International Union of Air Pollution Prevention and Environmental Protection Associations (IUAPPA)
Symposium and Korean Society for Atmospheric Environment (KOSAE) Symposium, 12th World Clean Air
& Environment Congress, Greening the New Millennium, 26-31 August 2001, Séoul, Korea.




URBAN FABRIC AND MEASURES OF VARIABILITY: NEIGHBORHOOD
EFFECTS ON PROXIMITY POLLUTION




C. WEBER1, J. HIRSCH1, L. SCHNELL1, G. PERRON2, J. KLEINPETER2, T. RANCHIN3

1 Laboratoire Image et Ville (Upres-A 7011 CNRS) - Uni. Louis Pasteur, Strasbourg, France. Faculté de
Géographie ­ Uni. Louis Pasteur, 3 rue de l'Argonne, 67 000 Strasbourg. FRANCE
2 ASPA, Strasbourg, France
3 ARMINES / Ecole des Mines de Paris, Groupe Télédétection & Modélisation, Sophia Antipolis




1.INTRODUCTION

This paper reflects the work achieved for demonstrating the potential influences of geometric features on
atmospheric pollutant distribution over metropolitan's areas. Actual methods seldom take into account
morphological aspects and structural characteristics of city of interest. Hence, influences of urban forms on
pollutant distribution are presently investigated in urban atmospheric pollution researches at various scales,
from local to global urban area.
In fact this multi-scale problem is generally studied on a single scale (local, meso or global) but the total
effect of streets network and urban units together are rarely mixed. The interaction of several scales of
pollutant observations (points-measurements, close surrounding area, streets and global urban area) is a
particularity of the present work. This project is applied to the city of Strasbourg, France. Its objectives are to:
·
precisely define the morphological elements which influence the pollutant distribution;
·
spatialize the measurements with the help of satellite data, taking benefits from previous results.

Therefore this research project explores:
·
the combination of different scales (from points-measurements to the complete urban community of
Strasbourg);
·
different methods (mathematical morphology, spatial analysis, interpolation, extrapolation, ...);
·
different modeling approaches (from a single street to the total urban area);
·
and different geographical and pollutants data sources (points-measurements, geographical databases,
remotely sensed data).

The benefits of the decreasing SO2 and NO2 values since several years in European cities are erased by an
overpass of the general norms fixed by the EU15. 37 millions of inhabitants are daily exposed to high
pollutants values in European cities (among those which are equipped with measurement network) [1]. Such
situations are not only representative of European countries, developing countries are also facing this
problem. The climatic urban profile must be more analyzed than previously, in order to get not only a precise
but located understanding over time but also to define complementary spatialized measurements campaign

-2-




events. The collected data would enrich the knowledge gathered over different observation scales: local to
global. Urban pollutant distribution needs rather adapted observation scales allowing studying internal
variability. Most of the transportation and distribution pollutant models are developed according a 1*1km
grid, so it seems to be worthwhile to precise explanation elements (land cover, built density, etc.) at a rather
large scale to enhance these results.

This research program deals with a better understanding of pollutants distribution processes at different
scales. The design of some quantitative indicators on measurements spots network is presented. They are
developed for comparison purpose and variability assessments in urban transportation pollutant studies. Thus,
they are exploited for the extraction of parameter inputs in the models of pollutant distribution, the
spatialization of collected data using remotely sensed data and the representation of pollution events
according to public and legal obligations


2.AIR QUALITY ASSUMPTION AND STRASBOURG CASE STUDY

Several studies propose to analyze transportation and pollutants distribution processes in a global
understanding rather than to consider only pollution heights [2]. It implies to observe the daily exposure to
pollution through a set of indicators allowing to enhance knowledge and estimation of urban pollution level
and to disseminate results throughout populations and politic decision-makers. This study deals with such
goals, observing air quality through the interactions between distribution processes and urban forms.

The new European directive on Air quality promotes control and management of crisis situation in case of
overlapping values of pollutants (30/12/1996) regarding European norms. Information has to be dispatched to
alert the population in case of such situation and to stop or to push off course traffic transportation. To handle
this situation, two objectives must be followed: the pollutant measurements must be defined at different
scales and the pollutant distribution has to be the spatialized. It means that over the urban area, we might be
in position to locate pollution peaks, and also to represent areas that are exposed to a too high level of
pollutant particles. Such delimitation of urban area is uneasy and rather delicate toward public authorities.

