(Innovative WEather Presentations on TELevision)


The participants in the DMU team are

Annemarie Bastrup-Birk, Jørgen Brandt and Zahari Zlatev from ATMI

On the contents of this WEB-site.

The main tasks, which will be solved by the specialists from the DMU team (the National Environmental Research Institute) participating in the EU project WEPTEL, are briefly and without any technical details described and discussed in this WEB-site. The tasks that have to be solved by the specialists from the other five teams are not discussed here but links are given to the common WEPTEL WEB-site where more details about their tasks can be found.

What is WEPTEL?

WEPTEL is an abbreviation of "Innovative WEather Presentation on TELevision". This is an EU-project with the same name under the ESPRIT-programme (WEPTEL's project-number is 22727), which is carried out by a consortium of six partners. The consortium will develop a set of tools which can be used by TV companies in order to improve the presentation of the weather forecasts; partly by applying more advanced graphical programs and partly by introducing new topics in the weather forecasts. The introduction of graphical products containing air pollution information is one of the main new topics which are proposed by the teams participating in the WEPTEL project. Such information in modern TV weather forecasts will tell the population from the area under consideration when high, and possibly harmful, air pollution levels are expected and, if necessary, what can be done in an attempt to reduce the damages caused, say, by very high ozone levels.

Who is participating in the WEPTEL-project?

Six teams are participating in this project (see the section "Partners" in the WEPTEL web-site): three teams from private or semi-private TV companies and three teams from research institutions. One of the teams is from DMU (the National Environmental Research Institute).

The participants in the DMU team are Annemarie Bastrup-Birk, Jørgen Brandt and Zahari Zlatev from ATMI (the Department of Atmospheric Environment).

The main tasks solved by the teams participating in WEPTEL.

All the tasks that have to be solved by the teams participating in WEPTEL are described in detail in a special report "The Project Programme" (this was the basis for signing a contract for creating the WEPTEL Consortium with the European Commission). This document has been prepared according to the requirements for the Esprit programme. It is possible to find in this document: (i) a description of the structure of the consortium WEPTEL, (ii) full information about all packages of tasks which must be solved by the participating teams and (iii) a list of all reports (deliverables) which must be prepared (with the deadlines for their publication and with the teams which are responsible for their preparation). It is not necessary in this WEB-site to discuss the WEPTEL programme in detail (as mentioned above, more details about the tasks solved from the other partners are given in the common WEB site for the WEPTEL project; in the section "Objective, approach and results"). The tasks solved from the DMU team will be discussed in the next section.

The main task for the DMU team.

The team from DMU is responsible for the production of data which can be used to present information about air pollution in the weather forecasts. This information will complement the other information in the weather forecasts in situations where some air pollution levels are high and, therefore, may be harmful for the human health.

The team from DMU must be able to produce input data about different air pollution items for de graphical programs which prepare visualizations (and animations when necessary) of the transport and the distribution of high concentrations from a series of harmful compounds in relevant European areas. It should be possible to used these visualizations in connection with the TV weather forecasts. This main task of the DMU team is solved in five stages (it should be possible to run the five stage procedure in an operational manner, i.e. every day, in the end of the WEPTEL project).



Short description of the contents of the different stages

1 Meteorological data, which are calculated by the KNMI's HIRLAM, are transmitted to the computer, on which the DMU air pollution models are run.
2 The transmitted data file are reformatted to the format, which is used in DMU air pollution models.
3 The DMU models are run in order to produce the desired forecast for the air pollution levels in the next two or three days.
4 The output data files, which are produced by the DMU models, are verified and visualized either by using DMU's graphical tools on a powerful ONYX2 Silicon Graphics computer or by applying the graphical modules, which are developed by the teams from NOB and RUL (on the same computer).
5 All relevant output files are reformatted to the format which is used by the TV companies and subsequently the obtained files are transmitted to them.

