Summary

Emission inventories are a tool to monitor the development of the emission situation in urban areas. Therefore periodic updates of the emission inventory are required. An update of an emission inventory is always combined with new or improved emission factors and models. In order to define the emission trend it is necessary to distinguish between methodological and real changes. This means using the most recent data (models and emission factors) and applying it to the new and old base year. Taking the emission inventory for the city of Graz as an example the effects of methodological changes on emission estimates are shown. Due to the fact that this emission inventory is also used for urban air quality modelling it has a high temporal and spatial resolution. Emphasis was therefore put on an accurate description of the time behaviour of the emission sources. The use of a geographical information system ensures an exact location of the emission sources.

Aim of the research

The first emission inventory for the city of Graz was created in 1989 (Sturm et al., 1994). The aim of this inventory was – besides the main purpose of quantifying emissions – to act as a base for an urban air quality model. A precondition for an emission inventory to serve as a basis for urban air modelling is a high temporal and spatial resolution - in this case – one hour and 250 x 250 m respectively. In addition it is necessary to define emission sources at a very detailed level in terms of pollutants, source strength, activity and location.

An update of the emission inventory was made for the base year 1995. The reason for this was on the one hand to make a new estimation of the emission quantities and in this way to define the trend for the last years, while on the other hand it was necessary to have an up-to-date emission inventory for on-going research projects dealing with pollution dispersion in the city of Graz.

Principal results

An update of an emission inventory is always combined with new or improved emission factors and models. In order to define the emission trend it is necessary to distinguish between methodological and real changes. That means, using present-day knowledge (models and emission factors) and applying it to the new and old base year. The methodological changes are mostly due to improved emission factors and emission models. The emission inventory is divided into three major source groups: traffic, residential combustion and industry. Each of these source groups is treated separately with different methodological approaches.

Emissions for 1995

The total emissions are a sum of the emissions estimated for the source groups traffic, residential combustion and industry. Emissions from biological processes (mainly VOC) were not considered. Total emissions on an annual basis are necessary to evaluate the emission situation of a certain area. In order to use the emission inventory as a basis for urban air quality modelling the time dependence of emission quantities is very important. Therefore it is necessary to calculate the emissions on an hourly basis. There are two possibilities to reach this goal. One is to do a time dependent disaggregation using the annual mean value and time dependent functions for monthly, weekly and daily periods. This would be a top down approach. This approach was used for industrial activities and residential combustion sources. In contrast to this it is possible to start the calculation on an hourly basis and end up with an annual value as a sum of hourly results. This approach was used for traffic emissions, where activity data is known on an hourly basis (bottom-up approach). Fig. 1 shows the VOC- emission quantities for traffic sources over the course of a day, separated into emissions from evaporation, cold and warm start, and hot emissions under running conditions.
 

Another requirement for urban emission inventories is a high spatial resolution of the emission sources. Here again bottom-up and top-down approaches can be applied. Bottom-up approaches are typical for line sources (traffic) or big point sources. Here the emission source is unambiguously defined by its co-ordinates. The opposite approach is the top-down one where statistical information from bigger units is disaggregated. This is typical for residential sources and small business activities. In order to serve as input for the dispersion model, the emissions have to be aggregated or dissagregated into grid cells according to the size of the calculation grid, i.e. 250 times 250 m uniform distribution in this case. An example of this is the spatial distribution of traffic emissions for a specific time period shown in Fig. 2.
 

Trend of the emissions between 1995 and 1988

As mentioned before, an update of an emission inventory is always based on an improved knowledge of emissions and activity data as well as on emission models. To avoid a misinterpretation of emission trends it is essential to use the same emission models and emission factors (where the relevant emission technology is taken into account) for the different base years. That means to identify the trend between two base years it is necessary to recalculate the old situation with the models used for the new one.

Taking the activity data of 1988 as a basis the differences between the estimates for 1995 and 1988 can be attributed to improvements in emission reduction technology (Table 1). For all sectors a remarkable reduction of the emission quantities was identified. The biggest reductions were found for CO and SO₂. The CO reduction was due to the improved emission standards for cars, while the SO₂ reduction is due to changes in fuel used and improved fuel quality in residential and industrial combustion processes.

Table 1: Trend in emission quantities between 1988 and 1995
 

Validation

It is not easy to validate emission inventories because an emission inventory is always based on simulations, estimations and only sometimes on real time measurements of emission quantities. One possibility for validation exists when looking at long term evaluations in air quality. Due to the fact that the air quality of the city of Graz is dominated by local emission sources, the emission reduction should be found also on the air quality side. Of course meteorological conditions were not the same during the different years, but the annual mean air quality values of the different years should show the same trend as the emission quantity does. Table 2 contains the average of all yearly mean air quality values of each of the six monitoring stations located within the city boundaries of Graz. The comparison of the yearly mean values (in Table 1 – emissions, in Table 2 – air quality) shows that there is a good agreement between both trends 1995 to 1988 (see Table 3).

Table 2: Yearly mean values of the air quality monitoring stations of Graz (average of all monitoring stations)
 

*NO_x as NO₂

Table 3: Trend in air quality (yearly mean values) and emission quantities
 

*NO_x as NO₂

Uncertainties

The update of the urban emission inventory of Graz provided the possibility to show how changes in methodology and emission factors can influence the results. The basis was the emission inventory generated in 1989 (base year 1988) (Sturm et al., 1994). During the update for the reference year 1995 some calculation methods had been updated so as to conform to the presently available state of the art. The same was done with technology dependent emission factors. Using the knowledge of 1995 (emission factors and methodologies) and applying it to the 1988 situation (activity data, fleet composition and ages, etc.) results in quite different emission quantities for the same reference year. The overall changes are in a range of -30% for NO_x and +20% for SO₂. That means NO_x was overestimated in the old emission inventory and SO₂ underestimated. The different NO_x emission sources showed different trends with a decrease in traffic emissions and an increase in residential combustion emissions. SO2 showed reductions for industrial activities and an increase for residential combustion sources. The full results are given in Sturm et al., 1998.

Main conclusions

A big question mark concerning the usability of an emission inventory for urban air quality modelling is always the accuracy of the emission of estimates. Taking the emission inventory for the city of Graz as an example, it was possible to show the importance of having the best and most accurate activity and emission data available. Especially when using statistical data, this data has to be adjusted to the local situation instead of using national statistics.

For industrial processes, the use of statistical values should be limited to sources which do not contribute excessively to the overall emission quantities for this specific emission sector. This is not the case when looking at small combustion units for space and water heating, and the calculation method for this source group is a top down approach and therefore based on statistics. In the case of the city of Graz SO₂ and CO₂ emissions are dominated by this source group, CO and VOC emissions contribute strongly to this group. Therefore, it is evident that the appropriate statistics in this field must be used.

References: