Laboratory of Heat Transfer and Environmental Engineering
Aristotle University Thessaloniki, GREECE
Mesoscale meteo and air quality
modelling
Technical Report
Deliverable D05.1
Final version
SUTRA project
Sustainable Urban
Transportation for the City of
Tomorrow
EVK4-CT-1999-00013
Deliverable D05.1
Project Deliverable: D05.1
Mesoscale Meteo and Air Quality Modelling
(Implementation Report and User Manual)
Programme name:
Research Programme:
Project acronym:
Contract number:
Project title:
Project Deliverable:
Related Work Package:
Type of Deliverable:
Dissemination level:
Document Author:
Edited by:
Reviewed by:
Document Versions:
Revision history:
First Availability:
Final Due Date:
Last Modification:
Hardcopy delivered to:
Energy, Environment and Sustainable
Development
1.1.4. - 4.4.1, 4.1.1
SUTRA
EVK4-CT-1999-00013
Sustainable Urban Transportation
D05.1
WP 05 Mesoscale meteo and air quality
modelling
RE (Technical Report)
RES (Restricted)
N. Moussiopoulos, K. Karatzas, A.
Arvanitis,
E.A.
Kalognomou,
I.
Theodoridou and E. Georgiadou
Laboratory of Heat Transfer and
Environmental Engineering
1.3 (FINAL)
2001 08 31
2002 06 30
2002 06 15
-2-
Deliverable D05.1
Executive Summary
The Mesoscale Meteorological and Air Quality modelling report aims to develop a
comprehensive overview of the photochemical air pollution modelling for long term
strategic planning that is being applied for the scopes of the SUTRA project. For this
reason, the report describes the tool used (the OFIS model) and the way it has been set-up
and implemented, provides a user manual, example data sets and calibration examples.
The overall aim is to provide a complete set of information that accompanies and fully
describes the OFIS tool and the actions taken in order to prepare the input data for the
model runs. This report is thus going to act as an implementation report and user manual
for the OFIS model.
Keywords: air quality modelling, air pollution
-3-
Deliverable D05.1
Contents
1
2
3
4
5
Urban Air pollution: basic principles...............................................................................5
Air Quality Management framework: the 96/62 EC directive .........................................6
Assessing and managing urban air quality: The use of models.....................................7
The Urban Ozone Problem ............................................................................................8
The Ozone Fine Structure model OFIS .........................................................................9
5.1
Introduction.......................................................................................................................... 9
5.2
The conceptual basis of OFIS ........................................................................................ 10
5.3
OFIS applications ............................................................................................................. 10
5.4
The OFIS User Manual ................................................................................................... 11
6 Ozone target and limit values in the EU ...................................................................... 11
6.1
Ozone pollution in Europe: current status .................................................................... 12
7 AQ modelling in the SUTRA project - The OFIS implementation ................................ 15
7.1
Genoa results.................................................................................................................... 16
7.2
Gdansk results .................................................................................................................. 19
7.3
Thessaloniki results ......................................................................................................... 21
7.4
Lisbon results .................................................................................................................... 24
7.5
Geneva results.................................................................................................................. 25
7.6
Tel Aviv results ................................................................................................................. 26
8 Conclusions for the OFIS application .......................................................................... 27
9 References .................................................................................................................. 28
ANNEX I: Ozone Fine Structure Model............................................................................... 30
-4-
Deliverable D05.1
1 Urban Air pollution: basic principles
Major air pollutants in urban areas or the so called “classical pollutants” are: Sulphur
dioxide (SO2), nitrogen dioxide (NO2), carbon monoxide (CO) and particulate matter (PM).
Other pollutants may also be important, for example, the volatile organic compounds
(VOC), characterised through the concentration of, for example, benzene or polycyclic
aromatic hydrocarbons (PAH), known to have adverse effects on human health such as
cancer, or the less dangerous for the health alkanes and aldehydes, extremely important
though since they contribute to the formation of photo-oxidants. Lead also used to be one
of the most commonly emitted pollutants but its significance has declined due to the
reduced lead contents of gasoline and the introduction of unleaded gasoline in most
European countries. Last but not least is the ground-level ozone, produced secondarily via
chemical reactions of nitrogen oxides and of non-methane volatile organic compounds in
the presence of sunlight.
During the last decades, the dominating sources of air pollution have changed in many
cities and populated areas from the combustion of high-sulphur coal and oil (causing for
example elevated SO2 and smoke concentrations) and from industrial processes to motor
vehicles and the combustion of gaseous fuels. In many Central/East European cities,
however, this shift is rather recent, and in some cities SO2 and smoke levels are still high.
The ambient concentration of air pollutants varies very much in time (daily, weekly and
seasonally, following the temporal profile of human activities resulting to air pollutant
emissions) and in space. It depends, apart from the morphological and meteorological
characteristics of the area concerned, upon the distance from dominating sources and the
location within a city. It is made up of the following contributions:

The natural background contribution;

The regional background contribution: Long-range transport of anthropogenic
emissions, as well as emissions from the cities themselves, leads to a regional
increase in the concentration levels of many pollutants and their chemical
transformation products;

The city background contribution: Concentration levels of a number of pollutants are
higher in cities than in the surrounding rural areas. This refers to the concentration
of pollutants at places within cities not directly influenced by sources such as
industry or traffic;

The traffic and industrial contribution: In busy streets and near industrial sources,
the concentration field is further elevated through nearby emissions. Traffic and
industrial concentrations refer to the concentration of pollutants at places directly
influenced by traffic or industry, the so called hot-spots.
High concentration levels, (the so called air pollution episodes), with a life-time of a few
days, are usually observed in urban areas when the large-scale synoptic weather situation
is unfavourable for dispersion and deposition, especially in case of enhanced regional
concentrations. Winter-type smog episodes occur during spells of cold winter weather
when a high pressure system persists for several days. Dispersion is limited due to low
wind speeds and a marked subsidence inversion. Winter-type air pollution episodes are
generally characterised by high concentrations of sulphur dioxide (SO 2) and particulate
matter (PM), mainly due to increased use of, and subsequent emissions from, fossil fuels
for space/domestic heating (RIVM, 1992). Summer-type smog episodes occur during warm
-5-
Deliverable D05.1
and sunny weather in the summer season. Under the influence of sunlight, ozone is formed
from nitrogen oxides and volatile organic compounds. At the same time the concentrations
of other secondary formed compounds are increased as well as those from primarily
emitted compounds such as traffic emissions (RIVM, 1992).
The actual occurrence and frequency of increased air pollution concentrations depends
primarily on the magnitude and the distribution of emission sources, on local topography
(e.g., flat terrain, basin or valley) as well as on the local meteorology (e.g., average wind
speed, frequency of calm weather conditions, occurrence of inversion layers) which
determines the degree of pollutant dispersion and mixing with cleaner air after the emission
took place; in Southern Europe, systems of local air circulation (such as land-sea breezes)
are particularly influential. Continental-size weather patterns (cyclones and anticyclones),
usually lasting a few days, can suddenly increase pollution loads on the regional scale,
resulting in air pollution episodes.
Following the above, it is apparent that the problem of air quality management should be
dealt with in a way capable of addressing the complexities of interactions between the
various physical, ecological, socio-economic and political aspects, components and actors
related with urban air quality, thus posing a considerable challenge to planners, policy and
decision makers and the general public. Moreover, there should be a distinction between
the problems that urban air quality management is dealing with and between problems of a
more generalised scale like climatic change. Climatic change, resulting from global
warming (a problem resulting from the emissions of the so-called greenhouse gases), is a
global environmental problem that is related to, but certainly not covered by, urban air
quality management.
2 Air Quality Management framework: the 96/62 EC
directive
The 96/62/EC directive on ambient air quality assessment and management, also called
“framework directive for air quality - FD”, was adopted by the European Council in
September 1996. The objectives of this Directive are to (van Aalst et. al., 1998):
 Define and establish objectives for ambient air pollution abatement in the Community
designed to avoid, prevent and reduce harmful effects on human health and the
environment as a whole;
 Assess ambient air quality in Member States on the basis of common methods and
criteria;
 Obtain adequate information on ambient air quality and ensure that it is made available
to the public inter alia by means of alert thresholds;
 Maintain ambient air quality where it is good and improve it in other cases.
Within these objectives, specific air quality management needs emerge. Thus, the most
important consequence of the FD to the built of AQMS is the need for including new
functionality aspects. As follows:
 Considerations of population density (definition of agglomeration, article 2 of FD)
 Geographical classification of polluted areas (FD, article 6, par. 2)
 Combination of monitoring and modelling techniques (FD, article 6, par. 3)
 Need for an integrated approach in order to achieve the aims of the directive (FD, article
7, par. 2)
 Information to be provided to the public and sensitive members of the community
(hospital, kinder gardens etc), when alert air quality thresholds are exceeded (FD, article
10).
-6-
Deliverable D05.1
 A detailed list of information to be included in the local, regional and national
programmes for improvement in the ambient air quality (FD, annex IV, and Dir.
99/30/EC).
 Continuous assessment of air quality in urban agglomerations.
In light of these “new” AQM needs, a contemporary AQMS should address all information
relevant with the problem at hand, provide access to appropriate tools and support
effective decision making. For this reason, the use of air quality models is essential, and is
introduced in the following chapter.