A heavy use of HPCN (High Performance Computing and Networking) is absolutely necessary in the efforts to perform successfully this five stage procedure. Networking is the kernel of the first and the last stage. High Performance Computing is essential for the very important Stage 3. Until the last two-three years, the efficiency of the computational process was only measured in MFLOPS (million floating point operations per second). This concept is now starting to be old-fashioned. The concept of GFLOPS (billion floating point operations per second) is now becoming more and more popular when the efficiency of the computations is measured. If all relevant physical and chemical processes must be adequately be described (and this is necessary since the uncertainty will be increased when this is not done), then it is necessary to use in Stage 3 a computer on which it is possible, in principle at least, to achieve computational speed of many GFLOPS. Moreover (and this is a very difficult task), it is also necessary to develop computer programs, by which one will actually be able to achieve such high computational speeds. The supercomputers available at UNI-C (Danish Computer Centre for Research and Education) are used in the runs of the DMU large-scale air pollution models. The access to all supercomputers from UNI-C is granted by SNF (the Danish Natural Sciences Research Council). A CRAY C92A vector processor is still used in the operational runs (a computational speed of about 0.5 GFLOPS is achieved when the models are run on this computer; this is practically an upper limit because the theoretical top-performance, which can be obtained on CRAY C92A, is 0.9 GFLOPS). Therefore, it is necessary to apply a more powerful computer. There is an IBM SP parallel machine at UNI-C. Its theoretical top-performance is 16 GFLOPS when 32 processors are used. Some modules of DMU's models are now running with a computational speed of about 2 GFLOPS on 32 processors. It is expected that it will be possible to achieve even greater computational speeds. This short analysis shows clearly that it is possible to run the DMU large-scale air pollution models on the available computers when all physical and chemical processes are described in a sufficiently adequate way (or, in other words, there is no need to introduce non-physical assumptions only in order to reduce the amount of the computational work and, thus, to make it possible to handle numerically the models on the available computers).

Some demonstrations


1. Damaging levels of the ozone concentrations

Ozone is one of the most harmful pollutants in the troposphere. High ozone concentrations can damage vegetation, animals and humans. According to the EU regulatives, the EU population should be informed when the hourly values of the ozone concentrations are greater than 90 pbb. Moreover, a warning must be given, again according to the EU regulatives, whenever these values exceed 180 ppb. Examples where both these critical levels are exceeded are given in figures 1 and 2 (below). It is clearly seen (figure 1) that in a large region of Europe the ozone levels were over 90 ppb in the day under consideration (July 30 1994). In parts of Netherlands also the second level, 180 ppb, is exceeded. More detailed information, obtained by zooming, is displayed in figure 2 (the values of the concentrations for each grid-squares can be seen in figure 2). This example illustrates the need of pollution forcasts in some critical situations.

Furthermore, the WEPTEL system has been run and tested for two well-known episodes: an ozone episode in July, 1994, and the European Tracer EXperiment (ETEX). The model results are shown here


Figure 1: Episode of high ozone concentrations on July 30, 17:00 UTC, 1994, in Europe.


Figure 2: Episode of high ozone concentrations on July 30, 17:00 UTC, 1994, in the north-western part of Europe. Ozone concentrations above the warning threshold are seen in the The Netherlands and in western parts of Germany.


2. Harmful effects on crops

The damaging effect on vegetation depend on the magnitude of a critical parameter, the cumulative ozone exposure (AOT40, Accumulated Ozone Threshold over 40 ppb). Illustrations will be given in the near future.

Information on the DMU large-scale air pollution models.

Two of the DMU large-scale air pollution models are used in the WEPTEL project: DEM (the Danish Eulerian Model) and DREAM (the Danish Rimpuff and Eulerian Accidental release Model).

DEM can be used and was used to study (i) the distribution of concentrations and depositions of different air pollutants in Europe and (ii) transport of air pollutants from different area in Europe to Denmark. The model is fully described in [5], [8] and [9]. Output results obtained by the model were compared both with measurement that are taken at many stations located in nearly all European countries ([5], [6] and [7]) and with measurements that are taken over sea ([4]). DEM was also used in a series of simulations with different scenarios; first and foremost in ozone-studies over long time-periods (five and seven years). Some of the results obtained in these experiments are described in [1] and [10]. DEM is until now the only large-scale air pollution model in the world which has been applied in ozone-studies over so long time-periods as the studies which are described in [1] and [10].

DREAM is a relatively new model. It is developed to study the transport of dangerous compounds (and, first and foremost, radioactive compounds) emitted in the atmosphere after a powerful release which is a result from a great accident (the Chernobyl catastrophe being a typical example. The reliability of the results that are calculated by using DREAM were carefully tested within the large ETEX (European Tracer Experiment) project. The model is described in [2] and [3] (for a short introduction, see DREAM).

The DMU air pollution models can be used in many different studies. The application of the models in the investigation of critical levels of concentrations and depositions and in the development of control-strategies for regulation of concentration levels and/or deposition levels is the most important contribution of the DMU team to another EU project under the ESPRIT programme, EUROAIR (its project number is 24618). A particular task, which must also be solved in connection with the EUROAIR, consists of a series of studies of relationships between emissions and concentrations (or depositions) in France. This is a special demand from INERIS, the French end-user in the EUROAIR project.

The development of large scale air pollution models in Europe and in the other parts of the world is very fast. DMU will also in the future be a preferred partner in large international projects in this field if and only if the DMU large air pollution models are continously kept at a high international level. This is why it is necessary to continously improve the models by the following procedures:

  1. implementing more reliable physical mechanisms,
  2. more robust chemical schemes,
  3. more accurate numerical algorithms, and
  4. more efficient parallel computations.

In other words, it is absolutely necessary to carry out some basic research. Some activities in this direction are briefly discussed below.

NATO (the North Atlantic Treaty Organization) is supporting (through its scientific programme) a basic research project: "Environmental Mathematics and Development of Large-scale Air Pollution Models" (grants ENVIR.CRG.930449 and OUTR.CRG.960312). The work on this project is carried out in cooperation with a Bulgarian team from the Bulgarian Academy of Sciences in Sofia.

Moreover, an ARW (Advanced Research Workshop) is also supported by a grant from NATO. This project is carried out in a cooperation with the Russian Academy of Sciences. The ARW will be held in July 6-10, 1998, in Sofia (Bulgaria). The organizers of the meeting are G. Marchuk (from the Russian Academy of Sciences in Moscow) and Z. Zlatev (from DMU). Some of the best specialists in air pollution modelling, numerical analysis and parallel computations will participate in this ARW and will discuss a wide range of problems that appear in large-scale air pollution modelling.

The Danish Research Academy is supporting another basic research project "Object Oriented Software for Air Pollution Models". This project will be started in the near future in cooperation with two other Danish Institutes: UNI-C (Danish Computing Centre for Research and Education) and IMM (the Institute for Mathematical Modelling at the Technical University of Denmark).


1. A. Bastrup-Birk, J. Brandt, I. Uria and Z. Zlatev: "Studying cumulative ozone exposures in Europe during a seven-year period". Journal of Geophysical Research, Vol. 102, (1997), 23917-23935.

2. J. Brandt, A. Bastrup-Birk, J. Christensen, T. Mikkelsen, S. Thykier-Nielsen and Z. Zlatev: "Testing the importance of accurate meteorological input files and parametarizations in atmospheric transport modelling using DREAM - Validation against ETEX-1". Atmospheric Environment, to appear.

3. J. Brandt, T. Mikkelsen, S. Thykier-Nielsen and Z. Zlatev: "Using a combination of two models in tracer simulations". Mathematical and Computer Modelling, Vol. 23, No. 10 (1996), 99-115.

4. R. M. Harrison, Z. Zlatev and C. J. Ottley: "A comparison of the predictions of an Eulerian atmospheric transport chemistry model with experimental measurements over the North Sea". Atmospheric Environment, Vol. 28 (1994), 497-516.

5. Z. Zlatev: "Computer treatment of large air pollution models". KLUWER Academic Publishers, Dordrecht-Boston-London, 1995.

6. Z. Zlatev, J. Christensen and A. Eliassen: "Studying high ozone concentrations by using the Danish Eulerian model". Atmospheric Environment, Vol. 27A (1993), 845-865.

7. Z. Zlatev, J. Christensen and Ø. Hov: "An Eulerian air pollution model for Europe with non-linear chemistry". Journal of Atmospheric Chemistry, Vol. 15 (1992), 1-37.

8. Z. Zlatev, I. Dimov and K. Georgiev: "Studying long-range transport of air pollutants". Computational Science and Engineering, Vol. 1, No. 3 (1994), 45-52.

9. Z. Zlatev, I. Dimov and K. Georgiev: "Three-dimensional version of the Danish Eulerian Model". Zeitschrift für Angewandte Mathematik und Mechanik, Vol. 76 (1996) S4, 473-476.

10. Z. Zlatev, J. Fenger and L. Mortensen: "Relationships between emission sources and excess ozone concentrations". Computers and Mathematics with Applications, Vol. 22 (1996), 101-123.

Related home pages at the Department of Atmospheric Environment

Home page for the NATO Advanced Research Workshop "Large Scale Computations in Air Pollution Modelling, ENVIR.971731", Sofia, Bulgaria.

Home page for EU-project ESPRIT, 24618 EUROAIR

Home page for the REMAPE workshop in Copenhagen, September, 1996 REMAPE

Home page for the DEM model.

Home page for the DREAM model.

This page is maintained by Annemarie Bastrup-Birk , Jørgen Brandt, Helge Rørdam Olesen and Zahari Zlatev

Document date: January 19, 1998

[Dep. Homepage] Department of Atmospheric Environment, NERI