WORLD METEOROLOGICAL ORGANIZATION
COMMISSION FOR BASIC SYSTEMS
OPAG DPFS
CBS-DPFS/CG-NERA/Doc. 7.1
(24. IV.2008)
_______
COMMISSION FOR BASIC SYSTEMS
Agenda item : 7
NUCLEAR EMERGENCY RESPONSE
ACTIVITIES
COORDINATION GROUP
ENGLISH ONLY
MELBOURNE, AUSTRALIA, 5-8 MAY 2008
Ensemble Atmospheric Transport Modelling (RTH Offenbach)
(Submitted by RTH Offenbach)
Summary and purpose of document
This document gives a short information on the RTH Offenbach
(DWD) activities regarding Ensemble atmospheric transport modelling
Action Proposed
The meeting is invited to discuss Ensemble atmospheric transport modelling in the context of ERA.
Ensemble atmospheric transport modelling at the Deutscher Wetterdienst
I.
THE EUROPEAN ENSEMBLE SYSTEM OR EMERGENCY DISPERSION MODELS
The emergency dispersion models used in different countries differ from one another
conceptually and in terms of results. This is basically due to the fact that the approach to the description
of the atmospheric dispersion process is not unique and to the fact that weather forecasts originating
from different NWP models are used for producing the dispersion calculations. Therefore, in the
framework of an EU-project a permanent internet-based platform was established at the JRC Ispra,
which can in real time collect dispersion forecasts provided by more than 20 models operational in the
EU and around the world. The results of these models are combined to a "multi-model" ensemble
dispersion forecast. Additionally, the platform provides statistical tools for the analysis of model results
that include the possibility to quantitatively estimate the variability of the forecast and its relative
uncertainty. The ENSEMBLE system collects forecasts of atmospheric concentration, deposition (wet
and dry) and precipitation over a domain reaching from Iceland to the Urals and from Norway to Northern
Africa 3. If necessary the domain can be shifted to other regions of the earth. Since 2000 GME-LPDM
and LME-LPDM results are transmitted to the ENSEMBLE system for exercise applications. Examples
of the evaluation of these experiments including a re-analysis of the ETEX-simulation are presented in
the next section.
II.
EXAMPLES OF ENSEMBLE APPLICATION
In the last seven years more than 20 exercises were performed in the context of the EUENSEMBLE activities featuring the release of one ore two volatile nuclides. The examples presented
here are obtained with tools provided by the ENSEMBLE system. They are typical for the display of
model result and the usage of multi-model ensemble techniques. The figures relate to exercise 4. The
scenario applied was based on the assumption that an accidental release of Cs-137 took place on 5
Feb 2005 at 12 UTC near Nantes (France) with an estimated source strength of 1000 TBq/hour and a
release period of 6 hours. Regarding the results of the German emergency dispersion model (GMELPDM, DE1) the radioactive plume was transported to an ENE direction covering the northern part of
Germany in the following 24 hours with surface concentration values estimated above 10 Bq/m³ (fig.
1). During the following day (48h forecast) the plume was simulated further to the east and left
northern Germany, but a part of the plume was transported to the south east of Germany. The
maximum values at the surface decreased below 1 Bq/m³ at that time (fig. 2). These estimations were
supported by the corresponding ensemble calculations, which were based on the then available
dispersion forecasts of other weather services (RSMCs and NMCs). The ensemble plots (fig. 3 and
fig. 4) show the percentage of models forecasting concentration values above critical threshold
values (1 Bq/m³ resp. 0.1 Bq/m³). The calculated agreement level is 60 to 80 %.
REFERENCES
1. GLAAB H., FAY B., and JACOBSEN I.,"Evaluation of the Emergency Dispersion Model at the
Deutscher Wetterdienst using ETEX Data", Atmospheric Environment 32, 4359-4366 (1998)
2. GALMARINI S. et al., "Ensemble Dispersion Forecasting - Part I: Concept Approach and Indicators.
Part II: Application and Evaluation", Atmospheric Environment 38, 4607-4632 (2004)
3. GALMARINI S. et al., "Can the Confidence in Long Range Atmospheric Transport Models be
increased? The Pan-European Experience of ENSEMBLE", Radiation Protection Dosimetry 109,
No. 1-2, 19-24 (2004)
Fig. 1:
Fig. 2 :
Surface concentration (Bq/m³) at T + 24 h
(GME-LPDM)
:
Surface concentration (Bq/m³) at T + 48 h
(GME-LPDM)
Fig. 3:
Agreement (9 models) on concentration threshold
level (1 Bq/m³) at T+ 24 h
Fig. 4:
Agreement (8 models) on concentration threshold threshold
level
(0.1 Bq/m³) at T+ 48 h