FIFTH INTERIM REPORT. Cost-effective Control of Acidification and Ground-Level Ozone. Part B: Ozone Scenarios

Transcription

1 FIFTH INTERIM REPORT Cost-effective Control of Acidification and Ground-Level Ozone Part B: Ozone Scenarios Fifth Interim Report to the European Commission, DG-XI Markus Amann, Imrich Bertok, Janusz Cofala, Frantisek Gyarfas, Chris Heyes, Zbigniew Klimont, Marek Makowski, Wolfgang Schöpp, Sanna Syri May 1998 IIASA International Institute for Applied Systems Analysis A-2361 Laxenburg Austria Telephone: Telefax:

2 Cost-effective Control of Acidification and Ground-level Ozone Part B: Ozone Scenarios Markus Amann, Imrich Bertok, Janusz Cofala, Frantisek Gyarfas, Chris Heyes, Zbigniew Klimont, Marek Makowski, Wolfgang Schöpp, Sanna Syri Fifth Interim Report to the European Commission, DG-XI This report constitutes Part A of the Second Interim Report on the Study Contract B4-3040/97/000654/MAR/B1 Economic Evaluation of Air Quality Targets for Tropospheric Ozone. May 1998 Interim reports inform on work of the International Institute for Applied Systems Analysis and have received only limited review. Views or opinions expressed herein do not necessarily represent those of the Institute or of its National Member Organizations. IIASA International Institute for Applied Systems Analysis A-2361 Laxenburg Austria Telephone: Telefax: info@iiasa.ac.at

3 Acknowledgments The authors would like to express their gratitude to Arne Drud from ARKI Consulting and Development, Denmark, for making the CONOPT library for the non-linear optimization available, Prof. Anton Eliassen from the Norwegian Meteorological Institute, Oslo, and Peringe Grennfelt from the Swedish Institute for Environmental Research, Gothenburg, for offering thoughtful comments on the modeling approach, David Simpson from the Nowegian Meteorological Institute for producing the large number of scenarios required as input to the regression analysis, Steffen Unger and Prof. Achim Sydow (GMD-FIRST, Berlin, Germany) for transferring the EMEP model to parallel computers, Jean-Paul Hettelingh and Max Posch from the Coordination Centre for Effects (CCE) at the National Institute for Public Health and the Environment (RIVM) in Bilthoven, Netherlands, for providing the data bases on ecosystems and population densities, Mark Barret from SENCO, UK, for assisting in the interpretation of the documents relating to the Auto/Oil 1 study, Mari Saether, Martin Lutz and Ger Klaassen from DG-XI for guidance in performing this study.

4 Cost-effective Control of Acidification and Ground-level Ozone Part B: Ozone Scenarios Table of Contents 1 STRUCTURE OF PART B OF THE FIFTH INTERIM REPORT 5 2 OZONE EXPOSURE CRITERIA The AOT60 as a Surrogate Indicator for Risk to Human Health The AOT40 as a Critical Threshold for Vegetation Protection 7 3 THE SITUATION IN 1990, THE EXPECTED IMPACTS OF THE CURRENT POLICIES AND THE MAXIMUM TECHNICALLY FEASIBLE REDUCTIONS Health-related Ozone Exposure Vegetation-related Ozone Exposure 13 4 TARGETS FOR THE REDUCTION OF GROUND-LEVEL OZONE IN EUROPE Basic Target Setting Rules A Compensation Mechanism Rationale Compensation Rules 19 5 OZONE REDUCTION SCENARIOS Main Input Assumptions Baseline Scenarios Targeted at Ground-level Ozone Scenarios Targeted at Health Protection (E1) Scenarios Targeted at Vegetation Protection (E2) The Ozone Strategy: Combined Targets for Health and Vegetation Protection (E3) The Ozone Directive: Exposure Limits for Health and Vegetation (E4) 44 6 SENSITIVITY ANALYSIS Allowing Unlimited Compensation (Scenario E5) Impacts of the Post Kyoto Energy Scenario (E6) 52 3