3 Assessing and managing urban air quality: The use of
models
In air pollution assessments information on all parts of the cause-effect chain have to be
collected. Not only a physical/chemical description of ambient air has to be presented in
such a way so as to compare it with threshold values, but also the relationship between
threshold values and the atmospheric emissions from sources (e.g. source categories,
countries, regions, economical sectors) should be quantified. For an optimal abatement
strategy to be developed it is essential that all three elements, (threshold or critical values,
ambient parameters and emissions) are available. Three types of instruments are used in
assessment studies: emission inventories (as a prerequisite for linking anthropogenic
activities with air emissions), air quality field measuring programmes and atmospheric
dispersion and transport models.
Field measurements form an important aspect of a system aiming at the description of air
pollution patterns in a given domain. Yet, observations are made at a limited number of
locations which are not necessarily representative for the entire area of interest.
Mathematical models may therefore prove useful for establishing consistent mass budgets
of emission, transport, transformation, and deposition of pollutants.
There are several examples for previous numerical simulations of air pollutant transport
and transformation in the local-to-regional scale which, broadly speaking, corresponds to
the mesoscale (Moussiopoulos et. al., 1995). In this context, it has already been
recognized that urban scale problems can only be treated successfully by the aid of
mesoscale air pollution models if either a large enough domain is considered or accurate
boundary conditions are established. Air pollution models require at input considerable
meteorological information. In the last years, two different approaches were followed in this
respect: Diagnostic wind field calculation, in conjunction with an empirical parameterization
for turbulence quantities, and prognostic calculation of both wind fields and turbulence
quantities. The former approach presupposes the availability of very detailed observed
data which would allow an accurate wind field reconstruction (Ratto et. al., 1994). This,
however, is under normal circumstances not feasible. Therefore, the latter approach, i.e.
the numerical simulation of the wind and turbulence patterns in the area of interest, is
nowadays widely preferred.
Both Eulerian and Lagrangian model types are being employed to describe the dispersion
of inert pollutants. Eulerian dispersion models predominate in the case of reactive
pollutants, typically ozone and its precursors (Peters et. al., 1995). Here it is usual practice
to apply the wind model first and the (photochemical) dispersion model subsequently.
In prognostic mesoscale models the large scale (temporal and spatial) distribution of all
problem variables is assumed to be known and is used to define initial and boundary
conditions. Major aim of these models is to describe how the problem variables are
affected by mesoscale influences (e.g. those associated with orography and
inhomogeneities in the surface energy balance). As a minimum requirement for a realistic
-7-
Deliverable D05.1
simulation of air pollutant transport in the local-to-regional scale, a prognostic mesoscale
air pollution model should include a reasonable parameterization with regard to the
dynamics of the atmospheric boundary layer. The latter depends on the turbulence
characteristics which may vary with both height and time.
Prognostic mesoscale models differ with regard to the treatment of pressure. If the
characteristic horizontal length scale (roughly corresponding to the grid spacing) is larger
than 10 km (which is not the case in the majority of urban air quality management
systems), nonhydrostatic effects (and thus also dynamical vertical accelerations) may be
neglected. In models adopting this approach, the so-called hydrostatic models, pressure
can be simply obtained from the hydrostatic equation. On the contrary, in nonhydrostatic
models the elliptic differential equation for pressure has to be solved, a fact usually
resulting in higher demands in computational resources. Nowadays efficient elliptic solvers
are available, and so the overall computational demand of a non-hydrostatic model is not
much higher than that of a hydrostatic model.
In most of the contemporary prognostic mesoscale models a transformation to terraininfluenced co-ordinates is performed to avoid difficulties in the formulation of the boundary
conditions at surface. Regarding the impact of the surface on wind flow and dispersion
characteristics, special care has to be taken to describe urban scale processes. Such
processes are in general much more complex than those at larger scales: Buildings and
other obstructions lead to very complex wind flow patterns in an urban area, while the
presence of large concentration gradients within cities makes it extremely difficult to find
representative locations for air quality monitoring stations. Additional difficulties may arise
from the typical intermittency of air pollutant concentrations in an urban area and from the
strong impact that concentration fluctuations may have with regard to chemical reactions
occurring in an urban airshed. Details on the overall structure of prognostic mesoscale
models are given in several previous articles and books (Physick, 1988; Pielke, 1984;
Schlünzen, 1994).
4 The Urban Ozone Problem
According to the report "Air Pollution in Europe 1997" of the European Environment
Agency, the EU ozone threshold value for the protection of human health (110 μg/m3, 8h
average) is exceeded substantially (Jol and Kielland, 1997). Based on measurements at
urban stations it was concluded that 80% of the EU urban population is exposed to these
exceedances. It is therefore evident that ozone related air pollution is of paramount
importance regarding urban air quality, this being the main reason for dealing with ozone
related air quality indicators in SUTRA.
For an effective reduction of high pollutant concentrations, an appropriate co-ordinated
strategy must be adopted. Such a strategy usually consists of a number of measures or
interventions which may be either of long term or of short term character. Long term
measures mainly aim at the reduction of mean pollutant concentration levels, whereas
short term actions focus on lowering peak concentrations.
The occurrence of air pollution episodes largely depends on dispersion and transformation
phenomena of pollutants in the atmosphere. It is well known that atmospheric transport of
pollutants, including ozone, is governed by phenomena occurring at different interacting
scales. In this context, larger scales provide boundary conditions for phenomena
characterized by the scale of interest. Thus, air quality in a local or regional scale domain is
-8-
Deliverable D05.1
not only influenced by primary pollutant emissions and secondary pollutant production
rates within the domain but also by the net pollutant transport across the boundaries.
Considering ozone, attention has to be drawn on its major precursor substances VOC and
NOx. VOC emissions originate partly from biogenic and partly from anthropogenic sources,
while the entire NOx emission is of anthropogenic origin, road traffic being one of the main
emitters.
Urban transportation is among the key issues when addressing the anthropogenic origin of
air pollution problems in cities. In many cases there are “city specific” factors that increase
these problems, like fast urban growth and increasing amount of commuter traffic of the
daily workforce from suburbs and surrounding areas, extreme asymmetry in seasonal
transportation demand due to tourism, a marked preference for individual cars, or
insufficient capacity and quality of the public transportation system, and many others.
The resulting environmental problems are equally severe in most cases, and even in the
cases where the overall environment is considered as reasonably clean, the contribution of
traffic to pollution dominates all other sources.
Finally, the trends in all cities indicate a worsening of the current problems: traffic volumes
increase and traffic congestion becomes an ever more frequent phenomenon.
Abatement strategies include legislative measures, such as the introduction of catalysts for
new cars and incentives favouring retrofitting of old cars, resulting in decisive reductions of
the average NOx emissions of car fleets. Nevertheless, increases in individual traffic often
counterbalance this reduction leading even to an increase of the total NOx emissions.
Hence, an increase of the background ozone levels at areas downwind of the major
precursor sources could be observed. The state of Baden-Wuerttemberg can be used as
an example: background ozone levels increased by 5 μg/m3 per year between 1984 and
1991 (Mayer and Schmid, 1992). Moreover, frequent exceedances of the ozone threshold
value of 180 μg/m3 initiated a discussion about the effectiveness of short term emission
interventions, such as driving prohibitions or speed limits during severe photosmog
episodes. For the first time, short term regional emission reduction concepts were
introduced in the states of Hessen, Schleswig-Holstein, Bremen and Niedersachsen as a
preparatory step towards a federal law which passed legislation in 1995 (URL 1).
5 The Ozone Fine Structure model OFIS
5.1
Introduction
What precursor emission reductions are needed in order to reduce sufficiently and costeffectively tropospheric ozone levels has to be addressed with suitable photochemical
simulation models. Until recently, urban photochemistry was analysed either crudely (with
simple box models) or by the aid of sophisticated 3D photochemical dispersion models.
The computational effort of the latter limits their applicability to episodic simulations which
only exceptionally can be considered as representative for longer time periods.
In contrast to these previous approaches, the OFIS model allows an adequate description
of urban photochemistry at a very low computational effort: This newly developed Eulerian
-9-
Deliverable D05.1
model is capable of simulating transport and photochemical transformation processes in an
urban plume. Thus it may be used for calculating urban scale ozone levels and for
detecting exceedances of ozone threshold values based on large scale anemological and
long range transport information over a longer time period (typically six months).
The Ozone Fine Structure (OFIS) model belongs to European Zooming Model system
(EZM), a comprehensive model system for simulations of wind flow and pollutant transport
and transformation (Moussiopoulos et. al., 1995). It is directly related to the photochemical
dispersion model MARS and makes use of the nonhydrostatic prognostic mesoscale model
MEMO (URL 2).
5.2
The conceptual basis of OFIS
The OFIS model was derived from well-tested full 3D models, and hence it retains
all elements necessary to achieve a realistic statistical evaluation of urban scale
ozone levels. The conceptual basis of OFIS is a 2D approach:


Background boundary layer concentrations are calculated with a threelayer box model representing the local-to-regional conditions in the
surroundings of the city considered. This model uses at input nonurban emission rates, meteorological data and regional scale model
results for pollutant concentrations (e.g. EMEP model results
(Simpson, 1993, 1995).
Pollutant transport and transformation downwind of the city (along the
prevailing wind direction) is calculated with a three-layer multibox
model representing a substantially refined version of MARS-1D
(Moussiopoulos, 1990).
The distinction of three individual layers of time depending thickness allows
adequately describing the dynamics of the atmospheric boundary layer. At the same
time, vertical transport is taken properly into account by considering the exchange
between adjacent layers.
