3 Methods - Costs

UK cost curves were developed by AEA Technology for SO₂, NOx and VOCs, and by the Imperial College Centre of Environmental Technology (ICCET) for NH₃.

3.1.1 Input data

1) UK inventory data

The starting point for the cost curves was emission estimates for 1990. For SO₂, NO_x and VOCs these were taken from the most up-to-date version of the UK’s National Atmospheric Emissions Inventory (NAEI 1997), and from both the NAEI and MARACCAS for NH₃.

2) UK growth rates

The 1990 emission estimates were projected forward to 2010 using provisional projections of industrial growth rates provided by DTI. Growth rates for fuel use were available for the major industrial and final-use sectors individually. The Central (growth), High (fuel price) scenario was used in all cases.

The projected NAEI data were then converted into the emission categories used in the RAINS model developed by IIASA. In a number of cases there is poor correspondence between NAEI source categories and those used in RAINS. For example, whereas IIASA splits industrial combustion into boilers and other devices, the NAEI does not. The conversion of NAEI sectors to those used by IIASA, therefore, involved some subjectivity. The emission estimates derived in this manner represent emissions in 2010 assuming no change in the levels of control from 1990 onwards, and will be referred to hereafter as the ‘uncontrolled emissions’.

3) UK forecasts

The cost curves also make use of UK forecasts of the impact of current legislation. The forecasts take account of the following legislation:

• Environmental Protection Act
• Solvents Directive
• Petrol Distribution (Stage 1) Directive
• Sulphur in Liquid Fuels Directive
• Large Combustion Plant Directive
• Directives on transport emissions up to and including the legislation from Auto-Oil I: 91/441/EEC, 94/12/EC, 93/59/EEC, 91/542/EEC (Stage 1 and 2), the European Conciliation Committee and Common Position agreements on vehicle emissions (Auto-Oil) Euro III (2000) and Euro IV (2005) and European Directives Relating to the Quality of Petrol and Diesel Fuels (2000 and 2005).

3.1.2 Development of the Cost Curves

Each cost curve starts with the uncontrolled emission. The first part of the cost curve includes those measures that are the most cost-effective means of achieving the UKREF emission. For example, the uncontrolled VOC emission for industrial adhesives is 86.7 kt, whilst the UKREF emission is 42.7 kt, a decrease of 44 kt. There are three options for the industrial adhesives sector - good housekeeping, the additional use of substitution, and the additional use of incineration, in ascending order of cost per tonne of VOC abated. The reduction in emissions from the uncontrolled emission to the UKREF emission can be achieved, as shown in Table 3.

Table 3. Development of the VOC cost curve for the industrial adhesives sector.

Measure	Marginal cost	Efficiency	Uptake	Reduction
	euro/tonne	%	%	kt
Housekeeping	10	10.5	100	9.1
Substitution	400	53.9	93	34.9
Substitution	400	53.9	0	-
Incineration	6000	57.9	0	-

The housekeeping measure is the most cost effective and is applied fully. However, this reduces emissions by only 9.1 kt. The substitution measure is the next most cost effective, but only needs to be applied in 93% of cases in order to provide the remaining 34.9 kt of reduction necessary to give a total reduction from the sector of 44.0 kt. Because of the need for only partial rather than complete implementation of substitution, it appears twice in the list, the extra allowing an additional 7% reduction beyond that required. Finally, incineration is the least cost-effective measure.

Figure 1 shows a typical cost curve. Starting from the uncontrolled emission, emissions are reduced at increasing marginal cost (indicated by the gradient) until the UKREF emission is reached. The marginal cost of abatement just above this emission level is lower than the marginal cost of abatement just before this point is reached. This is because the cost curve is constrained to the controls that will come in under UKREF, as a result of existing legislation. Thus, some of the most cost-effective measures might not be used until after the baseline emission is reached, because they address sectors that are not controlled in the UKREF scenario.