5 6.2.1 A Combined AOT60/AOT40 Gap Closure/Target Scenario An AOT60/AOT40 Target Scenario Interaction with Non-EU Countries (E7) 58 7 CONCLUSIONS Summary of the Scenario Results Reference Case and Maximum Feasible Reductions Scenarios for Reducing Ground-level Ozone Sensitivity Analysis Caveats Conclusions 67 8 REFERENCES 68 4

6 1 Structure of Part B of the Fifth Interim Report Part B of the Fifth Interim Report addresses scenarios for reducing ground-level ozone in Europe. Modeling methodology and databases used for the analysis are described in Part A of the Fifth Interim Report, while acidification-related emission control scenarios are the subject of Part C. Section 2 of Part B introduces the environmental endpoints for human health and vegetation selected for the optimization analysis. Using these endpoints, Section 3 explores the possible range for ozone reduction by outlining the anticipated results from the presently planned emission reductions and comparing them with the hypothetical exposure after full implementation of currently available emission control technologies. Section 4 discusses basic rules for setting environmental interim targets in order to arrive at a plausible, equitable and robust international distribution of costs and environmental benefits. Scenario analysis is presented in Sections 5 and 6. Section 5 illustrates the practical implications of the target setting rules. By complementing the gap closure concept with absolute targets for ozone exposure, joint health/vegetation targeted scenarios are presented for a range of ambition levels. Section 6 analyzes the robustness of the optimized emission reductions against variations in important input parameters and assumptions. The assessment addresses the impacts of modified compensation rules applied for the gap closure optimization, the implications of a Post Kyoto energy development, and the possible interaction of ozone control strategies in the EU-15 with measures in non-eu countries. Conclusions from the analysis are drawn in Section 7. Although there are some modifications to the target setting and compensation rules described in Section 3, readers who are familiar with earlier reports might find it most effective to turn immediately to Section 6 addressing the sensitivity analysis. 5

7 2 Ozone Exposure Criteria The analysis presented in this report addresses the protection of human health and vegetation against elevated ozone exposure. The appropriate exposure measures for environmental longterm targets for these categories are discussed in detail in the Draft Position Paper on Ozone prepared by the Commission s Services. For modeling and optimization purposes, however, the use of some of these original criteria proved to be complicated and impractical, and some surrogate indicators have been introduced instead. By no means the use of such surrogate indicators does question the original definition of the criteria. Furthermore, they must not be interpreted as actual damage estimates. The only reason for the surrogate indicators is to facilitate the modeling and optimization exercise. 2.1 The AOT60 as a Surrogate Indicator for Risk to Human Health Following the revised WHO Air Quality Guidelines for Europe (WHO 1997), the Draft Position Paper on Ozone prepared by the Commission s Services proposes a maximum eighthour average concentration of 60 ppb (120 µg) as the long-term environmental objective for the EU ozone strategy 1. The ultimate goal would be to eliminate all excess of this criterion. The modeling of European abatement strategies for individual days over a multi-month period is a rather ambitious task and is not entirely feasible at the moment. In order to simplify the modeling task, and particularly to find a manageable approach for the reducedform model implemented in the RAINS optimization, the target of no-exceedance of the WHO criterion (60 ppb as maximum eight hours mean concentrations) was converted into an AOT index, which could be handled in a similar way to the AOT40 for vegetation. As a result, an AOT60 (i.e., the cumulative excess exposure over 60 ppb, for practical reasons over a six-month period) of zero is considered as equivalent to the full achievement of the WHO criterion. Any violation of this WHO guideline will consequently result in an AOT60 of larger than zero. It is important to stress that this AOT60 surrogate indicator has been introduced purely for practical modeling reasons. Given the current knowledge on health effects it is not possible to link any AOT60 value larger than zero with a certain risk to human health. The only possible interpretation is that if the AOT60 is above zero, the WHO criterion is exceeded at least once during the six-month period. For the actual model exercise, the AOT60 of different emission control scenarios at a given site is calculated as a function of the emission levels of NO x and VOC in the various European countries (see the description of the 'reduced form' model in Part A). The reducedform model is derived from a statistical analysis of a large sample of results obtained with the full EMEP model. The EMEP model provides ozone levels at six-hourly intervals (0 GMT, 6 GMT, 12 GMT and 18 GMT) over a six months period. Following the findings of various studies for different parts of Europe (Künzle, 1995; Dumont 1998), the AOT60 has 1 The maximum is calculated from running eight-hour averages of the one-hour mean concentrations. 6

