)
Observation

Initial : 18 UTC, 11 July 2010 (T+06)
Forecast : dx =10 km, Num. of forecasts 41
Perturbation : Singular vector method

Control forecast Ensemble maximum Ensemble spread
Researches at MRI/JMA relating to regional EPS

1) Development and comparison of initial perturbation methods using mesoscale/global SV,
BGM and LETKF and lateral perturbation methods in the WWRP Beijing 2008 Olympics
Research and Development Project (B08RDP):
•Saito, K., M. Kunii, M. Hara, H. Seko, T. Hara, M. Yamaguchi, T. Miyoshi and W. Wong, 2010: WWRP Beijing 2008 Olympics Forecast
Demonstration / Research and Development Project (B08FDP/RDP). Tech. Rep. MRI, 62, 210pp.
•Saito, K., M. Hara, M. Kunii, H. Seko, and M. Yamaguchi, 2011: Comparison of initial perturbation methods for the mesoscale ensemble

prediction system of the Meteorological Research Institute for the WWRP Beijing 2008 Olympics Research and Development Project
(B08RDP). Tellus, 63A, 445-467.
•Kunii, M., K. Saito, H. Seko, M. Hara, T. Hara, M. Yamaguchi, J. Gong, M. Charron, J. Du, Y. Wang and D. Chen, 2011: Verifications and

intercomparisons of mesoscale ensemble prediction systems in B08RDP. Tellus, 63A, 531-549.
•Duan, Y., J. Gong, M. Charron, J. Chen, G. Deng, G. DiMego, J. Du, M. Hara, M. Kunii, X. Li , Y. Li, K. Saito, H. Seko, Y. Wang, and C. Wittmann,

2011: An overview of Beijing 2008 Olympics Research and Development Project (B08RDP). Bull. Amer. Meteor. Soc. (conditionally
accepted)
•Saito, K., H. Seko, M. Kunii and T. Miyoshi, 2011: Effect of lateral boundary perturbations on the breeding method and the local ensemble

transform Kalman filter for mesoscale ensemble prediction. Tellus. (conditionally accepted)

2) Validation of regional EPS and ensemble storm scale data assimilation:
•Seko, H., K. Saito, M. Kunii1 and M. Kyouda, 2009: Mesoscale Ensemble Experiments on Potential Parameters for Tornado Formation.
SOLA, 5, 57-60.
•Saito, K., T. Kuroda, M. Kunii and N. Kohno, 2010: Numerical Simulations of Myanmar Cyclone Nargis and the Associated Storm Surge Part

2: Ensemble prediction. J. Meteor. Soc. Japan. 88, 547-570.
•Aonashi, K. and H. Eito, 2011: Displaced Ensemble Variational Assimilation Method to Incorporate Microwave Imager Brightness

Temperatures into a Cloud-resolving Model. J. Meteor. Soc. Japan. 89, 175-194.
•Seko, H., T. Miyoshi, Y. Shoji and K. Saito, 2011: A data assimilation experiment of PWV using the LETKF system -Intense rainfall event on 28

July 2008-. Tellus, 63A, 402-414.

3) Development of a cloud resolving ensemble analysis/prediction sytem in the K-computer
project
• http://www.jamstec.go.jp/hpci-sp/kisyo/kisyo.en.html#kisyo_2
Activities in SA by Celeste Saulo

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
By Celeste Saulo

Super model ensemble system

http://www.master.iag.usp.br/ind.php?inic=00&prod=intercomparacaodemodelos/phps
By Celeste Saulo

SMES
• Now includes 54 model outputs, that combine
GCMs, with regional models (mainy ETA-BRAMS,
and WRF).
• Some models contribute with different members
of their ensemble (CPTEC models, for example)
• Products are available at station points (not
maps!)
• The web page is interactive, and the user can
select any model, and any synoptic station and
plot surface variables forecasts
By Celeste Saulo

Examples

• Surface temperature according to some models.
• Blue dots are the observations and red curve is the Super
Ensemble Mean (bias corrected!)
By Celeste Saulo
?
Complex case...

11-12 days of useful forecast

Note that this is particularly
difficult period to forecast!

Large
discrepancies
7-8 days of usefull forecast
By Celeste Saulo
Products developed at CIMA (CONICET-UBA) -
ARGENTINA
http://wrf.cima.fcen.uba.ar

Using the SMES, we routinely produce probabilistic precipitation forecasts over part of
South America. These products are calibrated using GTS data and/or CMORPH data
(other results available in Ruiz et al 2009 (W&F) and Cardazzo et al 2010
(Meteorologica)
TIGGE LAM @ ….
 DRIHMS (Distributed Research Infrastructure for Hydro-Meteorology Study) -First
open consultation meeting. Genova 14 Oct. 2010. “Limited Area Ensemble Forecasting
and TIGGE-LAM” Tiziana Paccagnella

 EGU 2010 Meeting, Vien. TIGGE, TIGGE LAM and the GIFS T. Paccagnella D.
Richardson D. Schuster R. Swinbank Z. Toth S. Worley

 4th HYMEX workshop, Bologna, 8-10 June 2010. “HyMeX and TIGGE-LAM”. Laurent
Descamps, Andrea Montani, Tiziana Paccagnella

 QPE QPF Meeting, 2010 Nanjing. Ongoing developments on LIMITED AREA
ENSEMBLE FORECASTING and TIGGE LAM Tiziana Paccagnella

 SOCHI Kick Off Meeting, 2011 Sochi. TIGGE LAM & SOCHI 2014 T. Paccagnella
R. Swinbank Z. Toth

 WWRP JSC Meeting 2011 Geneva
 EPS DA SRNWP meeting – 2011 Bologna
 LAM EPS meeting – 2011 Bologna

 Contribution to the definition of the GEO-WOW FP7 proposal
 Contribution to THORPEX mid-term report
 ERM – Sarah Jones
 TIGGE – R. Swinbank and Zoltan Toth

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
Thank you!

