Transcript The THORPEX Interactive Grand Global Ensemble (TIGGE) Richard Swinbank, Zoltan Toth and Philippe Bougeault, with thanks to the GIFS-TIGGE working group and other colleagues.

The THORPEX Interactive Grand
Global Ensemble (TIGGE)
Richard Swinbank, Zoltan Toth
and Philippe Bougeault,
with thanks to the GIFS-TIGGE working group
and other colleagues
TIGGE overview
 Introduction to TIGGE
 Objectives
 The TIGGE archive
 TIGGE-LAM
 Early results based on TIGGE data
 Intercomparison of ensemble forecasts
 Use of multi-model ensembles
 Towards the Global Interactive Forecast System
 Implementing THORPEX science
 The TIGGE User Workshop
 What users can learn about TIGGE this week
TIGGE
A key component of THORPEX: a World Weather Research
Programme to accelerate the improvements in the accuracy of 1day to 2-week high-impact weather forecasts for the benefit of
humanity
The TIGGE project has developed a database of global
ensemble forecasts collected in near real-time.
Objectives:
 Enhance collaboration on ensemble prediction, both
internationally and between operational centres &
universities.
 Facilitate research on ensemble prediction methods,
especially methods to combine ensembles and to correct
systematic errors
 Enable evolution towards a prototype operational system, the
“Global Interactive Forecast System”
For more about TIGGE, see http://tigge.ecmwf.int
TIGGE infrastructure
 Data collected in nearreal time (via internet) at
central TIGGE data
archives
Predictability
science
academic
Applications
NHMS
users
 Could be implemented at
relatively little cost
 Can handle current data
volumes (approaching
300 GB/day) within
available network and
storage capabilities
NCAR
ECMWF
EPS 1
EPS 2
CMA
 More information in
Baudouin Raoult’s
presentation tomorrow
EPS n
Summary of TIGGE database
Centre
Ensemble
members
Output data
resolution
Forecast
length
Forecasts
per day
Fields
(out of 73)
Start date
BOM
33
1.50º x 1.50º
10 day
2
55
3 Sep 07
CMA
15
0.56º x 0.56º
10 day
2
60
15 May 07
CMC
21
1.00º x 1.00º
16 day
2
56
3 Oct 07
CPTEC
15
1.00º x 1.00º
15 day
2
55
1 Feb 08
ECMWF
51
N200 (Reduced
Gaussian)
N128 after day 10
15 day
2
70
1 Oct 06
JMA
51
1.25º x 1.25º
9 day
1
61
1 Oct 06
KMA
17
1.00º x 1.00º
10 day
2
46
28 Dec 07
Météo-France
11
1.50º x 1.50º
2.5 day
1
62
25 Oct 07
NCEP
21
1.00º x 1.00º
16 day
4
69
5 Mar 07
UKMO
24
1.25º x 0.83º
15 day
2
70
1 Oct 06
TIGGE-LAM:
TIGGE for limited-area models
The TIGGE-LAM panel, chaired by Tiziana Paccagnella (ARPASIM), supports the development of the Limited Area Model
Ensemble Prediction System component of TIGGE.
This Panel works in close coordination with the GIFS-TIGGE WG, in
liaison with pre-existing LAM EPS initiatives and in coordination with
the THORPEX regional committees.
Aims:
 encourage a coordinated approach to LAM EPS for the mutual
benefit of WMO members;
 facilitate the interoperability of the different modelling systems
contributing to TIGGE;
 coordinate the archiving of limited-area ensemble forecasts – the
three TIGGE archive centres have agreed to host a sub-set of high
priority data;
 contribute to the definition of scientific issues related to LAM EPS
and particularly to TIGGE-LAM and propose specific initiatives to
address these issues to advance LAM EPS.