8 been calculated as the excess ozone over 60 ppb at 12 GMT and 18 GMT, accumulated over the entire period and multiplied by a factor of six. 2.2 The AOT40 as a Critical Threshold for Vegetation Protection In the absence of accepted dose-response curves applicable at the large scale, the analysis in this report uses the concept of critical thresholds as developed within the framework of the UN/ECE Convention on Long-range Transboundary Air Pollution. The Working Group on Effects of this Convention established two long-term related critical levels: For agricultural crops and herbaceous plant communities (natural vegetation), the critical level is set at an AOT40 of 3 ppm.hours for the growing season and daylight hours, over a five-year period; For forest trees, a critical level of 10 ppm.hours for daylight hours, accumulated over a six-month growing season, is proposed. The AOT40 is calculated as the sum of the differences between the hourly ozone concentrations in ppb and 40 ppb for each hour when the concentration exceeds 40 ppb, using daylight hours only. It has been shown elsewhere that for the currently prevailing European ozone regime the critical level for crops and natural vegetation is stricter than the critical level for forest trees. This means in other words, while the critical levels for forest trees are usually met when the critical level for crops and vegetation is achieved, the opposite statement does not hold. Based on this finding it has been decided to restrict the scenario analysis to the critical levels for crops and natural vegetation. If considered necessary, however, there are no methodological problems to prevent exploring scenarios for the achievement of the critical levels for forest trees separately. For the regression analysis of the reduced-form ozone model, the AOT40 has been calculated from the results of the full EMEP model by multiplying the excess exposures over 40 ppb at 12GMT and 18 GMT, accumulated over a three months period, by a factor of six. 7

9 3 The Situation in 1990, the Expected Impacts of the Current Policies and the Maximum Technically Feasible Reductions 3.1 Health-related Ozone Exposure It is documented elsewhere that actual ozone concentrations are strongly influenced (a) by the concentrations of the precursor emissions and (b) by the actual meteorological conditions. Excluding for the moment the meteorological influence, the following figures portray the anticipated (from the REF scenario) and the possible (from the MFR scenario) changes in AOT60 between 1990 and In doing so, the analysis uses the mean meteorological conditions of the five years 1989, 1990, 1992, 1993 and 1994, for which results of the EMEP model are available. Obviously, the data displayed in the maps cannot be directly compared with real observations, since these incorporate the specific meteorological conditions for the selected year. Figure 3.1 illustrates that for the emissions of 1990 and using the meteorological conditions of five years, in average the highest (rural) AOT60 of more than 6 ppm.hours occurred in northern France, Belgium and Germany. In many other parts of France, Germany and Benelux, the AOT60 was modeled in a range of 5-6 ppm.hours. Typical rural values in the UK and Austria were between 2 and 3 ppm.hours, while the highest AOT60 in Spain and Greece was between 1 and 2 ppm.hours. Portugal is estimated at 3 ppm.hours, while Scandinavia did not experience significant excess of the AOT Figure 3.1: Mean AOT60 for the emissions of 1990 taking into account the meteorological conditions of five years (in ppm.hours) 8

