Title of Example

Example

In Bristol the emissions database has been an evolutionary process. In 1995 an organisation called London Research Centre was commissioned by the UK government to compile emissions inventories for a number of the key cities in the country. Information was provided for a number of pollutants in a document which contained maps of Bristol covered in 5 KM squares. The more useful part of the process was the raw data provided on a spreadsheet of all the industrial sources. There were also data on the road sources but these were not used initially.

The inventory included emissions from all identifiable sources for six of the eight pollutants covered by the UK National Strategy, together with carbon dioxide and non-methane volatile organic compounds.

The eight pollutants were:

Oxides of nitrogen (NOx)

Sulphur dioxide (SO2)

Carbon monoxide (CO)

Carbon dioxide (CO2)

Non-methane volatile organic compounds (NMVOC)

Benzene

1,3-butadiene

Particulate matter < 10 µm in diameter (PM10)

Roads

In the early days traffic data were available for the whole of Bristol and an area surrounding the city from a traffic model database. There were also manual count data available from about 50 sites at the boundary of an inner, middle and outer cordon. These counts included 10 different vehicle types ranging from cars and motorbikes to multiple axle heavy goods vehicles.

The manual counts had the advantage that they were fairly accurate but had the disadvantage that they were only done at each site one day a year between 7:00 and 19:00 on a weekday.

The model data had the advantage that it covered the whole road network but the data were based on surveys, which had been conducted a few years earlier and were factored up for each year. It only calculated total traffic peak time traffic and did not distinguish between vehicle types.

Another problem with the traffic model was that the mapped links were in straight lines so a large amount of re-digitising of the roads had to be done for the air quality modelling.

Further analysis showed that the manual counts and the model data were often very different at specific locations. Air quality modelling with the traffic modelled

data showed an under prediction by about 1/3 when compared with the real time monitored data from the air quality monitoring sites around the city. For this reason the traffic model data were abandoned.

After the year 2000 more automatic counters were being installed round the city. Most of these were only able to count total volume of traffic whereas the manual counters could count specific vehicle types. However, the automatic counters were able to give an indication of the 24-hour variations and also the seasonal variations, which were important for the modelling process.

As the traffic counts were mostly on the major roads the minor roads were calculated and estimated as grids.

The road network from the traffic model showing the links as straight lines.

The early air quality models produced used two vehicle, heavy and light and one road type. As better data became available more vehicle types could be modelled and the roads were categorised to describe the hierarchy. Each road type could be modelled with a different proportion of vehicle types.

In the UK there have been several different sets of emission factors for vehicles. Some of the simpler versions gave total annual estimates for certain vehicle types. When using advanced dispersion models it is necessary to have speed related emissions factors for the range of vehicle types modelled. There were later developments of speed related emissions factors for a range of different vehicle types ranging from motorbikes to heavy goods vehicles for the main pollutants of concern, oxides of nitrogen (NOx) and particulate matter (PM10).

Other sets of emission factors have also included other pollutant gases such as VOC and CO2. The emission factors also have predictions of emissions from the different vehicle types in future years which is useful for running models for future years.

For modelling traffic a combination of all data sources were used. Traffic flow for different vehicle types used the manual count data as they were more accurate. The daily variation was calculated using the automated counts from the few sites, which had them, as this gave a 24-hour variation. This also gave annual variations which is also important as the school holiday season can have a noticeable drop in traffic flow in the city.

The road network used for the air quality models after re-digitising to the correct location and categorising into road types.

Industrial Emissions

The data from the London Research Centre were used initially for the point sources. Many of these were the larger industrial process but also some smaller less significant sources such as occasional oil burning incinerators at petrol stations were included.

The data compiled by the LRC was often indicating the maximum emissions permitted by the regulatory authorities and not necessarily the emissions actually released.

Regular meetings were set up with the regional office of the UK Environment Agency who are responsible for regulation the larger industrial processes. In this way more accurate data on the emissions from the individual process could be obtained and notification received when things changed. This meant that the emissions inventory could be updated regularly.

Any processes which emitted substances that were not required to be modelled by the Air Quality regulations were ignored.

The Bristol industrial emissions inventory database uses data from :

· Part A Processes:

Individual spreadsheets for each authorisation.

A summary spreadsheet detailing emissions for all processes in the Bristol area.

The Environmental Protection Act 1990 requires operators of Part A processes to obtain an authorisation for each scheduled process, to monitor the release of pollutants and to submit information in order to demonstrate compliance with the standards set by the Environment Agency.

· Part B Processes:

Individual spreadsheets for each authorisation

A summary spreadsheet detailing emissions for all processes in the Bristol area.

Prescribed processes designated for local authority control must not operate without an authorisation from the local authority. Operators must submit a detailed application for authorisation and provide conditions to ensure that the process is operated using the Best Available Technique Not Entailing Excessive Cost (BATNEEC).

Public registers at both the Environment Agency and Local Authorities have data for the Part A and Part B processes.

Database

The different parameters of the database are listed below:

Data ID (source)

Local Authority Process Significance rating[1] (A to D) for Part B’s only

Facility name

Operator

Address

Location (Ordnance Survey Grid Reference)

Telephone

Contact Name

Facility Authorisation

Process Type

Potential Air Pollutants

Status (of Authorisation)

Process Equipment / Abatement technology

Output (product) process rate (1)

Unit of (1)

Throughput (main raw material) process rate (2)

Unit of (2)

Typical operational hours/day

Typical operational days/week

Typical operational weeks/year

Release height

Release diameter

Stack gas exit velocity (m/s)

Stack gas exit flow rate (m³/s)

Stack temperature

Oxygen (%) for Part B’s only

Moisture (%) for Part B’s only

Instantaneous emissions for each pollutant (g/s)

Annual emissions for each pollutant (kg/yr)

Databases

In order to store the data in a convenient way a Microsoft Access database was created in 1996 to hold the variety of industrial information. This has evolved and been further developed by new members of staff.

The air quality model eventually chosen to do the three stage Review and Assessment required by the UK government was ADMS – Urban. The creators of this model CERC has recently developed a relational database to work with the model called EMIT. This is now used to hold all the industrial and transport data.

Conclusions

Developing emissions inventories is an evolutionary process. As better data become available more accurate modelling can be achieved. Although the early modelling may not be as accurate as desired it is important to try with the data available at the time and not wait for perfection. Any systems for collecting and storing data should be flexible to allow easy modification as more data become available.

Last Updated

13th January 2005