04_EEA_EUMETNET_WS_minutes_050407

							                                  Workshop on

 Real time air pollution data exchange and forecast in Europe
                    EEA, Copenhagen, 7-8 April 2005



                                 Draft minutes



Main conclusions
A successful first joint EEA/EUMETNET workshop was held with about 50
participants. Participants came from National Meteorological Services (NMSs),
National Environmental Agencies (NEAs), the European Environment Agency (EEA)
and its European Topic Centre on Air and Climate Change (ETC/ACC), the European
Commission (DG Environment and JRC), EUMETNET (Working group on
environment), GEMS (Global Earth-system Monitoring using Space and in-situ data),
PROMOTE (PROtocol MOniToring for the GMES Service Element) and various
regional/local organisations. Also the US Environmental protection agency (US EPA)
participated. See the participants list available on the ETC ACC web site (http://etc-
acc.eionet.eu.int/).

The workshop agreed on the following:
  The need for improved up-to-date information on air pollution for the public and
   for data exchange on real time monitoring and forecasting at the European level.
  The need to recognise different interests and needs of the involved groups: the
   public; the research community; monitoring and data experts and policymakers.
  The type of information to be provided to the public and many aspects of data and
   data exchange.
  A need for further discussion on the extent of access to information to the various
   different groups.
  EUMETNET WG ENV and EEA will jointly prepare a proposal for a plan (for
   2005-2008) for a European real time and forecasting information network,
   incorporating the main conclusions from the workshop and building on existing
   systems. The proposal will be developed with all interested parties. For EEA this
   is a key part of the “in your neighbourhood project”.
  The proposal will be submitted to the EUMETNET Council and EEA
   management board and/or national focal points. It should also be presented at the
   next EEA EIONET workshop on air quality management (autumn 2005). The
   proposal should be in line with EU proposals for revision of monitoring and
   reporting requirements, developed under the thematic strategy Clean Air for
   Europe (due to be published by the European Commission in mid 2005).
  A next joint workshop is proposed to be held in autumn 2006.
Overview
The main underlying reason for organising the joint EEA/EUMETNET workshop is
that the public should be properly informed on air pollution levels with harmful
effects on human health by means of up-to-date easily accessible information. The
workshop was organised jointly by EEA and the EUMETNET Working Group on
Environment (WG ENV) and in collaboration with key research projects GEMS and
PROMOTE. The workshop objectives were to discuss the required elements of a
European real time air quality information system, to develop a strategic vision how
to achieve such a system, to prepare recommendations for an EEA real time air
quality information system and a EFNet (Environmental Forecasting Network for
Europe, EUMETNET WG) proposal for a forecasting system and to exchange
experiences on national and local information systems.

In the workshop a number of presentations were given by the European Environment
Agency (EEA), its European Topic Centre on Air and Climate Change (ETC/ACC),
the European Commission (DG Environment and JRC), National Meteorological
Services (NMSs), National Environmental Agencies (NEAs), EUMETNET (Working
group on environment), GEMS (Global Earth-system Monitoring using Space and in-
situ data), PROMOTE (PROtocol MOniToring for the GMES Service Element),
various regional/local organisations (e.g. AIRPARIF) and the US Environmental
protection agency (US EPA). See the agenda available on the ETC ACC web site
(http://etc-acc.eionet.eu.int/). See for more details a short overview below and the
presentations available on the ETC ACC web site (http://etc-acc.eionet.eu.int/).

After the presentations discussions took place in 4 working groups, for more
information, see below. In a final session a discussion took place on future steps,
identification of opportunities for cooperation, responsibilities and time tables.

After the workshop the EUMETNET WG on environment had a separate session to
discuss their EFNet proposal.




                                           2
Working group discussions

The four working groups addressed the following (and some additional) questions:

Working group 1 (Data and data exchange)
 Which pollutants?
 Combine near real time data and forecasts into one system?
 Required frequency of exchange (hourly)?
 How many stations are needed (what are limitations?)
 Criteria to select stations and who should do so?
 Quality checks needed, by whom?
 Which data to archive, where/by whom?