For prescribing the thickness of the three layers, a 1D version of MEMO is utilised:
The vertical profiles of temperature, mean wind speed and turbulent exchange
coefficient as well as the mixing height are calculated both for the city surroundings
and the urban plume assuming Monin-Obukhov similarity at the lower boundary.
5.3
OFIS applications
Examples of previous OFIS model applications are available via the WWW:
Urban Plume Prediction and Application to Stuttgart (URL 1).
- 10 -
Deliverable D05.1
5.4
The OFIS User Manual
The OFIS model user manual, including a description of the input data required is available
in Annex I
6 Ozone target and limit values in the EU
In order to be able to proceed with an ozone-based urban air quality assessment, one has
to take into account the relevant legislative framework, as the latter provides specific
suggestions on the air quality “indicators” that should be used, the related limit and target
values, and the way that an AQ assessment should be planned, materialized and
integrated in the environmental management tasks performed by city or other
environmental authorities.
Currently EU member states are required to report ozone levels under the Council
Directive 92/72/EEC on air pollution by ozone. The ozone directives are daughter directives
of the EU's 1996 air quality framework directive (96/62/EC) and would require member
states to test ozone levels and take measures to reduce them to an indicative maximum
value.
The European Environment Agency (EEA) publishes reports that provide an overview of
the air pollution by ozone, including the data from the 15 EU member states as these are
reported to the EEA, under their obligations according to the framework directive. However
no reporting is required (yet) at city level.
According to the Council Directive 92/72/EEC, the health protection threshold is set at 110
μg/m3 for the mean values over eight hours, the population information threshold is set at
180 μg/m3 for the mean value over one hour, and the population warning threshold is set at
360 μg/m3 for the mean value over one hour.
On 9/6/1999 a proposal for a new Directive relating to ozone in ambient air was first
proposed (COM(99)125), which sets the target value for the protection of human health to
120 μg/m3 (highest 8-hour mean within one day, calculated from hourly running 8-hour
averages), and proposes that EU countries should aim to reduce exceedance s of the
target value to a maximum of 20 days per year by 2010, with the ultimate aim of achieving
no exceedances at all.
The information threshold remained set at 180 μg/m 3 for the mean value over one hour,
whereas the alert threshold (referred to as the population warning threshold Directive
92/72/EEC) is set at 240 μg/m3 for the mean value over one hour.
It is important to note that the fact that the 120 μg/m 3 target is proposed by the Commission
as a "target value" rather than as a mandatory "limit value", is in contrast with previous
"daughter" directives proposed under the EU's 1996 air quality framework directive. The
Commission argues that target values are justified since ozone pollution in one country is
often caused by pollution emitted in other countries, giving states only limited control over
pollution levels. However, it should also be noted that anticipating potential objections to
the stringency of its proposed ozone target level, the Commission notes that WHO
- 11 -
Deliverable D05.1
information shows that about 70% of hospital admissions can be attributed to ozone
concentrations of less than 160 μg/m3.
The amended proposal for a Directive relating to ozone in ambient air COM(2000)613,
changed neither the target values, nor the information and alert thresholds as these appear
in COM(99)125.
Finally, on 24/10/2001 the European Parliament announced that EU governments and
MEPs agreed on a new directive (which has not yet appeared in the Official Journal),
according to which the 120 μg/m3 target has not changed, however EU member states will
have to make sure that the target is not exceeded more than 25 times each year by 2010,
except where this is "not achievable through proportionate measures". As is mentioned
above, the Commission originally proposed 20 days but accepted that the weaker ceilings
agreed for precursor pollutants (nitrogen oxides and volatile organic compounds) now
make this unrealistic.
The member states will also commit to a non-binding long-term objective of ensuring that
the WHO standard of 120 μg/m3 is not breached at all by 2020, although governments had
wanted to avoid putting any date on this aim. Moreover, as ozone levels tend to be higher
in Southern Europe (due to high insolation), southern member states such as Spain,
Portugal, Greece and Italy attempted to set the agreement on a 40-day exceedance target,
but in the end, Swedish, Dutch and Danish pressure to maintain the Commission’s original
figure, led to the adoption of the lower compromise figure. The accord brought to an end
more than two years of institutional wrangling over the law.
6.1
Ozone pollution in Europe: current status
According to the EEA Topic report 13/2001 on Air pollution by ozone in Europe in summer
2001, and overview of exceedances of EC ozone threshold values during the summer
season April-August 2001 (de Leeuw and Bogman, 2001), Europe experienced slightly
more ozone pollution this summer than in 2000, largely due to better weather conditions.
However, overall the figures reveal little trend in summer ozone since 1995.
The highest "population warning" ozone threshold of 360 μg/m3 was not exceeded this
summer. However, Spain's highest recorded level was exactly 360 μg/m 3, while the
threshold was exceeded once in southern France in March.
As in previous years, there were frequent and widespread exceedances of the lower
"population information" threshold of 180 μg/m3. A preliminary evaluation of the AprilAugust 2001 period conducted for the European Commission shows that the public
information threshold was exceeded in 11 of the 15 EU Member States and in five out of 10
other European countries that supplied data at the EEA’s request. An exceedance
occurred in at least one of these 25 countries on 101 of the 153 days covered.
Italy recorded the highest number of days with exceedances of the public information
threshold (80), followed by France (58) and Spain (48). Of those countries reporting
exceedances, Poland had the fewest exceedance days (2). However, these numbers do
not necessarily give a fair comparison because of wide variations in the extent of different
countries’ ozone monitoring networks. Belgium and France both had the highest proportion
of stations reporting exceedances - 73%. The countries that recorded no exceedances of
the public information threshold this year were Bulgaria, Denmark, Estonia, Finland,
- 12 -
Deliverable D05.1
Ireland, Latvia, Lithuania, Norway and Sweden. For Ireland and Finland, 2001 marks the
fourth consecutive year without exceedances.
Differing definitions between the existing and new directive make it difficult to assess how
countries are currently performing against the future standards. However, the EEA notes
that about 5% of exceedances of the 180 μg/m 3 threshold this summer also exceeded the
future alert threshold of 240 μg/m3.
In common with previous annual reports, the EEA suggests a continuing trend in yearround ozone levels towards lower pollution peaks but higher background concentrations.
From 1994 to 1999, 120 stations show a significant upward trend and 12 show the reverse.
Falling European emissions of ozone precursor chemicals such as VOCs and nitrogen
oxides are suggested to account for the lower peaks. The rising background levels remain
mysterious but could be due to rising global precursor emissions.
Table 1. Summary of exceedances of the threshold for information of the public (1h ozone
concentration greater than 180μg/m3 ) during the summer 2001 (April to August) on a
country by country basis.
Country
No of
stations (2)
No of stations
with
exceedance
No of days
with
exceedances
(3)
Maximum
observed
concentratio
n (μg/m3)
Average
maximum
concentration
(μg/m3) (4)
Average
duration of
exceedances
(hour)
Greece (1)
13
6(46%)
12
273
205
2.0
Italy
314
59(19%)
80
353
203
3.5
Portugal
22
7(32%)
10
358
215
1.0
Switzerland
13
9(69%)
32
290
200
3.0
Poland
20
2(10%)
2
182
182
1.0
(1) Incomplete information was received
(2) Number of stations implemented in the framework of the ozone directive
(3) The number of calendar days on which at least one exceedance was observed.
(4) Average of all maximum concentrations recorded during exceedances
However, as mentioned in section 6: Ozone target and limit values in the EU, a revision of
the ozone directive is in preparation. In this proposed directive target values and long-term
objectives for protection of human health and vegetation have been defined which differ
from the values set in the current directive.
- 13 -
Deliverable D05.1
In the 1997 EEA annual report under the directive (de Leeuw et. al., 1999) an attempt was
made to evaluate the ozone data submitted under the current directive against the targets
and long-term objectives proposed in the new ozone directive. This analysis showed that,
due to the differences in the definition of the thresholds, the information collected under the
current directive is not adequate to assess exceedances of the newly proposed ozone
thresholds1. In the proposed directive the information threshold is defined similar to the
current directive as 1h average values of 180 µg/m 3. The proposed directive further defines
an alert threshold of 240 µg/m3 as 1h average and stipulates that for the implementation of
short-term action plans (as described in Article 7) the exceedance of the threshold should
be measured (or predicted) for three consecutive hours.
The available information allows an evaluation of the occurrence of exceedances of the
alert threshold but it is insufficient to evaluate whether ozone concentrations are above the
alert threshold during three consecutive hours.
1
The main difference between the current and the proposed directive is the time averaging. In the current
directive the eight-hourly average concentrations are calculated four times a day from the eight hourly values
between 0 and 8.00, 8.00 and 16.00, 16.00 and 24.00, 12.00 and 20.00. In the proposed directive the
maximum daily eight-hour running average is used. In the current directive the threshold for protection of
vegetation is defined as a daily averaged value; in the proposed directive, the value is calculated as AOT40,
that is, the sum of the difference between hourly concentrations greater than 80 μg/m 3 ( = 40 ppb) and 80
μg/m3 over a three-month period (May–July) using only the one-hour values measured between 8 a.m. and 8
p.m.