10 The emission controls calculated for the REF scenario for the year 2010 (NO x -46 percent, VOC -47 percent compared to 1990) are expected to have profound impacts on ozone exposure (Figure 3.2). Expressed as a mean over five years, the highest AOT60 in Europe would decline to about 3 ppm.hours, i.e., by about 50 percent. In many other regions there would be even higher improvements, leading to an average drop of the AOT60 by 60 percent. Even further cuts in the AOT60 could be achieved by the maximum feasible emission reductions (Figure 3.3). While keeping the emissions from the non-eu countries at their REF level, a 68 percent decline of NO x and 65 percent cut of VOC emissions in the EU-15 would bring highest AOT60 levels in Europe down to about 1.5 ppm.hours, which is percent below the 1990 levels. Since most of Europe would be able to achieve the WHO guideline, the average AOT60 would be 83 percent lower than in Figure 3.2: Mean AOT60 for the emissions of the REF scenario, taking into account the meteorological conditions of five years (in ppm.hours) 9

11 Figure 3.3: Mean AOT60 for the emissions of the MFR scenario, taking into account the meteorological conditions of five years (in ppm.hours) Although the AOT60 is a convenient index to model, it might be a difficult one to interpret and to link with generally understandable notions. A better measure in this respect is obviously the number of days on which the WHO criterion is exceeded. Figure 3.4 displays the regional distribution of the (the five years average) excess days for the emissions of the REF scenario. It is interesting to note that there is not a 1:1 relationship between the AOT60 and the number of days across all regions in Europe, indicating that the amount by which the 60 ppb criterion is exceeded varies over Europe. Whereas the highest AOT60 is expected for the northern part of Europe (France/Belgium/Germany), large numbers of days exceeding the 60 ppb threshold are also found in Italy, where the AOT60 is typically 20 to 30 percent lower than in northern Europe. A detailed analysis of the available monitoring results is presented in van Hout (1998). This phenomenon underlines the observation that ozone exposure shows different temporal characteristics in different parts of Europe, a fact which is important to take into account when designing emission control strategies. 10

12 Figure 3.4: Number of days exceeding the WHO ozone guideline value (60 ppb eight hour mean) for the emissions of the REF scenario. The map displays the second highest occurrence out of the meteorological conditions of five years Table 3.1 presents two different types of population exposure for the AOT60. The cumulative index reflects for each country the total exposure of a population and is expressed in person.ppm.hours. The RAINS model calculates these indices on a grid basis (using gridded data on AOT60 and population); in a second step these grid values are aggregated to the country level. The indices presented in this report use the AOT60 concentrations per grid, representing the rural ozone concentrations, and the total population per grid in Inaccuracies may occur for grids with major urban areas, where the rural ozone concentrations used for these analysis present an upper bound for the concentrations in the cities, and are lower than the concentrations occurring in the city plumes (Kindbom and Grennfelt, 1998). The average indicator reflects the average exposure of a person in a country, calculated from gridded data. It is important to stress that these indices may not be used to derive estimates of health damage, for which more detailed information is deemed necessary. In the context of this report, these indices provide relative measures to enable a quick comparison of different scenarios. 11