Complementarity of embedded scales of observation is obvious from the measurement spot to the global
urban area, but the relationships between these levels are complex and insufficiently understood. Several
problems rise up when dealing such questions, such as the measurements spots network. The representativity
of these measurement spots (urban, industrial, rural, traffic one, etc.) needs to be assessed according national
or international typologies. It has to be noticed that such typologies change over time, over countries and
even over region of the same country. The relationships between the measurement spots (its characteristics),
and the interpretation accorded to the measurement values collected, may change if some variability is
introduced. In this study, we cope with the observation of this representativity and we try to gather qualitative
and quantitative indicators able to describe the measurement spots.

In order to spatialize the values collected through the irregular measurement network, it is necessary to
perform a spatial representation of the collected values. To cope with this task, remotely sensed data are used
in order to guide this realization. Based on the relationships between pollutant and dust particles and the
spectral values of satellite images, previous studies [3] [4] [5] [6], have demonstrated that pollutants
spatialization might be approached by this way. This solution needs several indicators to perform the
relationships study, for instance collected values of course and also quantitative indicators on the close
surrounding area of the different measurement spots. In fact it is necessary to get a maximum of information
on this area in order to match with the spatial resolution of the satellite image used (Landsat TM, 30m*30m).
Thus the more information collected around measure locations, the more the possibility to understand these
specific relationships.



C. Weber

-3-




Several questions appear when dealing with those goals:
· Could some indicators be attached to measurement spots in order to get comparison values in a
quantitative format?
·
Could we observe specific behavior among the different spots?
·
Are they some impacts of the geometry of the measurement areas on the collected values?

Such results assume to provide some information on feasible urban variability at a large scale. The
relationships between spots, streets and general area might be more understandable with such explanatory
elements. Different types of indicators might be listed for instance: the class of measurement spot, the surface
of the surrounding area, its land-cover characterization, the volume of the spot according to the heights
around, the number of streets linked with the spot, the density of built surfaces, the density of population
concerned by the spot and consequently by the pollutant values collected...

The city of Strasbourg is located in Eastern France at the border of Germany. L'Association pour la
Surveillance et l'étude de la Pollution Atmosphérique en Alsace (ASPA) is the local organization in charge of
the air quality measuring network in the city of Strasbourg and vicinity. The case of Strasbourg-France
(306km2) is interesting because a large set of measurement spots is spread over the area. 14 spots are listed.
The measurement network is rather old and provides time series data since about 1980 according to national
typology requirements (figure 1). All the spots collect large or peculiar data sets.

Figure 1: Typology of the Pollutants and particles collected by the measurement spots




3. STEPS OF STUDY AND COLLECTED DATA




C. Weber

-4-




The different steps developed in this study are complementary. The general process articulates observation
and spatialization and anticipates the inputs/outputs the results might provide.

These steps focuse mainly on the measurement spots description liable to allow comparisons. The use of the
third dimension to characterize built area and design geometric indicators introduces urban forms in the
approach. The hypothesis here, is that at a micro scale, forms have some influence on urban climatic profiles.
Data sets collected (table 1) are numerous describing measurement spots, buildings and streets, socio-
economic data (population, density of population), relations with streets (removal of traffic source). It has to
be noticed that the streets have been studied also, but the results will not be developed here. Each building is
described by the location coordinates (x, y) and the z value (height) in a geographical database (BD TOPO©
IGN). The points-measurements sites were qualified according to morphological indicators and a
representativity indicator of the sites defined. The geographical database was also exploited for the extraction
of parameters as inputs of the models of pollutants transport [7].

The capacities of the GIS software (Arcview© 3D analyst) allow the obtaining of a 3D representation of the
measurement areas and also performing of several spatial analyses over buildings or heights for instance. This
is obtained by extrusion of the volumes through a hybrid process able to dissociate two buildings placed side
by side.

Table 1 - Measurement spots indicators

Measurement data
Close surrounding
area
Buildings
Socio-
economic data
Street data

Meteorological
Surface
Coordinates
Population
Distance to
Pollutant data
Volume
Height
Density
Traffic
Land-cover
Orientation

4.RESULTS

Performing GIS analyses provide different indicators about measurement spots. These indicators will be used
for two purposes: (1) the study of the representativity of measurement spots and its close surrounding area;
(2) the extraction of parameters for " virtual station "[8] in relation with the satellite data in order to describe
as precisely as feasible the location and the associate spectral values. They have to be defined according
comparison criteria among the spots of the same city but also regarding networks of other cities.