Working group 2 (information required to display)
 Present separate pollutants and/or an index?
 Use the same scheme for presenting (e.g. low/medium/high and colours)?
 Which level of detail for maps is needed for presentations?
 Display more details during episodes?
 Present for specific groups (children etc?)?
 Which language?
 Which media (internet, TV, SMS, other)?

Working group 3 (links between activities and organisations)
 What are the main ongoing local, national, European activities?
 Which organisations are involved and which are leading/coordinating?
 Which activities require coordination (e.g. to avoid duplication) or cooperation
  (e.g. to enhance synergy)?
 How can cooperation be arranged (e.g. regular meetings, next workshop)?
 How to keep all informed to the extent needed and feasible?

Working group 4 (forecasting, model development, boundary data)
 QA/QC issues
          o Operational (quality scores)
          o Model inter comparison
          o Output parameters and criteria
 Ensemble forecast
          o Ways to define ensemble
          o Prioritising
 Statistical corrections, data assimilation, etc
 Emission scenarios (short term abatement strategies)




                                          3
Working group 1 draft conclusions (Data and data exchange)
 Partners involved in data exchange are EUMETNET-community (NMS) and the
  EEA-EIONET community (NRC, NEA).
 End-users will be the public, policy makers and AP-experts; this might require
  two interfaces: one general website showing maps and graphs for the public and a
  – in the initial phase, restricted – website with download options for experts. The
  website must include a disclaimer saying that the data is provisional and not yet
  validated and it can therefore not be used for compliance checking.
 In the group discussion emphasis was on ground based measurements but in the
  (near) future an extension of the exchange platform with satellite and ground
  based remote sensing data (e.g. lidar) should be considered.
 Pollutants to be considered are ozone and PM10. In view of their important role in
  atmospheric chemistry, it is strongly recommended to include NO2/NOx. PM2.5
  should be included when the monitoring network is more mature.
 Near real time (NRT) data and forecasted data should be combined into one
  system, however, the first phase should start with NRT-exchange.
 The update frequency should be preferable once per hour. However, it is
  recognised that this is not always possible when non-automatic PM10 monitors are
  used and in winter months when the acquisition of ozone data has in some
  countries a lower frequency. The data should be archived for a limited period of
  time. The different users require different archive periods; for the public and for
  the use in data assimilation models, a period of several weeks is sufficient, for
  interpretation of satellite data, a couple of months should be available. In the first
  phase a 3-month archiving period is recommended.
 Rural and (sub) urban background stations but no hot-spot or traffic stations
  should be included. For rural O3 and PM10 stations the recommended density is
  one station per 75x75 and 150x150 km grid cell, respectively. The urban stations
  should be representative for the urban situation in a country. The MS are
  responsible for station selection. It is recommended that the stations included in
  the exchange form a subset of the stations available in AIRBASE. In this way a
  description of station and its surroundings is easily available. For ozone, it is
  recommended to select stations measuring both ozone and NO2.




                                           4
Working group 2 draft conclusions (Information required and its display)
The working group quickly developed a good atmosphere with an open discussion.
The moderator attempted to keep the wide ranging discussion focussed around the 7
questions. The questions were:
1. Present separate pollutants and/or an index?
2. Use the same scheme for presenting (e.g. low/medium/high and colours)?
3. Which level of detail for maps is needed for presentations?
4. Display more details during episodes?
5. Present for specific groups (children etc.)?
6. Which language?
7. Which media (internet, TV, SMS, other)?

1. Present separate pollutants and/or an index?
Yes definitely, an AQI should be presented. General consensus that the priority was to
present an AQI (derived) from local level and presenting the current situation. The
forecast situation would also be useful. There was discussion of whether to present
separate pollutants or not. It was agreed that they should be available if possible, but
not on the first level. Info about specific pollutants is needed for people managing
medical conditions who could become important users. The wider general public is
not that interested in specific pollutants. The idea of an expert interface was
considered and generally positively received.