13 Table 3.1: Population exposure indices (AOT60) for 1990, the Reference (REF) scenario and the maximum feasible emission reductions (MFR), mean values for five meteorological years. The table presents the cumulative population exposure for each country (in million person.ppm.hours) and the average exposure per person in each country (in ppm.hours). Note that the environmental long-term target is proposed at a level of zero. Cumulative population exposure (million person. ppm.hours) Average population exposure (ppm.hours) 1990 REF MFR 1990 REF MFR Austria Belgium Denmark Finland France Germany Greece Ireland Italy Luxembourg Netherlands Portugal Spain Sweden UK EU As shown in the table, in 1990 the average exposure was highest in Luxembourg, Belgium, France, Germany and the Netherlands; the highest cumulative exposure (due to the large population) occurred in Germany, France, Italy and the UK. The cumulative exposure of the population in the EU-15 countries is expected to decline by 58 percent as a result of the current policy. Larger improvements occur in Austria (-81 percent) and the Scandinavian countries (60-70 percent), while for the UK and Netherlands a decrease in AOT60 by about 40 percent could be expected. The maximum feasible emission reductions would reduce the exposure indices by 84 percent. It is important to mention that there are some areas where, despite - or because of - the anticipated emission reductions of the REF scenario, for individual years the AOT60 is expected to slightly increase as a result of current policy. Using mean meteorology, however, masks the increase occurring in individual years. The explanation for this increase is related to the ozone formation chemistry. Put in a rather simplistic way, very high NO concentrations (in areas with high NO x emissions) have, i.a., two effects: (a) they lead to the titration of ozone, i.e., the conversion of ozone and NO into NO 2, and (b) they cause a (partial) depletion of OH radicals. This resulting shortage of OH radicals at such high NO x levels limits ozone production. Reducing NO x emissions from such a high level will increase the available OH radicals, and more ozone will be produced, until NO x emissions are so low that the ozone production will be limited by the available NO 2 molecules. As indicated in Section 2.5 of Part A of this report, reducing NO x will lead for some time to increased ozone. Beyond a certain NO x reduction level, however, ozone will decline again. 12

14 Figure 3.5 supports this explanation by illustrating the emission densities in It is important to realize that the emissions in the areas where the increase occurs (UK, Belgium, Netherlands, etc.) are up to a factor of 10 higher than in other industrialized European regions (compare e.g., southern Germany). It is also important to realize that this ozone increase disappears for the maximum feasible emission reductions. This means that sufficiently high NO x reductions (which are considered as technically feasible) can overcome the temporary ozone increase everywhere ,y,, yy, yy,,,,, yyy,,, yyy,y,,,y,y,y,y,,,y,y,y,,,yy,y yy,y,,,y,y,y,, yy y y yy,,,,y,y y,,, yyy y,, yy,y,, yy,y,,, yy,y,, yy,y,,, y,, yy yy,, yy,y,,, y, y,y,y,y,y,, yy,,,y,, yy,,,,,,, yyy,,,,y,y,,, yyy,,, yyy,,, yyy,,, yyy,,, yyy,y,,,,, yyy,, yy,y yy,, yy,y,, yyy,,, yyy,y,, yy,y,,,,,,, yyy,,, yyy yy,y,y,y,,,, yyyy,,,,,, yyyy, y yy,, yy,y,, yy,y,y,, yy,y,,,y,, yy,y,,,, yy,y,y,y,, yy,y,,,y Unit:,y,y,,,, yy,,,,,,,, yy,,,, yyyy,, yy yyyy yyyy,,,, yy,, yy yy,y,, yy,, yy yy,, yy yy,y,, yy yy,y,y yy,, yy,y,, yy,, yy,, yy,, yy Figure 3.5: NO x emissions per EMEP grid cell in 1990 (in tons) 3.2 Vegetation-related Ozone Exposure Figure 3.6 displays the excess AOT40 (over the critical level of 3 ppm.hours) calculated for the emissions of the year 1990 using the five years mean meteorology. The map clearly shows that in most countries of the EU-15 the critical level for vegetation was exceeded. The only exceptions are parts of the Scandinavian countries. In an area extending from Paris over Belgium and Netherlands to Germany the excess AOT40 reached 16 ppm.hours, i.e., it exceeded the critical level by more than a factor of five. It is important to note that ozone levels in many areas, which do not experience significant excess of the AOT60, exceed the AOT40 criterion considerably. This applies particularly to the Mediterranean countries and some Alpine regions. 13

15 Figure 3.6: Excess AOT40 above the critical level of 3 ppm.hours for the year 1990 (using five years mean meteorology), in ppm.hours Figure 3.7: Excess AOT40 above the critical level of 3 ppm.hours for the Reference scenario in 2010 (using five years mean meteorology), in ppm.hours. 14