- 14 -
Deliverable D05.1
7 AQ modelling in the SUTRA project - The OFIS
implementation
Despite their diversity, cities and towns across Europe face the common challenge of
sustainable urban development. In the frame of the SUTRA project, the OFIS model is
applied to seven case study cities, for which the current tropospheric ozone concentration
levels, exposure, AOT values and other air quality related environmental indicators will be
estimated, and for which future development scenarios will be structured in order to
estimate future air quality.
The availability and implementation status of the OFIS urban air quality model assessment
application is presented in the following Table. The availability of the data does not
necessarily imply data quality. The inappropriateness of the data refers to the feasibility of
applying the model and reflects inherent difficulties for the data providers (city authorities)
in collecting QA/QC information for the purpose of environmental assessment, is not
restricted to the SUTRA project, and therefore does not refer to the validity of the model
results.
Table 2. Input data for OFIS simulations provided by city partners
Meteorological &
background
concentration data
Urban
emissions
Suburban
emissions
Rural
emissions
GENOA
✔
✔
✔
✔
GENEVA
✔
✔
✔
✔
THESSALONIKI
✔
✔
✔
✔
GDANSK
✔
✔
✔
✔
LISBON
BUENOS AIRES
TEL AVIV
✔
-
✔
-
✔
-
✔
-
✔
*
✔
✔
* Inappropriate data
Operational details of the OFIS air quality model can be found in deliverable D05.02.
Preliminary runs have been performed for the cities for which data are available and
appropriate, i.e. Genoa, Geneva, Thessaloniki, Gdansk, Lisbon and Tel Aviv (for the latter
some assumptions were used, to be presented in the relevant subchapter). It should be
underlined that, after a set of multiple tests and preliminary runs, it was decided to
complete the final runs per city, regardless of some input data problems that causes some
erroneous results, as it was evident that some city partners did not have the possibility to
provide with appropriate data, while the need to compile a final send of reference (base
case) scenarios for all cities was very pronounced, in order for the project to proceed. The
results as spatial graphs of maximum 8hour running average ozone concentration
exceeding 120μg/m3 (IND120) values are presented below.
- 15 -
Deliverable D05.1
7.1
Genoa results
IND 120
60
40
125
105
20
85
0
65
-20
45
25
-40
0
-60
-60
-40
-20
0
20
40
60
Figure 1. Number of days with maximum 8hour running average ozone concentration
exceeding 120μg/m3 (IND120), calculated by the OFIS model, for a 150x150km2 area
surrounding Genoa, and wind rose of prevailing wind during the summer semester of 1999
(draft reference scenario).
Figure 1 shows the initial estimation for the number of days exceeding 120μg/m3 – as daily
maximum 8hour average – for the six months period studied and the wind rose2
corresponding to this period. Instead of occurring in the urban ozone plume as expected,
exceedances are observed away from the city and the area influenced by it; this unrealistic
pattern is the consequence of air masses extremely rich in ozone and other pollutants
entering in the area from the domain’s boundaries. These boundary concentration values
2
Wind speed and direction distribution for the period of interest
- 16 -
Deliverable D05.1
of main photochemical pollutants were supplied by the partners and although they should
correspond to background levels, they were high enough to be observed in extremely
polluted urban plumes. Lower ozone concentrations – and consequently fewer
exceedances – are calculated in the urban area, as NO emissions consume ozone, and
downwind, as “cleaner” air is advected from the city. Based on the above findings, a
revised set of data was provided by the Genoa partners, resulting in the exceedances
presented in Figure 2.
IND 120
60
40
125
20
105
0
85
-20
65
-40
45
-60
25
5
-60
-40
-20
0
20
40
60
Figure 2. Number of days with maximum 8hour running average ozone concentration
exceeding 120μg/m3 (IND120), calculated by the OFIS model, for a 150x150km 2 area
surrounding Genoa, and wind rose of prevailing wind during the summer semester of 1999
(final reference scenario).
Figure 2 shows the final estimation for the number of days exceeding 120μg/m 3 – as daily
maximum 8hour average – for the six months period studied and the wind rose3
3
Wind speed and direction distribution for the period of interest
- 17 -
Deliverable D05.1
corresponding to this period. Exceedances are observed in combination with the urban
ozone plume, and the pattern is now more realistic, due to the refined boundary condition
data. Lower ozone concentrations – and consequently fewer exceedances – are calculated
in the urban area, as NO emissions consume ozone, and downwind, as “cleaner” air is
advected from the city. As a final check, comparisons were made between the results
presented here and the results that were made available from the GEA report (de Leeuw et
al., 2001). Thus, it should be noted that the exceedances calculated in the GEA report for
Genova were higher than the ones of the SUTRA reference scenario, the same standing
for the emission data, thus verifying the importance of the emission related ozone
production mechanisms in the area.
- 18 -
Deliverable D05.1
7.2
Gdansk results
IND120
60
40
125
105
20
85
0
65
-20
45
25
-40
0
-60
-60
-40
-20
0
20
40
60
Figure 3. Number of days with maximum 8hour running average ozone concentration
exceeding 120μg/m3 (IND120), calculated by the OFIS model, for a 150x150km 2 area
surrounding Gdansk, and wind rose of prevailing wind during the summer semester of
1999 (final reference scenario).
In contrast to the case of Genoa, few days of exceedance were observed for the area of
Gdansk. As shown in Fig. 2, these days are characterized by SE winds and high ozone
concentrations mainly above the urban area. Although there is a considerable occurrence
of NW winds and SW winds, no ozone plume is formed downwind of the city during the
corresponding days. Very low emission estimations for urban NOx and no emissions for
VOC at all, combined with zero boundary concentrations for ozone and VOC, resulted in
limited production of radicals which would interfere to the NO-NO2-O3 reaction chain to
favour ozone production. The small number of exceedances for most of the area is due to
the absence of VOC in the studied area. The high urban ozone levels occur in the days of
SE winds, as the pollution of air masses moving downwind of the city indicate in Fig. 2.
- 19 -
Deliverable D05.1
This happens also due to the scarcity of the emission data available; there are no NO
ground emissions for the urban area or the SE neighbouring cities of Elblag and Tczew to
“consume” ozone that NO2 – emitted in considerable amounts from these two cities –
photolysis produces. NW to Gdansk, in Gdynia and Sopot, comparable amounts of NO 2
are emitted, but ozone production is counterbalanced by NO emissions, which, for these
two cities, are available. As a final check, comparisons were made between the results
presented here and the results that were made available from the GEA report (de Leeuw et
al., 2001). Thus, it should be noted that the exceedances calculated in the GEA report for
Gdansk, and the relevant emissions, were the same with the ones of the SUTRA reference
scenario.
- 20 -
Deliverable D05.1
7.3
Thessaloniki results
IND120
60
40
125
105
20
85
0
65
-20
45
25
-40
0
-60
-60
-40
-20
0
20
40
60
Figure 4. Number of days with maximum 8hour running average ozone concentration
exceeding 120μg/m3 (IND120), calculated by the OFIS model, for a 150x150km2 area
surrounding Thessaloniki, and wind rose of prevailing wind during the summer semester of
1995 (draft reference scenario).
- 21 -
Deliverable D05.1
60
125
40
105
20
85
0
65
45
-20
25
-40
5
-60
-60
-40
-20
0
20
40
60
Figure 5. Number of days with maximum 8hour running average ozone concentration
exceeding 120μg/m3 (IND120), calculated by the OFIS model, for a 150x150km 2 area
surrounding Thessaloniki, and wind rose of prevailing wind during the summer semester of
1995 (final reference scenario).
In the case of Thessaloniki, exceedances are observed, as in the Genoa case, away from
the city influence and due to high ozone background concentrations exceeding in some
days the target value of 120μg/m 3. After the fist test run (Fiure 4), emission information
vere revisited and refined, resulting in the final reference scenario (figure 5). NO emissions
limit ozone production in the urban area as expected, but low VOC emissions prohibit
ozone formation downwind, similarly to the Gdansk case. The exceedance frequency
increases with the distance from the city centre, obviously in conjunction with the transition
from VOC to NOx limitation Although detailed and comprehensive emission data were
available for this case, the lumping of VOC from the original speciation to the one used by
the model was inadequate. NE of the city, the sea breeze effect is evident, as SW winds,
occurring in 40% of the studied summer days, advect poor in ozone air from the city. .As a
general remark, it should be noted that wind statistics in a city appear to drastically affect
the spatial pattern of the ozone exceedance frequency: In cases like Thessaloniki, where
- 22 -
Deliverable D05.1
the statistical predominance of certain winds is evident, OFIS results reveal the anisotropy
in ozone exceedance frequencies around the city centre. Finally, and in comparison to the
GEA results, emissions considered for the GEA report were higher than the ones
considered in the frame of the SUTRA scenario, yet the latter exhibits higher exceedances,
thus revealing the non-linear relationship of ozone levels and emissions, in the cases of
high photochemical sensitivity in a studied area (VOS vs NOx sensitivity), and the overall
photochemistry dominance in Ozone formation for the Thessaloniki area (as also
suggested by Güsten et al., 1997 and Zanis et. al., 2001)
- 23 -
Deliverable D05.1
7.4
Lisbon results
IND120
The analysis of the input data for Lisbon reveals that there are some basic background
concentrations missing. In comparison to the GEA results, emissions are lower. Overall,
no exceedances are observed and therefore the final reference scenario for Lisbon
should be considered as one with zero exceedances (this is why no Figure is provided
here). This results supports previous findings (Calheiros and Casimiro, 2001), that state
that for the year 1999, available ozone concentration level information suggest no
exceedances above the EU 1-hour threshold and no 8-hour exceedances of the WHO
guideline and EU thresholds. Nevertheless, the same reference also stated that
“preliminary analysis indicates that daily exposures in 1999 may have contributed up to
1.6% of all deaths ( 350 cases) and 1.9% of respiratory hospital admissions in Lisbon”,
thus revealing the complex relationship between exposure limit values and actual health
related impacts.