2. Use the same scheme for presenting (e.g. low/medium/high and colours)?
The question of a common AQI was an important element of the discussions.
Consensus was quickly reached that a common scheme based around a limited
number of colours should be developed. However it was also quickly accepted that
the development of a common AQI that would be widely accepted could be difficult.
In the initial stages, the approach by Citeair was seen as positive as well as the idea to
attempt to map different AQIs to each other, particulary for the values that link to EU
thresholds / limit values. The conclusion was that more work was needed to
investigate the feasibility.

3. Which level of detail for maps is needed for presentations?
The discussion focused on how to add value without duplicating what was being done
elsewhere. A portal element to link to other providers was seen as key. The
conclusion was clearly that no simple scale could be given, however the following
principles could guide the level of detail that any such system.
Cities need to be able to compare their AQ to other cities, the same applies to
countries and regions.
Where no national, regional or local provider exists, it would make sense for a
Neighbourhood type interface to fill the gaps.




                                            5
4. Display more details during episodes?
The answer was a clear yes, but not of more detail of the pollution itself. The time /
geography combination is important. The added value of a European level system
would be to highlight where episodes are occurring and when they are likely to end.
As such the identity and a portal to information providers for the affected areas is
important and should be brought to the fore during episodes.
Functionalities based around this could be to compare cities in terms of number of
episodes/days of threshold exceedances, etc.

5. Present for specific groups (children etc.)?
Three target groups were clearly identified:
  General public (simple colours)
  Experts (Data)
  Media (Data)
Within the general public group, some key sub groups were identified: People with
medical conditions; educational sector.
Within the experts group, some key subgroups were identified: national and local
service / data providers.
The purpose of providing air quality information is to enable people to take action to
improve health. For this reason the principle of open public access was seen as key. In
addition, the question of trust was also discussed. A key conclusion was that in order
to build confidence in the European picture, exisiting trusted sources, services and
science should be used. Transparency about the source used as well as continuity and
reputation of the European service provider was seen as key.
A brief discussion on the value of providing a service to the public by proxy of other
services, such as google, webservices or as data streams to the media. In general this
was seen as positive and potentially a very good way to raise awareness both of
current situations and forecasts.

6. Which language?
This topic was determined to be a too narrow focus and expanded to include other
cultural aspects.
It was agreed that the public side had a requirement for multilinguality. However it
was also agreed that there was no need to provide extensive information at the
European level as it would be easier to refer to other service providers.
The principle should be to direct users to the lowest local level as quickly as possible
to enable access to local data providers. In addition there was consensus that local
forecasts should be cascaded up to a European level in a similar way to the US
approach. There was general acceptance of the Neighbourhood principles and focus.
A diagram of information flow was drafted.
Barriers included: data ownership and availability.

7. Which media (internet, TV, SMS, other)?
Internet and media were seen as key. Information about other services could be
provided.