4.1Spot measurements definitions

Two main types have been developed relying on usual collected values and geometry of urban forms over
three " urban measurement spots ". The first is associated to the general behavior of the collected date,
meteorological data and pollutant information. They provide an idea of the variability of the spots values. The
second deals more with geometric information due to urban forms and heights use. The three urban spots
(Kleber, STG centre, STG est) have been chosen for the characteristics of the location (near the center of the
city) and the specificities of the urban forms of the close surrounding area.
4.1.1
Urban spots descriptors
The meteorological trends provide general behavior of the measurements spots, their nycthemeral and
seasonal rhythms. The land-cover data extracted from remote sensing data processing give information on the
land-cover/use types associated with the measures location and allow considering the associate spectral
values.
4.1.2 Geometric descriptors



C. Weber

-5-




In order to describe the close surrounding area and the possible relationships between forms and measured
values, we have defined a peculiar surface (S) corresponding to the " visible " area from the measurement
location and the total surfaces of the houses' fronts (f) concerned by the surface (f) [9]. Other information
are collected like the number of streets connected and type of street. This type depends on the number of
lanes, the height of buildings related to the width of the street, the type of buildings, etc. Another area has
been taken into account, defined as the exposed population density inside a 1km radius around the location
(A). This area have been divided into five embedded circles, each of them being able to allow extraction of
data on building densities per distance, heights of buildings per distance and sectoral information (figure 2).

Figure 2: Geometric information on close surrounding area




Volumes have been also used to describe the measurement spots. In particular they allow extraction of
comparison features enabling to look at the difference between a stuffy volume or on the contrary an open
one. Using a reference volume (VR) = (Hmoy*A) where Hmoy is the mean height of the surrounding
buildings) it is possible to obtain several indicators linked to urban forms (figure 3), like:
(V) related to S (Hmoy*S) and the form rate (TV = V/VR*100). These descriptors provide quantitative
features that complete the observations on measurement spots.

Figure 3: Volumetric descriptors




5.DISCUSSION AND CONCLUSION



C. Weber

-6-




This study is still on going and must be developed on various directions: the interactions between measure
location, surrounding area, street and global urban area. We have set down that some relationships might be
identified and described in order to improve our knowledge on pollutant transportation and distribution. We
have extracted several indicators that can be used to characterize measurement spots and the close
surrounding area, in different ways. This approach has to be developed to assess measurement spots
representativity inside a network but also through different networks. These indicators are also useful to
reinforce the relationships between locations and spectral values due remotely sensed data.

6. ACKNOWLEDGEMENTS

This project is supported by the French Ministry of Research through the research program Action Concertée
Incitative Ville.


REFERENCES

[1]
European Environmental Agency , 1997, Towards sustainable development for local authorities.
Approaches, experiences and sources. 65 p. + annexes
[2]
Institut de Veille Sanitaire, 1999, Surveillance Epidémiologique Air& Santé. 19 p.
[3]
Basly L. and Wald L. Remote Sensing and Air Quality in Urban Area. In : Proceedings of the 2nd
international workshop on telegeoprocessing, TeleGeo'2000, Sophia Antipolis, France, May 10-12,
2000, pp. 213-219.
[4]
Basly, L. Télédétection pour la qualité de l'air en milieu urbain. Thèse de Doctorat en Sciences et
Technologies de l'Information et de la Communication, mention " Propagation, Télécommunications,
Télédétection ", Université Nice ­ Sophia Antipolis, 2000, 186 p.
[5]
Wald L. and Baleynaud J.-M. Observing air quality over the city of Nantes by means of Landsat
thermal infrared data. International Journal of Remote Sensing, 20, 5, 1999, pp. 947-959.
[6]
Basly L., Wald L. and Baleynaud J.-M. Satellite mapping for urban air pollution. In : Proceedings,
EARSeL Symposium 1998 "operational remote sensing for sustainable development", Enschede, The
Netherlands, Nieuwenhuis G., Vaughan R., Molenaar M. eds, Balkema, Rotterdam, 1999, pp. 133-
139.
[7]
Schnell L., 2000, Mise en place et évaluation d'une base de données 3D sur l'emprise de la
communauté urbaine de Strasbourg. Rapport de stage DESS ATS. ULP. 60 p. et annexes
[8]
Weber C., Hirsch J., Perron G., Kleinpeter J., Ranchin T. , Ung A., and Wald L., 2001 Urban
morphology, remote sensing and pollutants distribution: an application to the city of Strasbourg,
France. To appear In Proceedings of the 12th World Clean Air & Environment and IUAPPA World
Congress, Seoul, Korea, 21-31 August 2001.
[9]
Durrenberger M., 2000, La qualification des sites de mesures de la pollution de l'air dans la
communauté urbaine de Strasbourg (CUS). Mémoire de Maîtrise de géographie. ULP. 80 p.




C. Weber

-7-




C. Weber