- 24 -
Deliverable D05.1
7.5
Geneva results
IND120
60
40
185
165
20
145
125
0
105
85
-20
65
45
-40
25
5
-60
-60
-40
-20
0
20
40
60
Figure 6. Number of days with maximum 8hour running average ozone concentration
exceeding 120μg/m3 (IND120), calculated by the OFIS model, for a 150x150km 2 area
surrounding Geneva, and wind rose of prevailing wind during the summer semester of
1995 (final reference scenario).
In the case of Geneva, the combination of low wind speeds and high emissions resulted in
high ozone exceedances (170 days for IND120). This is attributed also to the sensitivity of
ozone formation in correlation to VOC and NOx emissions, as discussed in detail for
Switzerland in Andreani-Aksoyoglou et. al., 2001. Last but not least, it should be noted that
high ozone concentration have already been reported for Geneva in literature, without
sufficient explanation on the cause-effect chain. (Neininger, 1997).
- 25 -
Deliverable D05.1
7.6
Tel Aviv results
IND120
In the case of Tel Aviv, input data were supplied by the city partner with delay due to
problems in collecting and compiling necessary information. Although ground level
emission information was not made available, the assumption that the elevated emission
data provided are identical with the (missing) ground level ones was used. The
corresponding results (representing vague input data), do not suggest any INT120
exceedances. These results correspond to OFIS model input data quality and availability
and their use in the frame of the SUTRA project should be considered as problematic.
- 26 -
Deliverable D05.1
8 Conclusions for the OFIS application
As it is evident from all the cases examined and described above, the provision of
complete and correct data is essential for gaining proper model results, or any results at all.
The difficulties of the city partners to provide these data although considerable and timeconsuming are steadily being overcome as experience is gained and exchanged among
them. Moreover
 In order to apply the OFIS model for a number of scenarios deviating from the
“reference” scenario, the new emission information is required following the
specifications provided for OFIS model inputs.
 Scenarios related with changes in emissions that do not affect the total sum of the
latter are not expected to have a significant impact in the OFIS-based assessment
of urban air quality.
 On-line integration is not possible in the timeframe of SUTRA as it requires efforts
not available within the project.
In addition to the above, and in order to support the formulation and treatment of emission
related scenarios from the SUTRA city partners, some suggestions follow focusing on the
traffic related emission percentile that dominates urban emissions in all cases. The
introduction of new internal combustion engine technologies in combination with the
reduced emission standards already planned in Europe and the use of new type of fuels,
suggest that car emissions will change in the future. This alteration is of paramount
importance when coming to the estimation of ozone levels with tools like OFIS, as it is
expected to affect VOC emissions and their splitting, e.g. their qualitative and quantitative
distribution. It should be noted that the first research work results on this issue appeared in
literature in the mid 90ies, suggesting, e.g. that the use of oxygenated compounds as fuel
components would appear to be helpful in reducing atmospheric ozone formation due to
the modification of VOC emissions (Bowman and Seinfeld, 1996). On the other hand,
future technologies like the direct injection and the variable compression ratio in
combination with high pressure charging are expected to result in alteration of emissions,
yet any reduction is expected to occur fro the use of after treatment technologies (catalytic
converters). As the latter will more effectively reduce VOCs that are oxidized more easily (,
and aromatic hydrocarbons are not included in that category while being an important
ozone precursor, it is evident that VOC emissions (and its splitting) for the cars of the future
can not be treated with adequate accuracy and on the basis of sufficient scientific
evidence. Thus, and for SUTRA purposes, the application of the VOC spitting already used
when studying the reference scenarios will advance project work, as it will allow for
comparisons on a “common basis”, being among the only safe and scientifically sound
approaches currently used.
- 27 -
Deliverable D05.1
9 References
Andreani-Aksoyoglou S., Lu C.H., Keller J., Prevot A.S.H. and Chang J.S. (2001),
Variability of indicator values for ozone production sensitivity: a model study for
Switzerland and San Joaquin Valley (California), Atmospheric Environment 35, 55935604.
Bowman F.M. and Seinfeld J. H. (1996), Atmospheric chemistry of alternate fuels and
reformulated gasoline components, Fuel and Energy Abstracts 37, Page 134.
Calheiros J.and Casimiro E. (2001), http://www.siam.fc.ul.pt/health/health05_01.ppt
de Leeuw F. and Bogman F. (2001), Air pollution by ozone in Europe in summer 2001,
EEA Topic report 13/2001, European Environment Agency, Copenhagen.
de Leeuw F., Moussiopoulos N., Bartonova A., Sahm P, Pulles T., Visschedijk A. (2001),
Air quality in larger cities in the European Union, A contribution to the Auto-Oil II
programme (GEA report), European Environment Agency, Topic report 3/2001.
de Leeuw F., Sluijter R. and de Paus T. (1999), Air pollution by ozone in Europe in 1997
and summer 1998, EEA Topic report 3/1999, European Environment Agency,
Copenhagen.
Gusten H., Heinrich G., Monnich E. and Weppner J., Cvitas T. and Klasinc L., Varotsos C.
A. and Asimakopoulos D. N. (1997), Thessaloniki '91 field measurement campaign--II.
ozone formation in the greater Thessaloniki area, Atmospheric Environment 31 11151126.
Jol, A. and Kielland, G. (eds), (1997), "Air Pollution in Europe 1997", EEA, Copenhagen,
1997.
Mayer H. and Schmid J. (1992), Trendanalyse von Immissionsdaten in BadenWurttemberg, LFU Karlsruhe.
Moussiopoulos N. (1990), Influence of power plant emissions and industrial emissions on
the leeward ozone levels, Atmospheric Environment, 24A, pp. 1451-146.
Moussiopoulos N., Berge E., Bohler T., de Leeuw F., Gronskei K., Mylona S. and Tombrou
M. (1995), Models for ambient air quality and pollutant dispersion/transport: State of
the art - Needs and trends, Report MA3-2, European Topic Centre on Air Quality.
Moussiopoulos N., Sahm P. and Kessler Ch. (1995), Numerical simulation of
photochemical smog formation in Athens, Greece-a case study, Atmospheric
Environment 29, 3619-3632.
Neininger B. (1997),
http://www.bbw.admin.ch/abstracts/abstr2000/abstracts/cost/c97.0044.html.
- 28 -
Deliverable D05.1
Peters L.K., Berkowitz C.M., Carmichael G.R., Easter R.C., Fairweather G., Ghan S.J.,
Hales J.M., Leung L.R., Pennel W.R., Potra F.A., Saylor R.D and Tsang T.T. (1995),
The current state and future direction of Eulerian models in simulating the
tropospheric chemistry and transport of trace species: a review, Atmospheric
Environment 29, 189-222.
Physick W.L. (1988), Review: Mesoscale modelling in complex terrain, Earth-Science
Review 25, 199-235.
Pielke R. A. (1984), Mesoscale Meteorological Modelling, Academic Press.
Ratto C.F., Festa R., Romeo C., Frumento O.A. and Galluzzi M. (1994), Mass-consistent
models for wind fields over complex terrain: The state-of-the art, Environmental
Software 9, 247-268.
RIVM (1992), The Environment in Europe: a Global Perspective, Bilthoven, report no.
481505001.
Schlünzen K.H. (1994), Mesoscale modelling in complex terrain – an overview on the
German nonhydrostatic models, Contr. Phys. Atmos. 67, 243-253.
Simpson D. (1993), Photochemical model calculations over Europe for two extended
summer periods: 1985 and 1989, Atmospheric Environment 27A, 921-943.
Simpson D. (1995), Biogenic emissions in Europe 2: Implications for ozone control
strategies, J. Geophysics Research 100 D11, pp. 22891-22906.
URL1: http://lhtee.meng.auth.gr/lhtee_old/ofis/
URL2: http://155.207.20.121/mds/bin/show_long?OFIS
van Aalst R., Edwards, L., Pulles, T., De Saeger, E., Tombrou, M. and Tonnesen D.
(1998), Guidance Report on Preliminary Assessment under EC Air Quality Directives,
http://eea.eu.int/frdocu.htm.
Zanis P., Zerefos C. S., Gilge S., Melas D., Balis D., Ziomas I., Gerasopoulos E.,
Tzoumaka P., Kaminski U. and Fricke W. (2001), Comparison of measured and
modelled surface ozone concentrations at two different sites in Europe during the
solar eclipse on August 11, 1999, Atmospheric Environment 35, 4663-4673.
- 29 -
Deliverable D05.1
ANNEX I: Ozone Fine Structure Model
OFIS
Version 1.0
User’s Guide
Laboratory of Heat Transfer and Environmental Engineering
Aristotle University Thessaloniki
February 2001
- 30 -
Deliverable D05.1
1 Standard model version – short description
The OFIS model was derived from well-tested full 3D models, and hence it retains all
elements necessary to achieve a realistic statistical evaluation of urban scale ozone levels.