                                            6
Working group 3 draft conclusions (Linkages between ongoing activities
and partner organisations)
  Key organisations to involve are national (+some local), NMSs, NEAs, EEA,
   EUMETNET (WG ENV), GMES (ECMWF), PROMOTE.
  There is a need for countries to organise cooperation on real time monitoring and
   forecasting, which are often performed by different organisations. Env Agencies
   are within Env Ministries and forecasting NMSs are usually under other
   Ministries (e.g. Ministry Transport). There are examples of well functioning
   cooperation in several countries.
  It is important to identify which organisation is responsible for communication to
   the public and for reporting to the European level (both mandatory/legal and
   voluntary). For EEA the contact points are always the National Focal Point
   (Agencies) and National Reference Centres for air quality, organised through
   EIONET.
  Differences between European programmes can be understood as follows: GEMS
   is a research programme aiming to deliver by 2009; PROMOTE is a research
   programme already delivering (as a prototype). There is no clear organisational
   framework for European cooperation on operational forecasting at the moment.
   Hence the EUMETNET WG ENV proposal was developed (EFNET).
  To be a success and have acceptance throughout Europe, EFNET must be
   inclusive, also including e.g. non-NMSs that are involved in forecasting.
  Europe may learn from the US, on the need to set up a clear framework of what is
   the goal, who does what, by when and how (a simple diagram of information flow
   will help much to convince management).
  Proposed is that EUMETNET WG ENV, jointly with EEA will revise the current
   EFNET proposal and develop into a 3-year plan, as an ad-hoc activity. The
   proposal should be submitted to the EUMETNET Council and EEA management
   board/national focal points (and DG Environment). It should also be presented at
   the next EEA EIONET workshop on air quality management (autumn 2005). For
   the next EIONET AQ workshop also EUMETNET WG, GEMS and PROMOTE
   will be invited.
  A possibility to develop in parallel a COST Action for EFNET will be explored.
  EEA will write its proposal for near-real time data exchange the coming months
   and will keep EUMETNET WG ENV informed (the proposal should feed into the
   joint proposal). An important goal is to make near real time data more visible on
   the web site, and to enhance country participation.
  There is a need to be in line with upcoming proposals for revision on monitoring
   and reporting (CAFÉ), but it is already clear that near-real time data exchange will
   not become mandatory.
  Next workshop is proposed for autumn 2006.




                                          7
Working group 4 draft conclusions (Further WGs for instance on
forecasting, model development/ integration, boundary data)
The main discussion was about the WP5 (system and model development). The tasks
were split to short-term ones, which are active from the very beginning and long-term
ones, which become more elaborate at the phase 2.
Short-term tasks: Real-time and scientific QA/QC
Both O3 and PM should be included into the network species list from the beginning.
Real-time QA/QC is an automatic procedure, done in real-time, immediately when
new model data and/or measurements become available. It is simple, quick, includes
only basic quality scores, which are then presented to public. The results of this
QA/QC are stored for long-term archive and used for selection of interesting cases
(also from the point of view of the model performance) for the scientific QA.
Scientific QA is based on past analysis of performance of the models. It is organized
as a similar way as model inter-comparison exercises. Selection of cases is done on
the basis of results of the real-time QA/QC, with re-run of all participating models for
these episodes. Scientific QA includes in-depth analysis of cases, behaviour of
individual models, precursor analysis, etc. Methodology of the first QA exercise
(EFnet year 1) should be comparatively simple and supporting the phase 1 of the
network establishment. Further sessions will build on its experience.
Long-term tasks: All feasibility studies
Ensemble forecasting
As soon as phase 1 is over, the multi-model ensemble is in place. The task is to
establish the missing links, evaluate the obtained dataset and compute the ensemble
output. At least the uncertainty estimates are coming right away. In case of reasonable
agreement (and variety!) of the multi-model ensemble, the mean model will be also of
high use.
Building the special tool for the ensemble computation is one option. Specific
configuration largely depends on the user needs (demand is determined from the WP
6 “User requirements”).
Existing experience of multi-model ensembles is high use: JRC - ENSEMBLE site
(however, direct utilization of that tool is not possible due to its strong orientation to
emergency modelling). A new FP6 project ENSEMBLES: climate modelling. Other
experience: CAFÉ (EuroDelta) – analysis of past events. GEMS regional air quality
ensemble approach: a tool producer?
Emission scenarios:
Important part of EFnet, which largely extends the network capability and
distinguishes it from many other activities – at comparatively low price.
   Feasible scenarios are defined together with users.
   Initially the scenario analysis goes during past analysis exercises (scientific QA).
   A few selected scenarios are put to the list of “feasible” ones to be run in real-time
    in case of forecasted strong events.
   An idea: BAT 2020.



