delivering spatial and temporal data on a scale that
better relates to ground-level monitoring.
Design Methodologies for Monitoring
Networks
All monitoring techniques, from measurement of
ambient pollution levels to computer modeling,
rely on some assumptions and estimates in the
assessment of pollutant concentrations, exposure
to pollution, risk, and damages from pollution.
Sophisticated design methodologies for monitoring networks and campaigns have been developed
to ensure that the uncertainties are minimized.
These methodologies help to ensure that the
most useful information can be extracted from
what are often limited data.
Established design methodologies include
kriging, the use of indices of multivariate effectiveness [136], and the location-allocation method
[137]. Each method requires an established area to
be monitored and attempts to optimally cover that
area with a given number of monitoring stations.
Kriging involves placing monitoring stations such
that mean square prediction error is minimized
[137]. The method employing indices of multivariate effectiveness is based on Shannon’s [138]
information theory, as used by Silva and Quiroz
[136] which tests various configurations of monitors to optimize the information content of data
collected. The location-allocation method begins
with the creation of a “demand surface,” a surface
representing the area to be studied, with higher
demand areas represented by peaks in the surface.
Values on the demand surface are increased by
variability in pollution concentrations and by
proximity to sensitive populations [137]. To
inform the demand surface, an initial pollution
surface (representing the distribution of pollution
over an area) must exist based on previous or
ongoing monitoring. Once the demand surface is
finalized, monitoring stations can be placed at the
highest demand locations in the area.
Design methodologies for monitoring networks and sampling campaigns are a relatively
new development, with many previous networks
and campaigns set up in a subjective manner.
However, past campaigns and networks lacking
rigorous design principles have been effective in
that they led to the first generations of information
and significant discoveries relating to air pollution
(and thereby led much of the effective legislation
on air pollution). Further, these campaigns provided information that is necessary for the implementation of some of the newer design
methodologies such as the location-allocation
methodology, and they have brought the state of
air pollution science to the point where monitoring system design methodologies can be created.
However, now that design methodologies have
been established, their routine employment can
improve the information content of monitoring
data and the cost-effectiveness of monitoring
efforts.
Overcoming Oversimplifications in
Monitoring
Often, the practicalities of air pollution monitoring necessitate compromises in aspects such as
increased time resolution, assumptions about population behavior, and assumptions of relative uniformity over areas. As methods become more
sophisticated, some of these assumptions can be
eliminated.
Conventional studies usually assume that the
basis of exposure to air pollution is the concentration in a given location, often outdoors. However,
people move through various exposure situations
in their daily lives, from indoors to “hotspot”
intersections and through areas of differing land
use. A dynamic activity-based population model
has been suggested to develop a more realistic
sense of actual individual exposure [139]. Use of
this model captures exposure resulting from people’s travel in and out of urban areas, rather than
assuming a constant number of people within
cities, as has been the case in conventional studies
[139]. Beckx et al. [139] use of this model showed
that conventional static modeling of exposure is
underestimating population exposure to air
pollution.
Some very small areas in cities, such as high
traffic intersections, can have much higher pollution concentrations than the bulk concentrations
in the wider area being monitored through conventional means. Gokhale and Khare [140]
416
Air Pollution Monitoring and Sustainability
better relates to ground-level monitoring.
Design Methodologies for Monitoring
Networks
All monitoring techniques, from measurement of
ambient pollution levels to computer modeling,
rely on some assumptions and estimates in the
assessment of pollutant concentrations, exposure
to pollution, risk, and damages from pollution.
Sophisticated design methodologies for monitoring networks and campaigns have been developed
to ensure that the uncertainties are minimized.
These methodologies help to ensure that the
most useful information can be extracted from
what are often limited data.
Established design methodologies include
kriging, the use of indices of multivariate effectiveness [136], and the location-allocation method
[137]. Each method requires an established area to
be monitored and attempts to optimally cover that
area with a given number of monitoring stations.
Kriging involves placing monitoring stations such
that mean square prediction error is minimized
[137]. The method employing indices of multivariate effectiveness is based on Shannon’s [138]
information theory, as used by Silva and Quiroz
[136] which tests various configurations of monitors to optimize the information content of data
collected. The location-allocation method begins
with the creation of a “demand surface,” a surface
representing the area to be studied, with higher
demand areas represented by peaks in the surface.
Values on the demand surface are increased by
variability in pollution concentrations and by
proximity to sensitive populations [137]. To
inform the demand surface, an initial pollution
surface (representing the distribution of pollution
over an area) must exist based on previous or
ongoing monitoring. Once the demand surface is
finalized, monitoring stations can be placed at the
highest demand locations in the area.
Design methodologies for monitoring networks and sampling campaigns are a relatively
new development, with many previous networks
and campaigns set up in a subjective manner.
However, past campaigns and networks lacking
rigorous design principles have been effective in
that they led to the first generations of information
and significant discoveries relating to air pollution
(and thereby led much of the effective legislation
on air pollution). Further, these campaigns provided information that is necessary for the implementation of some of the newer design
methodologies such as the location-allocation
methodology, and they have brought the state of
air pollution science to the point where monitoring system design methodologies can be created.
However, now that design methodologies have
been established, their routine employment can
improve the information content of monitoring
data and the cost-effectiveness of monitoring
efforts.
Overcoming Oversimplifications in
Monitoring
Often, the practicalities of air pollution monitoring necessitate compromises in aspects such as
increased time resolution, assumptions about population behavior, and assumptions of relative uniformity over areas. As methods become more
sophisticated, some of these assumptions can be
eliminated.
Conventional studies usually assume that the
basis of exposure to air pollution is the concentration in a given location, often outdoors. However,
people move through various exposure situations
in their daily lives, from indoors to “hotspot”
intersections and through areas of differing land
use. A dynamic activity-based population model
has been suggested to develop a more realistic
sense of actual individual exposure [139]. Use of
this model captures exposure resulting from people’s travel in and out of urban areas, rather than
assuming a constant number of people within
cities, as has been the case in conventional studies
[139]. Beckx et al. [139] use of this model showed
that conventional static modeling of exposure is
underestimating population exposure to air
pollution.
Some very small areas in cities, such as high
traffic intersections, can have much higher pollution concentrations than the bulk concentrations
in the wider area being monitored through conventional means. Gokhale and Khare [140]
416
Air Pollution Monitoring and Sustainability