Its conceptual basis is a coupled 1D/2D approach: Background boundary layer
concentrations are computed with a multi-layer box model for a domain of typically
150150 km2 with the considered city in the centre and rural area all around. For each day
of the period considered, pollutant transport and transformation downwind of the city is
calculated with a multi-layer multi-box model assuming the prevailing wind direction for the
respective day, an initial plume width according to the city diameter and a plume widening
angle of 30°.
In order to adequately describe the dynamics of the atmospheric boundary layer, OFIS
allows in the vertical direction for either a static non-equidistant grid level distribution or a
number of individual layers with time dependent thickness. At the same time, vertical
transport is accounted for by considering the exchange between adjacent layers. For
prescribing the thickness of the layers, a 1D version of MEMO is utilised: The vertical
profiles of temperature, mean wind speed, turbulent exchange coefficient and mixing height
are calculated both for the city surroundings and the urban plume assuming MoninObukhov similarity at the lower boundary. The multi-layer methodology is identical to the
one adopted in the multi-layer model MUSE (Sahm et al., 1995) which is another simplified
version of MARS (a few layers instead of “normal” discretisation in the vertical direction;
semi-implicit solver instead of a fully implicit one).
The mathematical analysis is based on the coupled, two-dimensional advection-diffusion
equations for the ensemble averaged quantities of reactive species. The equations are
solved by operator splitting according to the method of lines, that is by solving the
advection dominated terms separately from the diffusion dominated ones (in vertical
direction) and the chemical reaction terms. The concentration trends due to advection,
vertical diffusion and entrainment are then treated as source terms in the chemical reaction
equation system. The latter is solved in OFIS with a backward difference solution
procedure, i.e. by applying the Gauss-Seidel iteration scheme (Kessler, 1995). The model
uses a variable time step with an upper limit for the integration time increment (e.g. of 300
seconds).
Due to the modular structure of OFIS, chemical transformations can be treated by any
suitable chemical reaction mechanism, the default being the EMEP MSC-W chemistry
which has been described in detail in Simpson (1995). The dry deposition process for both
the city surroundings and the urban plume is calculated following the resistance model
concept (Sahm, 1998).
2 Extended model version to account for inhomogeneous
situations
Considering emission inhomogeneities in the city surroundings
Assuming that the background boundary layer concentrations are independent of the wind
direction is most likely erroneous in densely populated areas as high emitters (like another
city or industrial areas) could well be located upwind of the considered urban area and thus
affect air quality in the latter. By extending the domain of the three-layer multibox model
well upwind of the city centre, large emitters can now be taken into account in the vicinity of
the urban area depending on the prevailing wind direction of the respective day.
- 31 -
Deliverable D05.1
Accounting for local circulation systems
The assumption of one prevailing wind direction for a specific day does not allow resolving
local circulation systems. The extended version of OFIS accounts for a local circulation
system such as the sea-breeze in coastal areas by inversing the wind direction of the
urban plume in the lower two layers (i.e. up to the mixed layer) in the afternoon hours of
days with weak synoptic forcing and off-shore wind direction. The inversion is simulated by
interpolating the wind speed from positive values to negative values between two
subsequent hours.
3 Input requirements
OFIS requires data about the geography of the city, its population figure, emissions, the
atmospheric trace gases and data about the meteorology. These data can be obtained e.g.
by CORINAIR and EMEP. A summary of the input data files is shown in Table 1.
Table 1: Input files needed by the OFIS model
Class
a
b,c
d
d
d
Optional:
o.a
o.b
o.c
File description
Control File
Meteorological and boundary concentration data file
Urban ground level and elevated emission data file
Suburban ground level and elevated emission data file
Rural ground level and elevated emission data file
Neighbouring cities location, ground level and elevated emission data file
Land use data file
Meteorological file containing local circulation system information
OFIS input requirement classification:
a.
b.
c.
d.
o.a.
City longitude, city latitude, city population, urban city area, suburban city
area. Such information about the size, position and population can be found e.g. in
HEGIS (Health and Environment Geographical Information System) or similar
databases.
For each day of the year, large scale pressure forcing (synoptic conditions),
regional scale model results for meteorological quantities. Such data can be
obtained from models like EMEP (European Monitoring and Evaluation Programme)
or LOTOS (cost ~7500 Euro/year), OR from appropriate measuring stations and
following the necessary assumptions.
For background boundary layer concentrations, regional scale model results for
pollutant concentrations. Such data can be obtained from and models like EMEP
(European Monitoring and Evaluation Programme) or LOTOS (cost ~7500
Euro/year), OR from appropriate measuring stations and following the necessary
assumptions.
Urban and suburban emissions as well as rural emissions for an area 150x150
km2 around the city centre. Such data can be assembled from emission databases
like CORINAIR and models like EMEP (European Monitoring and Evaluation
Programme) or LOTOS (cost ~7500 Euro/year).
For considering emission inhomogeneities in the core city surrounding the
distance and direction of each neighbouring city to the centre of the OFIS domain
- 32 -
o.b.
o.c.
Deliverable D05.1
has to be specified as well as the neighbouring city’s ground level and elevated
emission strength like described under item d.
For taking into account biogenic emissions and dry deposition, the composition of
the rural land cover of an area 150x150 km2 around the city has to be provided.
Such data can be obtained from land cover databases like CORINE or GLCC
(Global Land Cover Characteristics for land use).
For accounting for a local circulation system such as the sea-breeze in coastal
areas, the date (for a sea-breeze day) and the offshore wind direction on that day
has to be provided.
4 Meteorological and boundary concentration data
For each day of the year, the large scale pressure forcing (synoptic conditions) is needed.
More specifically, OFIS requires diurnal average values for ground level wind speed and
direction, temperature and vertical temperature gradient as well as the lateral concentration
of the 66 EMEP species (shown in Table 4). The meteorological and boundary
concentration values are stored in one file in free format, each column separated by blanks
or commas. The data format typically used to provide this data file is shown in Table 2, a
file containing example values is shown in Table 3.
In the first column, a sequential record number has to be provided. The second and third
columns contain the date in form month and day, respectively. The next two columns
contain data on wind speed and direction, column number six and seven data on ground
level temperature and the temperature lapse rate (gradient), respectively. Columns 8
through 74 contain the concentration of each of the 66 EMEP species at lateral inflow (in
ppb).
Table 2: Typical data format used to provide the meteorological and boundary values input
file
Record
1
2
3
4
5
6
7
8-74
Format
I5
I5
I5
F7.1
F7.1
F7.1
F7.4
F9.4
Unit
[-]
[-]
[-]
[m/s]
[]
[C]
[K/m]
[ppb]
Description
row number
month of the year
day of the month
daily average wind speed
daily average wind direction
daily average temperature
daily average temperature gradient
daily average concentration of species at inflow
- 33 -
Deliverable D05.1
Table 3: Sample meteorological and boundary concentration input file
1 2 3 4
5
6
7
8
1 4 1 13.6 358.0 12.6 .0064 .o6
2 4 2 9.8 344.8 11.5 .0064 .o6
3 4 3 8.0 331.7 13.2 .0063 .o6
4 4 4 5.8 287.4 13.8 .0063 .o6
5 4 5 4.0 245.9 13.1 .0063 .o6
6 4 6 3.4 257.8 13.4 .0063 .o6
7 4 7 6.8 250.4 14.7 .0063 .o6
... ... ... ... ... ... ... ... ... ... ...
... ... ... ... ... ... ... ... ... ... ...
... ... ... ... ... ... ... ... ... ... ...
... ... ... ... ... ... ... ... ... ... ...
... ... ... ... ... ... ... ... ... ... ...
... ... ... ... ... ... ... ... ... ... ...
... ... ... ... ... ... ... ... ... ... ...
... ... ... ... ... ... ... ... ... ... ...
... ... ... ... ... ... ... ... ... ... ...
26 4 26 12.6 229.3 17.3 .0064 .o6
27 4 27 7.8 256.9 17.8 .0063 .o6
28 4 28 7.7 261.0 17.1 .0063 .o6
29 4 29 2.5 25.6 17.4 .0063 .o6
30 4 30 5.9 353.6 18.0 .0063 .o6
31 5 1 12.6 8.3 18.1 .0064 .o6
32 5 2 14.9 6.0 19.0 .0063 .o6
...
...
...
...
...
...
...
...
...
9
.o6
.o6
.o6
.o6
.o6
.o6
.o6
...
...
...
...
...
...
...
...
...
.o6
.o6
.o6
.o6
.o6
.o6
.o6
10
.o6
.o6
.o6
.o6
.o6
.o6
.o6
...
...
...
...
...
...
...
...
...
.o6
.o6
.o6
.o6
.o6
.o6
.o6
11
.o6
.o6
.o6
.o6
.o6
.o6
.06
12
.o6
.o6
.o6
.o6
.o6
.o6
.o6
13
.06
.06
.06
.06
.06
.06
.06
14
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.o6
.o6
.o6
.o6
.o6
.o6
.o6
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
.06
For following characters, see below *
Temperature gradient [K/m]
Temperature [C]
Wind direction [o]
Wind velocity [m/s]
Day of the month
Month of the year
Successive number
In the current version the reaction mechanism EMEP is implemented. The EMEP
reaction mechanism consists of 66 reactive species which have to be provided at lateral
inflow(in [ppb]):
NO
C3H6
H2O2
PAN
OXYO2
ISOPRENE
BURO2H
MACRO2
ISONO3H
NO2
O-XYLENE
H2
BURO2
MAL
NITRATE
ETRO2H
MPAN
ISNIRH
SO2
HCHO
CH3O2
MEKO2
MALO2
ISRO2
PRRO2H
CH2CCH3
CO
CH3CHO
C2H5OH
C2H5OOH
OP
MVK
MEKO2H
ISONO3
CH4
MEK
SULPHATE
ETRO2
OH
MVKO2
MALO2H
ISNIR
C2H6
O3
CH3O2H
MGLYOX
OD
CH3OH
MACR
MVKO2H
NC4H10
HO2
C2H5O2
PRRO2
NO3
CH3COO2H
ISNI
CH2CHR
Table 4: List of chemical species used for boundary concentration data
Column
no.