                                             8
Data support
   Emission data and scenarios
   Boundary conditions for local-scale models? Output from which model? What
    parameters? Species? Decide together with local-scale modellers.
   Global boundaries: if someone is running global/hemispheric model, it can be
    used as a boundary for the other partners.
Small-scale model applications: low probability to have more than one model for the
specific city. Provide boundary conditions?




                                         9
Presentations Session I
 Jacqueline McGlade, EEA Executive Director, introduction and welcome

    Reasons for organising the joint EEA/EUMET workshop and the objectives of the
    workshop.

Identified needs and gaps, incentives from perspective of decision
making and forthcoming legislation
   Andrej Kobe, European Commission, DG Environment (Thematic Strategy on air
    pollution and the revision of air quality directives anticipating future information
    needs)
    The role of EC/the CAFE programme was presented followed by an overview of
    the Thematic Strategy on Air Pollution, revision of Air Quality legislation,
    INSPIRE and the Shared Information System.
   Frank de Leeuw, ETC Air and Climate Change (Limitation of current air quality
    data in Europe)
    Current air quality monitoring costs were shown. Furthermore, a number of
    shortcomings regarding data were listed. These can be remedied by: Improving
    reporting cycle, optimise networks, improve analytical methods and apply new
    interpolation/assimilation techniques.
   Gary Foley, U.S. Environmental Protection Agency (The future outlook for air
    quality forecasting in the US)
    Summary points: 1) US EPA is seeking better ways to measure the ultimate
    success of its regulatory programs, 2) CDC‟s Environmental Public Health
    Tracking program is seeking compatible air quality data to inform public health
    actions, 3) there are new possibilities for improving the way we characterize air
    quality and exposure, 4) EPA is building partnerships with public and private
    sectors, 5) EPA is building a database of high-resolution spatial maps of air
    quality over the U.S., and 6) EPA would like to work with EU in exploring the
    linkage between better air quality indicators and forecasts and human exposure
    and health.
   Andreas Skouloudis, European Commission, JRC (New data monitoring needs for
    assessing the impact of air-pollution on human health)
    The policy framework for air pollution linked to human health was presented. It
    was highlighted that current monitoring networks depict: 1) high infrastructure
    and measuring equipment costs, 2) static location and limited number of stations,
    and 3) low spatial representativeness. New monitoring ADA 4 system was
    introduced.
   Tim Haigh, EEA (In your neighbourhood project and real-time air quality
    information)
    The project „In your neighbourhood‟ and EEA‟s ideas behind it was presented. It
    can be seen as an INSPIRE pilot project. Air quality is in focus during 2005/2006
    (ozone website).




                                           10
Presentations Session II

EFNet objectives, proposed and related activities
 Ingo Jacobsen, Deutscher Wetterdienst (EFNet proposal objective and idea)
  Basic idea of EFNet was shown: 1) collect environmental forecasts and
  measurements, and 2) make them available to the public and various agencies and
  organisations. A number of products/outcome was listed in the presentation
  including project work packages.
 Mikhail Sofiev, Finnish Meteorological Institute (Added-value of multi-model
  simulation)
  The reason for ensemble modelling was presented and exemplified (single- and
  multi-model). Multi-model ensemble is a new tool in air quality assessment. The
  outcome from the ensemble usage shows a better stability than individual models.
  EFnet ensemble approach would be to build and understand the multi-model
  ensemble for O3 and PM. If successful, generate the ensemble air quality forecast
 Jørgen Brandt, National Environmental Research Institute (DK), (Integrated
  modelling and monitoring for use in forecasting)
  Integrated modelling and monitoring was exemplified on the basis of Danish
  experiences. Since 1998 three-days air quality forecasts at the different scales
  have been produced: European/regional, urban background and urban street.
  Ongoing developments include models with even higher resolution (1 km x 1 km
  at national scale) and implementation of data assimilation techniques.
 Peter van Velthoven, Royal Netherlands Meteorological Institute (Evaluation of
  air quality simulations with satellite observed NO2 columns and ground-based
  observations)
  Conclusions of the evaluation between SCIAMACHY satellite, Chimère model
  and ground-based observations was shown:
  Yearly mean:
  - very small bias SCIAMACHY - Chimère and Chimère - surface
  - Correlation coefficients ~0.7
  NO2 plumes similar
  Differences in details:
  - Seasonality (Chimère higher in winter)
  - Individual days
  - Detailed distribution
 Laurence Rouil, INERIS/Daniel Martin (FR), (PREV‟AIR system)
  The PREV‟AIR modelling system is a cooperative system for large scale air
  quality forecasting and mapping over Western Europe and France; developed in
  2003 upon an initiative of the Ministry of Ecology. The operational purposes are:
  in France, since 2004, public information related to pollution episodes (threshold
  exceedances) can be driven by forecasts (not only observations). Planned
  developments for 2005-2006 were highlighted.