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
SUPPORTING AND COMPLEMENTARY SLIDES

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
Slide by C. Marsigli

SRNWP-EurEPS-2013-2017
Cooperation on European Limited-area Ensemble Prediction
Systems

EurEPS - A proposal by the SRNWP Expert Team on Predictability and EPS

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
Slide by C. Marsigli The new EurEPS proposal

To be presented at the EUMETNET Assembly on November 22nd
 In the Roadmap for the Forecasting Capability Area of
EUMETNET the creation of a Eur-EPS Programme is
envisaged
 A major cooperative effort is required to develop a
capability for convection-permitting ensembles in
order to address prediction of severe or high-impact
weather in a probabilistic framework
 EurEPS - Phase I is proposed to be carried out in 2013,
in order to identify properly all the needed technical
facility, the requirements for Research and
Development to design properly this innovative kind
of systems and the potential framework for running
the Phase II as a demonstration project

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
Slide by C. Marsigli
The new EurEPS proposal
Phase I

Phase I will:
 build whenever possible on existing know-how and expertise,
available within the various European NWP Consortia;
 plan the use of results from SRNWP-I and SRNWP-V project for
verification and interoperability of EPS;
 schedule scientific and technical experiments for later to design a
convection-permitting ensemble system;
 consider the most suitable initial state, its ensemble spread, the models
errors and lower-boundary uncertainties, the available nesting
techniques;
 meet the requirements of all the involved partners, to provide a
suitable system, applicable by the different Meteorological Centres,
which will permit to significantly improve the forecasts of high impact
weather
 Analyze and evaluate:
 state-of-the-art of LAM-EPS scientific methodologies for very high-
resolution ensemble, including technical constraints;
 available and future computer resources scenarios;
 consider the necessary products for key end-users
 deliver the EurEPS Phase II project proposal for the EUMETNET Advisory
Committees

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
Slide by C. Marsigli
The new EurEPS proposal
Phase II

 Phase II would be executed over a 4-year time
frame (2014–2017) as a demonstration project
that, outside of a fully operational context,
would hopefully show the usefulness of its
results towards fundamental case studies and
the likely implementability of a full EurEPS
project with a distributed load of
responsibility among Members when
adequate computing resources might be
allocated.

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
Slide from A. Montani
LAMEPS BC project
Aim: generation of EPS boundary conditions (at 6 and/or 18 UTC) for
use of LAMEPS community
The issue: to make progress in terms of increasing the efficiency of archiving and retrieving
EPS data.
The solution: to archive the data on the native reduced Gaussian grid of the ECMWF IFS, but
only in a region of interest over the Northern Atlantic and Europe
The next step: to check if the regionally archived EPS data can be used by the LAMEPS
groups.

tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
tpaccagnella@arpa.emr.it 9th GIFS TIGGE WG Geneva 31 Aug – 2 Sept 2011
NIMR LAM EPS: Domain and Configurations
Ensemble Configurations
Domain
Model(member name) Configurations

▪ Horizontal grid : 18 km (181 Ⅹ211)

MM5 (M1~M4) ▪ Vertical layer : 35 sigma layers

GSMM5 (G1~G4) ▪ Initial, lateral boundary condition : GDAPS

▪ 4times/1day, 60hr forecast (interval:1hr)

▪ Horizontal grid : 15 km (251 Ⅹ251)

WRF ARW(W1~W4) ▪ Vertical layer : 35 sigma layers

WRF NMM (N1) ▪ Initial, lateral boundary condition : UM

▪ 4times/1day, 60hr forecast (interval:1hr)

▪ Horizontal grid : 15 km (251 Ⅹ251) GSMM5: A Generalized Sigma coordinate
system for the MM5 has been established to
▪ Vertical layer : 40 sigma layers
explain uncertainties which resulted from
WRF ARW (K1, I1) ▪ Initial condition : KLAPS(K1), GSI(I1) topographic effect (Zangl, 2003)

▪ lateral boundary condition : UM  GSI : Gridpoint Statistical Interpolation (GSI),
▪ 4times/1day, 60hr forecast (interval:1hr) replaced the Spectral Statistical Interpolation
(SSI) 3DVAR system at the National Centers for
▪ Horizontal grid : 12 km (540 Ⅹ432)
Environmental Prediction (NCEP)
▪ Vertical layer : 38 charney-phillips
tpaccagnella@arpa.emr.it
UM (U1)
(continued)
Model Member Micro physics Cumulus PBL

M1 WSM5 re-KF2

M2 WSM5 KF2
MM5 YSU
M3 Schultz KF2

M4 Reisner2 KF2
G1 Schultz KF2
G2 Reisner2 KF2
G3 Schultz Grell
G4 Reisner2 Grell
W1 WSM3 KF2
W2 WSM3 BMJ
GSMM5 YSU
W3 WSM5 KF2
W4 WSM5 BMJ YSU
WRF ARW
WRF NMM N1 Ferrier BMJ MYJ
K1 WSM6 KF2

WRF ARW-DA I1 WSM6 KF2 YSU
MOSES-II
UM U1 Improved Mixed phase CMODS
non-local K
 re-KF2 : revised Kain-Fritsch 2 by Lee et al. (2006)

 GSMM5 : YSU PBL version (NIMR)

 MM5 : YSU PBL version (NIMR), WSM5 (using WRF WSM5, NIMR)

 UM : regional model forecast from NWP operation

tpaccagnella@arpa.emr.it