Code
Molecular weight
Species name
8
9
10
11
12
13
14
15
NO
NO2
SO2
CO
CH4
C2H6
n-C4H10
C2H4
30
46
64
28
16
30
Nitric oxide
Nitrogen dioxide
Sulfur dioxide
Carbon monoxide
Methane
Ethane
n-Butane
Ethene
28
- 34 -
C2H4
HNO3
CH3COO2
GLYOX
N2O5
OXYO2H
ISRO2H
MACRO2H
Deliverable D05.1
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
C3H6
C8H10
HCHO
CH3CHO
C4H8O
O3
HO2
HNO3
H2O2
H2
CH3O2
C2H5OH
SO42CH3O2H
C2H5O2
CH3COO2
PAN
BURO2
MEKO2
C2H5OOH
ETRO2
CH3COCHO
PRRO2
HCOCHO
OXYO2
C14H21NO3
MALO2
C14H22O
OH
OD
NO3
N2O5
C5H8
HNO3
ISRO2
C4H6O
MVKO2
CH3OH
CH3COO2H
OXYO2H
BURO2H
ETRO2H
PRRO2H
MEKO2H
MALO2H
42
106
30
44
72
48
33
63
36
2
32
61
96
Propene
O-Xylene
Formaldehyde
higher aldehydes
Methyl ethyl ketone (MEK)
Ozone
Hydroperoxy radical
Nitric acid
Hydrogen peroxide
Hydrogen
Methyl alcohol
Ethanol + higher alcohols
Sulphate
61
Etylperoxy radical
61
62
63
C4H6O
70
65
66
67
68
69
70
71
72
73
74
MACRO2
MPAN
CH2CCH3
ISONO3
ISNIR
MVKO2H
CH2CHR
MACRO2H
ISONO3H
ISNIRH
Peroxyacetyl nitrate
72
Methyl glyoxal (MGlyox)
206
17
Glyoxal (Glyox)
Peroxyradical formed from o-Xylene + OH
MAL
Peroxy radical from (MAL)
Octyl phenole (OP)
Hydroxyl radical
62
108
68
63
Nitrogen trioxide
Dinitrogen pentoxide
Isoprene
Nitrate
70
Methyl vinyl ketone (MVK)
Peroxyradical from MVK
Methanol
251
32
76
Hydroperoxide from OXYO2
ISRO2H
Hydroperoxide from MALO2 (Peroxyradical from MAL
= OH)
Methacrolein (MACR)
ISNI (Organic nitrate)
Hydroperoxide from ISRO2 + HO2 (ISRO2:
Peroxyradical from isoprene=OH
Peroxy radical from methacrolein = OH
Peroxy methacryloyl nitrate
41
Isoprene-NO3 adduct
Alkyl peroxy radical from ISNI (organic nitrate)
- 35 -
Deliverable D05.1
5 How to create the meteorology/boundary concentration
input file
The average value for wind speed can be obtained as follows:
Values for the temperature and temperature gradient are calculated in the same manner.
Furthermore, OFIS assumes one wind direction valid for 24 hours. Based hourly data or
similar, the unit vectors of the wind direction should be added and the wind direction
determined.
- 36 -
Deliverable D05.1
6 Emission data
In the emission data files hourly values of urban, suburban and rural emissions of 15
species (which must be supplied in a prescribed order) shown in Table 6, separated in
ground level or elevated (stack) emissions are taken into account.
Table 6: List of chemical species used for emission data
Column no.
Code
Molecular weight
Species name
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
NO
NO2
SO2
CO
C2H6
C4H10
C2H4
C3H6
C8H10
HCHO
CH3CHO
MEK
C2H5OH
CH3OH
C5H8
30
46
64
28
30
58
28
42
106
30
44
72
46
32
68
Nitric monoxide
Nitrogen dioxide
Sulfur dioxide
Carbon monoxide
Ethane
n-Butane
Ethene
Propene
o-Xylene
Formaldehyde
higher aldehydes
Methyl-ethyl-ketone (MEK)
Ethanol + higher alcohols
Methanol
Isoprene
For each sector (urban, suburban and rural), a separate file in free format (each column
separate by blanks or commas) with two blocks each has to be provided. The first block
contains hourly ground level emissions in kg/h, the second block contains hourly elevated
emissions in kg/h for each specie. In each block, the first column contains the hour (00 to
23), the second and third column the hourly emissions of NO and NO 2, respectively. The
fourth column contains the hourly emission of SO 2, the fifth column the hourly emission of
CO. Column no. 6 through 16 contain hourly emissions of various NMVOC. A sample
urban emission input file is shown in Table 7.
- 37 -
Deliverable D05.1
Table 7: Sample urban emission data input file
1
2
hrs no
0
4.24835
1
3.27099
2
3.27099
3
3.27099
4
3.27099
5
3.27099
6
9.44808
7
9.44808
8 12.13582
9 12.13582
10 12.13582
11 12.13582
12 12.13582
13 12.13582
14 12.13582
15 12.13582
16 12.13582
17 12.13582
18 12.13582
19 12.13582
20 12.13582
21 12.13582
22
4.24835
23
4.24835
hrs no
0
4.24835
1
3.27099
...
...
...
...
...
...
...
...
...
...
22
4.24835
23
4.24835
3
no2
.74971
.57723
.57723
.57723
.57723
.57723
1.66731
1.66731
2.14161
2.14161
2.14161
2.14161
2.14161
2.14161
2.14161
2.14161
2.14161
2.14161
2.14161
2.14161
2.14161
2.14161
.74971
.74971
no2
.74971
.57723
...
...
...
...
...
.74971
.74971
4
5
so2
co
6.18636 11.32104
5.86436
5.92388
5.86436
5.92388
5.86436
5.92388
5.86436
5.92388
5.86436
5.92388
10.44331 11.39477
10.44331 11.39477
11.32881 26.23696
11.32881 26.23696
11.32881 26.23696
11.32881 26.23696
11.32881 26.23696
11.32881 26.23696
11.32881 26.23696
11.32881 26.23696
11.32881 26.23696
11.32881 26.23696
11.32881 26.23696
11.32881 26.23696
11.32881 26.23696
11.32881 26.23696
6.18636 11.32104
6.18636 11.32104
so2
co
6.18636 11.32104
5.86436
5.92388
...
...
...
...
...
...
...
...
...
...
6.18636 11.32104
6.18636 11.32104
6
Ethane
.46992
.35768
.35768
.35768
.35768
.35768
.48767
.49111
.79953
.79787
.79714
.79682
.79674
.79714
.79670
.79638
.79634
.79779
.79904
.79921
.79840
.79718
.47044
.47000
Ethane
.46992
.35768
...
...
...
...
...
.47044
.47000
7
n-butane
3.68186
2.94262
2.94262
2.94262
2.94262
2.94262
3.72445
3.72788
5.76054
5.75889
5.75816
5.75783
5.75775
5.75816
5.75771
5.75739
5.75735
5.75880
5.76006
5.76022
5.75941
5.75820
3.68238
3.68194
n-butane
3.68186
2.94262
...
...
...
...
...
3.68238
3.68194
8
Ethene
.49163
.41861
.41861
.41861
.41861
.41861
.51851
.51851
.71932
.71932
.71932
.71932
.71932
.71932
.71932
.71932
.71932
.71932
.71932
.71932
.71932
.71932
.49163
.49163
Ethene
.49163
.41861
...
...
...
...
...
.49163
.49163
9
10
Propene o-xylene
.49081
1.50020
.43157
.82645
.43157
.82645
.43157
.82645
.43157
.82645
.43157
.82645
.50138
1.50670
.50138
1.50670
.66428
3.35951
.66428
3.35951
.66428
3.35951
.66428
3.35951
.66428
3.35951
.66428
3.35951
.66428
3.35951
.66428
3.35951
.66428
3.35951
.66428
3.35951
.66428
3.35951
.66428
3.35951
.66428
3.35951
.66428
3.35951
.49081
1.50020
.49081
1.50020
Propene o-xylene
.49081
1.50020
.43157
.82645
...
...
...
...
...
...
...
...
...
...
.49081
1.50020
.49081
1.50020
- 38 -
11
HCHO
.06084
.04699
.04699
.04699
.04699
.04699
.06309
.06331
.10136
.10126
.10121
.10119
.10118
.10121
.10118
.10116
.10115
.10125
.10133
.10134
.10129
.10121
.06087
.06084
HCHO
.06084
.04699
...
...
...
...
...
.06087
.06084
12
CH3CHO
.05381
.04487
.04487
.04487
.04487
.04487
.05620
.05620
.08079
.08079
.08079
.08079
.08079
.08079
.08079
.08079
.08079
.08079
.08079
.08079
.08079
.08079
.05381
.05381
CH3CHO
.05381
.04487
...