Presentations Session III

Other activities in Europe and USA
  Tony Hollingsworth, ECMWF (UK), (The GEMS project on Monitoring and
   Forecasting Atmospheric Composition & possible interfaces with EFNet)
   GEMS objectives, deliverables, modelling and assimilation approach was listed.
   Daily and monthly-seasonal products from GEMS Operational (both global &


                                         11
    regional) were shown for greenhouse gases, reactive gases and aerosol, and UV_B
    forecast products.
   Eleni Paliouras, German Aerospace Center (DLR) Gerrit de Leeuw, TNO, (Air
    Quality Services of PROMOTE and potential areas of support for EFNet)
    Overall framework for PROMOTE was presented and current air quality services
    and products were listed. Beneficial outcomes of PROMOTE for EFNet was
    discussed.
   Richard Wayland, U.S. Environmental Protection Agency (Airnow: U.S. EPA's
    real-time and forecast air quality program)
    AIRNow as centralised national air quality network was explained in following
    elements: 1) collection of data and forecasts, 2) control/management of data, 3)
    communication of data. Finally the benefits/results/lessons learned was discussed.
   Francisco Ferreira, Universidade Nova de Lisboa (PT), (Qualar - from near to
    real-time air pollution data for Portugal to Ozone and PM10 forecast)
    The QUALAR information system (on-line database) and the data applied for it
    was explained. The model results achieved are very satisfactory (correlation
    always higher than 0.973).
   Karine Leger, AIRPARIF (FR), (AIRPARIF‟s initiatives at the regional scale,
    CITEAIR‟s products at a European level)
    The presentation was split into two: 1) AIRPARIF‟s initiatives at the regional
    scale, and 2) CITEAIR‟s products at European level. Re 1): Internet service for
    current day and tomorrow, monitoring data from each station in near real time,
    traffic emissions from the street each hour for 22 000 road sections in Ile-de-
    France. Re 2) The objective of CITEAIR is to development common approaches
    and sustainable solutions that can be applied throughout Europe by establishing
    „Best Practice‟ of air quality and traffic management through case study reviews
    of existing procedures, make a guidebook on common air quality reporting.
    Furthermore, decision makers, professionals and the public should be informed on
    the environmental situation through a Common Operational Website (COW) and
    an air quality index.
   Alexander Baklanov, DMI (DK), (Integrated Systems for Forecasting Urban
    Meteorology, Air Pollution and Population Exposure (FUMAPEX))
    The 5FP EC project FUMAPEX (integrated systems for forecasting urban
    meteorology, air pollution and population exposure) was presented and project
    objectives highlighted.
   Geir Endregard, NILU (NO), (The key results from the APNEE and Metropolis
    projects)
    Objectives and key findings from the APNEE and Metropolis projects were
    explained. Conclusions are: New techniques enhances information effect, on-line
    data needs to be max 1-2 hours old, georeferenced data might be positive,
    forecasts the most valuable content, simple text and colour symbols always work.
    A Norwegian example showed a dissemination solution implemented for local air
    quality.




                                          12