16 Figure 3.8: Excess AOT40 above the critical level of 3 ppm.hours for the maximum feasible emission reductions in 2010 (using five years mean meteorology), in ppm.hours. The emission reductions of the Reference scenario will generally lead to a decline of the excess AOT40, but will not significantly increase the protected area (Figure 3.7). Peak levels are in a range of ppm.hours. The maximum feasible emission reductions are expected to achieve a 50 percent and higher cut of the excess AOT40 in most regions (Figure 3.8). Table 3.2 introduces two vegetation-related exposure indices. The cumulative vegetation exposure index is calculated as the excess AOT40 (i.e., the AOT40 in excess of the critical level of 3 ppm.hours) multiplied by the area of ecosystems that is exposed to the excess concentration. The index is calculated on a grid resolution, considering agricultural land, natural vegetation and forest areas. The average vegetation exposure index reflects the average excess AOT40 (over all grids in a country). The estimate of these indices is based on rural ozone concentrations. 15

17 Table 3.2: Vegetation exposure indices for 1990, the Reference Scenario in 2010 and the maximum feasible emission reductions Cumulative vegetation exposure index (million hectares.excess ppm.hours) Average vegetation exposure index (excess ppm.hours) 1990 REF MFR 1990 REF MFR Austria Belgium Denmark Finland France Germany Greece Ireland Italy Luxembourg Netherlands Portugal Spain Sweden UK EU In 1990, France, Germany, Spain and Italy experienced the highest cumulative indices, while, for instance, the UK had a significantly lower value. On average, the highest exposure was experienced in France, Luxembourg, Belgium, Germany and Italy. The current reduction measures are expected to decrease the indices for the EU-15 by 37 percent, which is significantly lower than the expected decline of the health-related exposure indices (-58 percent). While areas with already low indices achieve a percent reduction (Ireland, Sweden, etc.), the expected improvement in Belgium, the Netherlands and the UK ranges only between five and 15 percent. Again, as for the AOT60 this low improvement is caused by the features of the ozone chemistry in high-no x regions. The maximum feasible reductions would overcome these effects and lead to a 63 percent reduction for the EU-15 as a whole. 16

18 4 Targets for the Reduction of Ground-level Ozone in Europe The assessment in the preceding section clearly demonstrates that the occurrence of and the reduction potential for ground-level ozone show distinct spatial differences over Europe. Furthermore, there is robust evidence that the presently available technical emission control measures will not be sufficient to meet the environmental long-term targets (the no-damage levels) everywhere in Europe within the next one or two decades without interfering with the business as usual expectations on economic development and energy consumption. In such a situation the choice of an equitable environmental interim target becomes crucial for deriving a balanced emission control strategy. 4.1 Basic Target Setting Rules The analysis of the Fourth Interim Report identified two basic concepts for setting interim targets, i.e., prioritizing measures in highly polluted areas by imposing uniform absolute exposure limits over the entire area as one principle, and postulating equal relative improvements in relation to the situation in a base year (the gap closure concept) as the alternative option. It was clearly shown that these two different conceptual approaches imply fundamentally different spatial distributions of environmental benefits and emission abatement efforts over Europe. The discussions on the EU and UN/ECE levels following the Fourth Interim Report concluded that a combination of both principles could most likely lead to internationally acceptable solutions and should be further explored. Furthermore, based on model results for five different meteorological years it was demonstrated that actual ozone levels do not only depend on the levels of precursor emissions, but also to a significant degree on the specific meteorological condition. Emission control strategies addressing an extreme situation might therefore look rather different than strategies tailored towards the improvement of typical situations. For the purposes of strategy development, the discussions following the presentations of the last Interim Reports concluded that, when considering the uniform ozone limit target, the most difficult situations should be excluded from the analysis. In practice, the strategy should be constructed in such a way that it would meet the absolute AOT targets in four out of five years. It is important to stress that the major motivation for this four out of five principle in the context of strategy development is the concern to avoid reliance on the model performance for extreme (and perhaps rare) situations. By no means should this principle prejudge the selection of meaningful criteria against which the compliance should be checked. The Fourth Interim Report revealed that the original definition of the gap closure target, i.e., closing the gap between the 1990 situation and the environmental long term target everywhere by the same percentage, would push areas with comparatively low ozone exposure to costly emission reductions, while less burden would be placed on more polluted regions. It is important, however, to recall that model uncertainties are, for a number of reasons, largest for just these low ozone ranges. In order to maximize the robustness of results 17