...
...
...
...
.05381
.05381
13
MEK
.11549
.06681
.06681
.06681
.06681
.06681
.11776
.11776
.25163
.25163
.25163
.25163
.25163
.25163
.25163
.25163
.25163
.25163
.25163
.25163
.25163
.25163
.11549
.11549
MEK
.11549
.06681
...
...
...
...
...
.11549
.11549
14
C2H5OH
1.15575
.81647
.81647
.81647
.81647
.81647
1.53060
1.53060
2.46363
2.46363
2.46363
2.46363
2.46363
2.46363
2.46363
2.46363
2.46363
2.46363
2.46363
2.46363
2.46363
2.46363
1.15575
1.15575
C2H5OH
1.15575
.81647
...
...
...
...
...
1.15575
1.15575
15
CH3OH
.00704
.00704
.00704
.00704
.00704
.00704
.00782
.00782
.00782
.00782
.00782
.00782
.00782
.00782
.00782
.00782
.00782
.00782
.00782
.00782
.00782
.00782
.00704
.00704
CH3OH
.00704
.00704
...
...
...
...
...
.00704
.00704
16
ISOPRENE
.25063
.14699
.14699
.14699
.14699
.14699
.26143
.26143
.54644
.54644
.54644
.54644
.54644
.54644
.54644
.54644
.54644
.54644
.54644
.54644
.54644
.54644
.25063
.25063
ISOPRENE
.25063
.14699
...
...
...
...
...
.25063
.25063
Deliverable D05.1
Typical procedure on how to create the emission files
A so far typical procedure for obtaining the required emission inventories is based on
CORINAIR and the GEA (Generalised Exposure Assessment) methodology.
CORINAIR (Coordination d'Information Environmentale) is an emission inventory for NO x,
SO2, and NMVOC (Non Methane Volatile Organic Compounds). Emissions are given for
11 distinct emission sources (SNAP4 groups, see Table 8). Besides data are also
distinguished concerning to the level, i.e. ground or elevated level (car emissions or
industrial stacks) and thus are separated in two height levels. The CORINAIR database
gives information about yearly total emissions (TONS/YEAR) per SNAP group activity and
administrative unit. These areas have to be spatial and temporal resolved. The NMVOC
have also to be separated.
Table 8: Main SNAP group categorisation
SNAP
no.
Description
1
2
3
4
5
6
7
8
9
10
11
PUBLIC POWER, COGENERATION AND DISTRICT HEATING PLANTS
COMMERCIAL, INSTITUTIONAL AND RESIDENTIAL COMBUSTION PLANTS
INDUSTRIAL COMBUSTION
PRODUCTION PROCESSES
EXTRACTION AND DISTRIBUTION OF FOSSIL FUELS
SOLVENT USE
ROAD TRANSPORT
OTHER MOBILE SOURCES AND MACHINERY
WASTE TREATMENT AND DISPOSAL
AGRICULTURE
NATURE
Diurnal
variation
Group
4
4
3
3
4
2
1
1
2
2
2
Converting emissions of administrative districts to urban emissions
For calculating urban data it is assumed that the emissions are equally spread per head
and therefore the urban emissions can be derived from the district emissions with the
following rule:
Urban emissions  District emissions *
Urban inhabitant s
Inhabitant s of administra tive district
Information about the size, position and population can be found in HEGIS (Health and
Environment Geographical Information System). It contains information about 1742
European cities with more than 50.000 citizens.
Splitting in urban and suburban sector
In OFIS the shape of one city is supposed as circular, the suburban is zonular of the same
size. Emissions of the city are separated in urban and suburban emissions and
correspond to the urban and suburban area, respectively. The relation urban : suburban
emission is typically assumed as 2:1.
4
SNAP: Selected Nomenclature for Air Pollution, nomenclature developed in the frame of CORINAIR to
relate emissions of air pollutants to relevant source sectors, sub-sectors and activities.
- 39 -
Deliverable D05.1
Separation of NMVOC / NOx
CORINAIR provides NMVOC and NOx emissions. For calculating the chemical reactions
these two sum parameters have to be separated into their single constituents. In OFIS,
NOx is separated in NO and NO2 at the ratio of 85 : 15. NMVOC can be itemized as
follows:
Table 9: Typical NMVOC split per SNAP group categorisation
SNAP
Ethane
n-butane
Ethene
Propene
o-xylene
HCHO
CH3CHO
MEK
C2H5OH
CH3OH
ISOPRENE
1
12.97
30.573
16.097
8.545
2.772
1.758
1.933
0.773
22.285
0.387
1.907
2
12.97
30.573
16.097
8.545
2.772
1.758
1.933
0.773
22.285
0.387
1.907
3
12.97
30.573
16.097
8.545
2.772
1.758
1.933
0.773
22.285
0.387
1.907
4
0.146
0.971
1.875
0
0.828
0
0
0.355
93.472
0.079
2.273
5
0
82.01
2.406
10.96
2.385
0
0
0
2.239
0
0
6
5.356
35.988
0
0
27.302
0
0
3.666
19.97
0
7.717
7
4.857
30.871
8.625
6.997
36.759
1.635
1.056
0
8.753
0
0.138
8
5.7
18
12
4.6
10.6
5.9
4
0
39.2
0
0
9
48.413
48.413
0
0
0
3.175
0
0
0
0
0
10
12.97
30.573
16.097
8.545
2.772
1.758
1.933
0.773
22.285
0.387
1.907
11
0
0
0
0
0
0
0
0
0
0
0
Diurnal variation of emissions
To allocate emissions to diurnal variations the 11 SNAP groups are outlined in 4 groups.
The classification results from the plants size (traffic plays a different role). Four diurnal
variations are distinguished: group 1 (traffic; further mobile sources and machines), group
2 (use of solvents; processing of waste material and waste sites; agriculture; nature),
group 3 (industrial combustion; manufacturing processes) and group 4 (power plants;
combustion plants; refineries and spreading of fossil fuel).
Table 10: Typical diurnal variation valid for Germany.
hour
00
01
02
03
04
05
06
07
08
09
10
11
12
13
14
15
16
17
18
19
20
21
22
23
road transport
0.016
0.013
0.008
0.005
0.004
0.006
0.015
0.044
0.070
0.068
0.058
0.053
0.052
0.053
0.053
0.056
0.061
0.069
0.072
0.065
0.051
0.042
0.035
0.028
small stationary
0
0
0
0
0
0
0.066
0.092
0.092
0.066
0
0
0.092
0
0
0
0.066
0.092
0.092
0.092
0.092
0.092
0.066
0
- 40 -
medium stationary
0
0
0
0
0
0
0.063
0.063
0.063
0.063
0.063
0.063
0.063
0.063
0.063
0.063
0.063
0.063
0.063
0.063
0.063
0.063
0
0
large stationary
0.042
0.042
0.042
0.042
0.042
0.042
0.042
0.042
0.042
0.042
0.042
0.042
0.042
0.042
0.042
0.042
0.042
0.042
0.042
0.042
0.042
0.042
0.042
0.042
Deliverable D05.1
The diurnal variation of the emissions for the 4 groups are taken from the AOP1
programme (Air Quality Report of the Auto Oil Programme) for 7 countries (F, FRG, GR, I,
NL, E and UK). For other EU countries similar diurnal variations can are presumed.
Consequentially urban and suburban emissions are yielded concerning to temporal and
also to spatial resolution resolution.
Snapemissi on( snap, i, daytime, k ) 
annual _ emission ( snap, i, k )
* diurnal var iation ( snap, daytime)
365
i : pollutant
k : elevated or ground level
(Sub )urban _ emission (i, daytime , k ) 
 Snapemission (snap, i, daytime , k)
Snap
Rural emissions
As mentioned above CORINAIR uses emission data based on administrative units. In
opposite OFIS covers a square area with the town its centre. This makes the CORINAIR
data scarcely transferable to OFIS. Therefore emission data usually provided by EMEP
are used. These data are available for whole Europe in a grid, each cell 50 km *50 km in
size. A city is regarded in the centre of the middle cell of 9 cells. To calculate the boundary
emissions the emissions caused by the city is subtracted by the overall emission occurred
in these 9 cells. In order to obtain a classification in SNAP groups also for the EMEP data,
these emissions are weighted with the ratio urban emission / overall emission. In analogy
to the previous section, rural emissions can be assigned to diurnal variations.
7 References
Kessler Chr. (1995), Entwicklung eines effizienten Losungsverfahrens zur Beschreibung
der Ausbreitung und chemischen Umwandlung reaktiver Luftschadstoffe, Verlag
Shaker, Aachen, pp. 148.
Sahm, P. and Moussiopoulos, N., (1995), “MUSE - a multilayer dispersion model for
reactive pollutants”, in Air Pollution III (H. Power, N. Moussiopoulos and C.A. Brebbia,
eds), Computational Mechanics Publications, Southampton, Vol. 1, pp. 359-368.
Sahm P. and Moussiopoulos N., (1998), A new approach for assessing ozone exposure
and for evaluating control strategies at the urban scale, Proceedings of Air Pollution
98, Genoa, Italy, September 1998.
Simpson D., Guenther, A.B., Hewitt, C.N. and Steinbrecher, R., (1995), “Biogenic
emissions in Europe 1: Estimates and uncertainties”, J. of Geophys. Res. 100,
22875-22890.
- 41 -