Early Results from TIGGE
Ensemble
forecasts Z500
Spaghetti
Diagram
Courtesy Mio
Matsueda
Comparison of RPSS N. hem Z500
z at 500hPa (cf_as_an)
10 categories, cases 20061201-20070228_N90, area n.hem
ecmwf ukmo jma
z at 500hPa (cf_as_an)
10 categories, cases 20071002-20071115_N45, area n.hem
ecmwf ukmo ncep msc jma cma bmrc kma
1
ON07
(45c)
Ranked Probability Skill Score
0.9
0.8
0.7
0.6
0.5
0.4
ECMWF
UKMO
NCEP
MSC
JMA
CMA
BMRC
KMA
DJF07
(90c)
0.9
Ranked Probability Skill Score
1
0.8
0.7
0.6
0.5
0.4
0.3
0.3
0.2
0.2
0.1
0 z at 500hPa (cf_as_an)
010 categories,
1 2 3 cases
4 520070328-20070528_N62,
6 7 8 9 10 11 area
12 n.hem
13 14 15 16
0.1
1
1
Ranked Probability Skill Score
0.8
0.7
0.6
JJA07
(84c)
0.9
Ranked Probability Skill Score
AM07
(62c)
0.9
0.8
0.7
0.6
0.5
0.5
0.4
0.4
0.3
0.3
0.2
0.2
0.1
0
0z at 1500hPa
2 (cf_as_an)
3
4
5
6
7
8
9 10 11 12 13 14 15
10 categories, cases 20070607-20070831_N84,
area n.hem
fc-step
(d)
ecmwf ukmo jma cma bmrc
fc-step (d)
ecmwf ukmo ncep jma
0.1
0
1
2
3
4
5
6
7
8
9
10 11 12 13 14 15 16
fc-step (d)
from Park et al, 2008
0
0
1
2
3
4
5
6
7
8
fc-step (d)
9
10
11
12
13
14
15
Comparison of extra-tropical cyclone tracks
Ensemble mean error:
Position
(verified against ECMWF
analyses)
Ensemble mean error –
Propagation speed
Propagation speed bias
Courtesy Lizzie Froude
Comparison of multi-model ensemble with ECMWF
MGCE51
10 members from
each of 5 centres
MCGE168
All 12Z forecasts
MCGE327
All forecasts for
each day
Courtesy
Mio Matsueda
Multi-model combination: MSLP & T2m
Brier skill-scores of multi-model
combination of bias-corrected
ECMWF, Met Office and NCEP
forecasts.
Mean Sea Level Pressure and
500 hPa (not shown) show only
small benefits from combining
different models
2m temperature shows more
benefit (1-5 days improvement in
lead time)
More sophisticated weighting led
to only small additional skill
Courtesy Christine Johnson
MSLP > mean
T2m > mean
T2m > 90 %
Similarity of ensembles
between
D
between  MSE
D has small values (high similarity,
blue) if the between model variance is
small compared to the mean-squareerror of the multi-model mean.
mslp
S
temp
between
between  MSE
D+2
D+2
D+10
D+10
Applications: flood prediction
Predictions with a 5-day lead-time for a flood event on River
Jiu in Romania
from Pappenberger et al, 2008
Towards the Global Interactive Forecast
System (GIFS)
 The objective of the GIFS is to realise the benefits of THORPEX
research by improving the delivery of improved forecasts of highimpact weather.
 As a first step, the GIFS-TIGGE working group set up a pilot
project for the exchange of real-time tropical cyclone predictions
using “Cyclone XML” format.
 Further plans for the development of GIFS will be presented by
Zoltan Toth, and discussed later in the workshop.
The TIGGE User Workshop
 Overview of the TIGGE project
 The TIGGE archive and how to access it:
 Presentation during TIGGE-B session, Tuesday afternoon
 Hands-on user demonstrations in coffee/poster breaks
 Opportunity for you to tell us what you need
 Presentations of early results from TIGGE
 In TIGGE-A/B/C/D sessions and throughout the symposium.
 Plans for GIFS
 Presentation during TIGGE-C, Thursday afternoon
 Opportunity to get involved in developing improvements to
operational forecasting of high-impact weather: focus group
discussion during Thursday coffee/poster break.
Conclusions
 Since October 2006, the TIGGE archive has been
accumulating regular ensemble forecasts from leading
global NWP centres.
 The archive is a tremendous resource for the research
community at large, and in particular the science
working groups of THORPEX.
 TIGGE provides the basis for research and
development projects targeted at specific applications
of severe weather forecasts (health, energy, flood
warning, wind storms, fire weather, etc…).
 Products to enhance the prediction of high-impact
weather will form the basis of the development of the
Global Interactive Forecast System.
TIGGE website: http://tigge.ecmwf.int