|
There are no
generally valid rules for network design. It is determined mainly by the
overall monitoring objectives and resource availability. Although monitoring
systems can have just a single, specific objective, it is more common for them
to have a broad range of targeted programme functions. No network design can hope
to completely address all the possible monitoring objectives listed below:
Monitoring Objectives
Determining compliance with national or EU limit
values/standards
·
Determining
population exposure and health impact assessment;
·
Informing the
public about air quality and raising awareness;
·
Identifying
threats to natural ecosystems;
·
Providing
objective inputs to air quality management, traffic and land-use planning;
·
Source
apportionment and identification;
·
Policy
development and prioritization of management actions;
·
Development/validation
of management tools (models, Geographical Information Systems etc.);
·
Assessing
point or area source impacts and
·
Trend
qualification, to identify future problems or progress against
management/control targets.
Network
design
The design of
the air quality monitoring network basically involves determining the number of
stations and their location, and monitoring methods, with a view to the
objectives, costs and available resources. (See Larssen, 1998 in the
Further reading list).
There are two main
basic approaches to determine the number of stations and locations: to
locate stations in a regular geometric grid covering the city. The grid size
thus determines the number of stations and to locate stations at sites
considered to be representative for more defined local environments, exposure
situations or source activities, such as urban background. While the first
approach was used earlier, e.g. in Germany, the latter approach is now in more
general use, and also the one prescribed by the Directives.
The typical
approach to network design, appropriate over city-wide or national scale, thus
involves sitting monitoring stations or sampling points at carefully selected
representative locations, chosen on the basis of required data and known
emission/dispersion patterns of the pollutants under study. This approach to
network design requires considerably fewer sites than grid strategies and is,
in consequence, cheaper to implement. However, sites must be carefully selected
if measured data are to be useful. Moreover, modelling and other objective
assessment techniques may need to be utilized to ‘’fill in the gaps’’ in any
such monitoring strategy.
Another
consideration in the basic approach to network design is the scale of the
air pollution problem:
·
The air
pollution is of predominantly local origin. The network is then
concentrated to within the urban area. Example: CO and benzene.
·
There is a
significant regional contribution to
the problem. More emphasis then on the regional part. Example: ozone, PM.
·
Large scale
phenomena, such as winter smog
episodes in NW Europe, or photochemical pollution episodes in the
Mediterranean. Even more emphasis on the regional part of the network.
The number
of sites depends of course upon the size and topography of the urban area,
the complexity of the source mix and again upon the monitoring objectives. The
Directives specify a minimum number of stations to be established dependent
upon the population, and it also indicates what types of areas should be monitored
(representing average as well as hot-spot exposure situations).
Some time
should be invested into determining the number and location, to ensure that the
network, which will normally be established to be operated over a long period
(many years), will serve its purpose most effectively. A basic procedure of
several steps should be followed:
·
Start with a
map showing main pollution related features such as urban central district,
residential areas, areas of dense traffic, the main road network, large industrial
plants and areas;
·
Use a
(preliminary) emissions inventory as a support to find the most polluted areas;
·
Carry out
preliminary dispersion modelling to identify polluted areas;
·
Carry out
surveys using inexpensive methods, such as passive samplers;
·
Consider that
different pollutants have different spatial scales of variability (e.g. CO
concentrated near streets; NO2 ozone and PM more evenly
distributed).
The further
process of location and number is subjective. Some guidance is given e.g. in
the Guidance
on Assessment under the EU Air Quality Directives report, and in the UK
Technical Guidance Documents (see Further reading list below).
The site
classification scheme used by the Commission is a guide that should be used
in the network design, so that any stations can be classed according to that
scheme, and such that the network covers as many as possible of the station
classes:
·
Level 1: Type
of station: traffic, industrial, background;
·
Level 2: Type
of area: urban, suburban, rural.
(See the EC Exchange of Information
(EoI) Decision: Council
Decision 97/101/EC, and amendment: Commision
Decision 2001/752/EC ).
Some countries
have developed station classification schemes of their own, following the same
basic principles as above, but deviating somewhat, e.g. UK and France (see
examples and links in the further reading section below).
Each station
should be described in terms of meta data, which includes data such as
coordinates, type, specific location area of representativeness, additional
data such as traffic data, etc. (see also EoI as above, and its guidance
document: Guidance
report on the Annexes to Decision 97/101/EC ).
Monitoring
involves assessing pollutant behaviour in both space and time. A good network
design should therefore seek to optimise both spatial and temporal coverage,
within available resource constraints.
The first
target is to maximizing spatial coverage and obtaining representative
measurements. Once priority pollutants are selected, the sampling methods must
be capable of a time resolution consistent with the pollutant averaging times
specified in guidelines.
The compounds
to be measured and the reference methods used are prescribed by the
Directives.
An air quality
monitoring network must, in addition to the air pollution monitoring part, also
comprise a meteorology (dispersion parameter) monitoring part.
The
meteorological data are needed for at least two reasons:
·
For the
interpretation of the temporal and spatial variation of the data from the air
quality monitoring, there is an obvious need for meteorological data: wind
speed and direction; parameters describing atmospheric turbulence and
stability, such as temperature profiles (measurements at two or more heights),
or direct turbulence measurements; mixing height; and ground air temperature.
·
The
meteorological data should provide hourly spatial fields of the
meteorological/dispersion parameters, either by interpolation, or using a
wind-field model. The calculation of dispersion parameters from the
meteorological parameter measurements to be used in the dispersion models,
usually require the use of a meteorological pre-processor.
Methods
Continuously
operating automatic analysers may be used to assess compliance with short- or
long-term guidelines. Well-recognised semi-automatic methods such as
acidimetric SO2 samplers, will be perfectly adequate for measurement
against daily standards or criteria. For automatic analysers or samplers to
reliably measure ambient pollutant concentrations, it is essential that these
pollutants are transferred unchanged to the instrument reaction cell. The
sampling manifold is a crucial and often overlooked component of any monitoring
system, which strongly influences the overall accuracy and credibility of all
the measurements made.
Integrating measurement methods
such as passive samplers, although fundamentally limited in their time
resolution, are useful for the assessment of long-term exposure, as well as
being invaluable for a variety of area-screening, mapping and network design
functions. Problems can arise, however, when using manual sampling methods in
an intermittent, mobile or random deployment strategy. Such an approach is
usually adopted for operational or instrumentation reasons, or simply because
it would not be possible to analyse the sample numbers or data produced by
continuous operation. Intermittent sampling is still widely used world-wide.
However, this sampling strategy may be of limited utility in assessing diurnal,
seasonal or annual pollutant patterns or, indeed, for a reliable assessment of population
exposure patterns. |