EFFECT:
Local Air Quality Prediction and Effective Traffic Demand Management

Nick Hodges
Special Projects Officer, Leicester City Council
3rd Floor, York House, 91 Granby Street, Leicester, United Kingdom, LE1 6FB.
Email: nhodges@lccatc.demon.co.uk

INTRODUCTION

EFFECT is a project funded under the European Union's Fourth Framework Programme. It stands for "Environmental Forecasting For the Effective Control of Traffic". The project commenced on 1st January 1996 and will finish on 31st December 1998. The partner cities in the project are Leicester (UK), Maidstone (UK), Gothenburg (SE) and Volos (GR). A project web site is based at http://www.effect.gr

Each of these cities were chosen because of the availability of particular expertise, available infrastructure and contrasting conditions. With a population of 120,000, Volos is Greece's third largest port and is situated between its sea and mountains. These factors, combined with the traffic conditions, have given rise to concerns over air quality, resulting in a political desire to improve the situation. Volos' Mediterranean climate is in stark contrast to the Scandinavian city of Gothenburg with its population of 430,000. Gothenburg, one of the leading cities in the field of air quality management, also hosts the ARENA transport/IT test site. Leicester has a long history of using state-of-the-art traffic control systems and has received political support for implementing management strategies within the 470,000 population conurbation. Maidstone, with its 120,000 population, also has an advanced traffic control system and is currently implementing a county-wide strategic pollution monitoring system under the LIFE funded MIST Programme.

OBJECTIVES

The objective of the project is to predict poor local air quality in real-time and then to instigate effective traffic demand management strategies (TDMS) to reduce pollution levels in particular problem areas. This is being achieved through the implementation of local Environmental Management Boards (EMB - a consultation forum for assessing the acceptance of alternative strategies), and the innovative integration of air quality modelling with real-time information on traffic flows, pollutant concentrations and meteorological conditions to highlight pollution "hotspots." Each of the partner cities has a different emphasis within the project and acts as either a development (validation) or demonstration site.

The overall system comprises a number of modules, as shown in Figure 1.

figure 1

Figure 1: Overview of the EFFECT system

The following list describes the components being developed or demonstrated at each site:

The availability of traffic data and air quality modelling lies at the heart of the system. In the two UK sites, the traffic data is obtained from the demand-responsive SCOOT (Split, Cycle and Offset Optimisation Technique) urban traffic control system. In Gothenburg, the ARENA traffic monitoring and modelling system is used. In Volos, a fixed-time urban traffic control system is used in conjunction with flow counters.

The air quality modelling in all four sites is being carried out using the AIRVIRO system marketed by SMHI in Sweden. The Eulerian Grid Model is used for the prediction of gaseous particle dispersion, except in Volos where the Gaussian Model is deployed. A Street Canyon model is also available. The EFFECT system architecture will allow the use of any other similar air quality model.

THE AIRVIRO SYSTEM

At the heart of the system is a computerised map containing the topography and land-use across the area. Next, AIRVIRO requires information about the major sources of air pollution in the area. These fall into three main categories - point sources, area sources and line/road sources.

Point sources include factories, offices, shops, and large residential blocks where all emissions are concentrated in a small area, such as a chimney or stack. Dynamic and static information is collected for each source. Dynamic information includes formulae that describe the emission as a function of outdoor temperature or as a function of day, week and time. Static information covering chimney height, exhaust gas temperature, coordinates, etc. describes the exact location of the source, but does not affect the level of emissions being produced. In Leicester, the Pollution Control Group had a high response rate to questionnaires that were sent to factories and companies with chimneys that they felt produced significant emissions.

Area sources are used to describe emissions which are assumed to be evenly spread over a wide area. These are usually housing or industrial estates where it would not be practical to include every chimney as an individual point source. The information needed is similar to that for point sources. The Energy Team in Leicester City Council had access to information about typical domestic heating sources in housing estates and seasonal fuel consumption figures, and was then able to give an indication of emissions according to weather conditions.

The information required to define road sources includes the physical location of each road and the flow profile and classification of the vehicles on each road. This information can be obtained from a traffic model or from direct observations. Emission factors for the vehicles (idling, accelerating, cruising and decelerating) are also needed, and were obtained from relevant research literature. In Gothenburg, the coastal sea lane is also an important line source.