3.1.3 Development of regional cost curves

The 1997 NAEI comprises spatially disaggregated (or regional) inventories of individual pollutants and regional cost curves were, therefore, developed by using these regional inventories rather than a UK inventory. In all other respects, the cost curve was identical, i.e. the same cost data were used and the same growth rates were assumed. The regional inventories were developed using the same data as used to map emissions. (see section 3.2 below)

3.1.4 Uncertainties in the cost curves

There are a number of areas of uncertainty in the cost curves. These may be divided into uncertainties regarding uncontrolled emissions, uncertainties regarding the UKREF emissions and uncertainties regarding the costs and effectiveness of the control options used in the cost curves.

The UKREF emissions data for the cost curves are taken from the 1997 NAEI. The NAEI has been developed over a period of more than ten years, allowing the emissions data to achieve what is believed to be a good level of accuracy. The inventory for SO₂ is probably accurate to ±10% while the NOx and VOC inventories are perhaps accurate to ±25%. However, the NAEI is continuously being updated and improved, as the emission sources and processes are better understood or more accurate data become available. This allows historic data to be revised retrospectively and it is to be expected that the 1990 data in the NAEI published in coming years might vary from the data used in this study.

The UKREF emissions are derived by making assumptions about the percentage reduction in emissions from each emitting sector as a result of legislation. Although these assumptions are made on the basis of a wide body of information, the approach is simple and there is scope for considerable uncertainty. Risk analysis, carried out on an earlier set of VOC forecasts, suggested that the uncertainty in the reduction achieved was likely to be of the order of ±10%. A similar figure could be assumed for NOx and SO₂. It should be said that the emission forecasts used to generate the UKREF emissions are fairly conservative, in that no account is taken of improvements in technology, except where required by environmental legislation. Thus, it might be expected that the UKREF emissions are more likely to be overestimated rather than underestimated.

Figure 1. Typical cost curve for the abatement of air pollutants subject to existing legislation.

The final source of uncertainty is the cost effectiveness data derived from the IIASA methodology. This methodology has undergone extensive peer review throughout Europe and is considered by the authors to be the most comprehensive and accurate set of cost effectiveness data available. Nonetheless, the data will be subject to quite large uncertainty. The effectiveness of control measures is likely, if anything, to be overstated, since IIASA often assume that control measures can be used for all processes within a sector whereas this is not always likely to be true.

No attempt has been made to quantify the likely impact of these uncertainties in numeric terms. However a qualitative assessment is made below (Table 4). In all cases there are some biases to overestimation and some to underestimation.

Table 4. Sources of uncertainty in abatement cost estimates, and potential effects on the outcome of the analysis. ++ and -- denote cases where the real costs could be significantly higher or lower respectively than those estimated here. + and – identify cases where uncertainties may still influence the results, but less significantly.

Source of uncertainty	Impact on cost of reaching a given emission ceiling
Scenario uncertainties
Uncontrolled emissions data	+/-
UKREF	+/--
Option uncertainties
Effectiveness of control measures	++/-
Cost of control measures	+/-

3.2 Mapping of emission totals

UK emissions of NO_x, SO₂, and VOC were mapped using existing data available from the NAEI. These include point source data for power stations and other large combustion plant, incineration plant, iron and steel processes, cement and lime processes, chemical processes, refineries, large solvent users and whisky distilleries. Road transport emissions were mapped using information on traffic on major roads. Other sources were treated as area sources and were distributed using regional statistics, such as fuel use or population.

Emission estimates for ammonia for the UK were obtained from three sources: the RAINS and MARACCAS models and the NAEI. Total estimates from MARACCAS (which was taken as the main source of data for NH₃ emissions and costs in this study) and the NAEI agreed well. However, significant differences were noted when these estimates were compared with RAINS, although this was not surprising, given that knowledge of ammonia emissions is improving rapidly, in response to the demands of legislation, such as the proposed Protocol and Directive.

Ammonia was mapped taking account of both agricultural and non-agricultural sources. The principal source of data for this was the ITE land cover database. Ammonia emissions are dominated by livestock production. Non-agricultural sources are numerous; including industry, emissions from catalytic converters, pets, horses, wild animals etc. There is continuing uncertainty concerning non-agricultural emissions of ammonia for most of the EU countries.

Chapter 2          Chapter 4