The road network data was obtained from a TRIPS traffic model in Leicester, a CONTRAM model in Gothenburg, and a SATURN model in Maidstone and Volos. The traffic model gave the information on the location of the roads, the number of vehicles in the peak and off-peak periods, and the speed limit of the roads. The roads were then (manually) assigned a particular road-type, e.g. inbound radial. Each road-type was then defined in terms of the typical daily and seasonal flow profile. This was based on actual traffic counts from representative roads. Local vehicle registration details were obtained and a typical vehicle fleet assigned to the road-type. The vehicle fleet definition took into account the type of vehicle and the fuel used. Standard emission rate information for the different vehicle types was then applied.

Once the AIRVIRO system was established, it was then possible to carry out air quality modelling to assess the effect of the various traffic demand management strategies developed in the project. However, it is not possible to be totally confident about the accuracy of the concentration figures produced unless some air quality and meteorological monitoring is also undertaken.

Air Quality and Meteorological Monitoring

As stated above, it is also advisable to collect and store real-time air quality and meteorological data so that predictions can be compared with actual levels. The four sites in the EFFECT project use the AIRVIRO system to perform this function. There is a wide range of monitoring equipment used at project sites. The air quality monitors fall into two categories - high precision, high cost sensors in urban background sites, and low precision, low cost sensors deployed at roadside sites. The former cost about ECU 10,000 per pollutant monitored, whilst the latter, also at ECU 10,000 will pay for the monitoring of both CO and NO2 by the latter. A PM10 component is being developed. It is at the roadside that levels are at their highest, and average factors of over three (roadside versus background) for CO and NO2 have been observed.

Meteorological monitoring is also advisable, as weather conditions are well known to have an important impact on air quality. A meteorological mast was used measuring vertical and horizontal wind speed and direction (at 12m) together with air temperature (at 2m and 10m to assess vertical air movement). As a result, a database of local weather conditions is built up. It is possible to carry out air quality predictions for various weather scenarios.

Traffic Demand Management Strategies (TDMS)

As stated previously, the traffic demand management strategies are being developed using the SATURN traffic model in conjunction with AIRVIRO. They will be demonstrated in Maidstone and will also be transferred to and demonstrated in Volos. The EFFECT PLUS project is providing TDMS for Leicester, whilst Gothenburg will use ARENA with EFFECT for TDMS. These strategies may include holding traffic outside the city, metering, gating, advisory diversions or closing roads. Information will be given to the drivers using tools such as Variable Message Signs, radio broadcasts using RDS/EON (radio data system/enhanced other network) technology and PROMISE 2 terminals.

The appropriate strategy will be selected by the traffic control system operator based on the available air quality modeling information and their own experience of traffic conditions in the network. The effectiveness of the implemented strategy can then be assessed using modeling if monitoring has also been carried out.

Real-time Links Between Traffic Monitoring and Air Quality Modeling

It was stated above that the AIRVIRO system is given information about the average daily and seasonal flow profiles of the traffic. However this information is only as good and up-to-date as the model that generated it. It would clearly be better if the traffic data were continuously updated to reflect on-street conditions "now." Consequently, one aim of the EFFECT project was to provide a link between the traffic control system and the AIRVIRO system. This link was developed for the Leicester SCOOT system and then transferred to Maidstone and Gothenburg for demonstration. Thus, instead of using the modelled values, AIRVIRO can produce air quality predictions that are based on actual traffic levels. The results of these predictions are updated and displayed every hour. Work on the development of live traffic models is proceeding in Gothenburg with the CLEOPATRA project, and in the UK with the UTMC R&D programme to enable alternative TDMS's to be assessed.

The EMMA/EFFECT Cycle

In Leicester, the EFFECT project is being carried out in conjunction with another EU-funded project called EMMA. This stands for "Integrated Environmental Monitoring, Forecasting and Warning Systems in Metropolitan Areas" (see also presentation paper given in the "Air Quality Management" section of these Proceedings). The focus of EMMA is on predicting air quality and providing information to the public as well as to network managers. The UK Meteorological Office provides 60-hour weather forecasts direct to the AIRVIRO database to enable 24 and 48 hour air quality forecasts to be made. This tie-in with EFFECT means that if the air quality conditions are known for the following day, then appropriate demand management strategies can be selected to reduce the impact of a potential air quality problems.

EMMA uses the weather forecast together with the enhanced AIRVIRO database to forecast tomorrow's pollution levels. The forecast is used by the network manager to select an appropriate TDMS, including regulatory restrictions (e.g. the closure of roads, such as occured in Paris in October 1997). The forecast is also broadcast by radio during the afternoon and evening when people return home. Given this information, they can choose how or when they will travel into the city the following day. In addition, the forecasts are available on the Internet at http://www.mdx.ac.uk/emma. The next morning, EFFECT will predict the current pollution levels, allowing the network manager to review the TDMS for the afternoon if necessary. In January 1998, a particulates episode was predicted, and radio broadcasts and variable message signs were used to advise travellers. The following day, public reaction was monitored. Surveys were carried out on the A47 Leicester Environmental Road Tolling Scheme corridor and in the city centre using both on-street and telephone questionnaires. The surveys revealed that over 30 percent of those questioned were aware of the problem, with 14 percent taking some action. Over 50 percent were already committed public transport users. Figure 2 describes the overall process which rolls forward from day to day.

figure 2

Figure 2: The EFFECT/EMMA cycle

DISCUSSION

Appendix I seeks to summarise the key elements described above. Level A illustrates that the map grid locates data within the emissions database which also takes account daily and seasonal variations in the rates of emissions, e.g. am/pm peak traffic flows; daily and seasonal heating patterns. Level B shows that the air circulation patterns and the differing effects of surface roughness (open space, estate or street canyon) are taken into account when the Eulerian Grid or Gaussian Model dispersion calculations are made. The meteorological data allows the model to assess the mixing height (crucial when temperature inversion is anticipated), which controls the envelope within which the dispersion takes place. When this mixing height varies during the day, pollution concentration levels will respond by decreasing or increasing. A strong weather front will often remove much of the airborne pollutants by wind or rain. Level C emphasises that each initial Air Quality Map produced by the models shows the relative concentrations of modelled pollutants at rooftop level (near ground level over open spaces). In the city centre, the street canyon effect has to be considered, to assess the likely concentrations at pavement level (these can be several times greater by, say, a factor of three). With time, the database of background information will allow the model to be calibrated to provide predictions of pollution levels, rather than just comparisons between levels. Research suggests that NO2 levels increase with vehicle speed, whilst CO generally reduces until a critical speed is reached, after which there is a small increase. PM10 appears to follow the CO pattern and produces many of the Winter episodes. The chemical reactions between NO2 and O3 affect the ratio of these gases and the total levels vary throughout the day. Background monitoring can assist in prediction. Air quality models such as ADMS, EMMA, NILU and REGOZON can provide predictions for Ozone. Since each pollutant has differing effects on health, the network manager has many competing factors to reconcile when selecting a traffic demand strategy. Currently, many opt to reduce the total pollution level.

Both short and long-term TDMS techniques are being explored. These include raising public awareness; provision of pedestrian and cycle networks; 'woonerf' protected housing areas and traffic calming (to slow or divert traffic); priority for peak traffic flows (am/pm); metering and gating of traffic flows with or without bus priorities (which relocate queues and then control the flow of traffic); corridor or area control, or linking of traffic signal operation; 'park & ride' by bus or rail; pedestrianisation; controlled parking or access zones; and advisory diversions using variable message signs.

CONCLUSION

This paper has outlined some of the key factors involved in the monitoring and prediction of local air quality to provide effective traffic demand management. Clearly the extent to which TDMS can be implemented depends on the existing conditions (air quality and traffic congestion, resources available etc). However, a useful start can be made by installing a basic air quality monitoring network together with a simple local meteorological mast. These will lay the foundations for future development, whilst also assisting in the raising of awareness among the public and network managers. Adding a simple air quality model will allow basic "what if" analyses to be made. As an understanding of the interaction of the various factors is developed, then the system can be enhanced using the EFFECT and EMMA concepts to provide more sophisticated tools. Whilst precision measuring instruments may accurately record conditions at a measuring station, one must appreciate that high-precision accuracy in predicting air quality levels elsewhere is not achievable no matter how much is invested. In the end, the information has to be tailored to the needs of the end-user, which generally requires considerable simplification (i.e. radio or television broadcasts will offer Good, Poor or Bad). However, the work in the EFFECT and EMMA projects has indicated that worthwhile improvements in air quality can be achieved by proper management of the city environment.

REFERENCES

"AIRVIRO Users Manual", Swedish Meteorological and Hydrological Institute.

"EFFECT", project proposal to European Union.

"EMMA" Air Quality Monitoring Handbook

APPENDIX I

figure 3

REC * PROGRAMS * TELEMATICS * DETERMINE

HOME PAGESEARCH