19 obtained from the currently available ozone optimization model the introduction of a confidence interval was accepted and not to let model results below this limit influence the actual strategy development. Experiments suggested that this confidence interval for the present AOT60 optimization model be set at 0.4 ppm.hours. With the new data set it turns out that, despite the introduced confidence limit, the traditional definition of the gap (between 1990 level and the long-term target) would still impose unbalanced reductions close to the maximum technically feasible in areas with ozone targets between 0.4 and 1 ppm.hours. The major reasons for this effect are the higher relative contribution of background ozone at low ozone levels, and an increasing noise in the regression coefficients of the ozone model for low AOT60 levels. As a practical solution, the calculations in the Fifth Interim Report slightly re-define the gap closure: Instead of relating the excess exposure (gap) to the long-term target of 0.0 ppm.hours, the new definition uses the model confidence range of 0.4 ppm.hours as the reference point. This revised formulation results in a relaxation of the environmental interim targets, which is, however, in practice only significant for targets below 1 ppm.hour. It must be stressed that the main argument for this change is the concern to avoid overstressing the reliability of the available ozone optimization model. Areas with less severe ozone problems should not be forced to costly and far-reaching measures, which cannot be firmly defended on the basis of the available models. 4.2 A Compensation Mechanism Rationale Earlier analysis (see, e.g., Amann et al., 1997) demonstrated that the optimal allocation of emission controls may be strongly influenced by the need to exactly meet specified environmental targets at a few single grid cells, while for the majority of grid cells the targets are usually over-achieved. The sensitivity of the optimization results towards modifications of the environmental targets of these binding grids was the subject of numerous discussions in the past. It was argued that the requirement to achieve stringent targets in isolated areas could possibly imply unbalanced high costs without yielding adequate benefits. This concern is even more pronounced when the targets are not related to absolute exposure levels, but to interim targets on the way towards the ultimate environmental objective. Both the Council Conclusions on the Acidification Strategy (8387/97 ENV 146 PRO- COOP45 - COM(97) 88 final) and the UN/ECE Working Group on Strategies (EB.AIR/WG.5/54) requested the analysis of alternative concepts, where environmental targets for single ecosystems are not allowed to drive the overall optimization system to extreme solutions. It is important to understand that, in the search for alternative solutions, the starting point is an already cost-minimized outcome of an optimization. Consequently, further cost savings (or more exactly higher benefit/cost ratios) can only be achieved by compromises (modifications) of some of the environmental targets (i.e., constraints in the equation system). In principle, one could construct two alternative concepts for modifications of the environmental constraints: 18

20 Single environmental constraints (targets) could be simply relaxed, e.g., if the costs for achieving them exceed a certain threshold. In this case the optimization process will violate the environmental target and result in lower costs and in lower environmental improvements. Cost savings are achieved through less environmental quality. It has been shown before that, at least for the case of acidification, the ratio between cost savings and the area of unprotected ecosystems (e.g., ECU saved/hectare of unprotected ecosystems) is rather linear over the reasonable range of interest, and does not show a distinct knuckle point. Alternatively, a compensation mechanism could be constructed, which maintains the overall level of environmental improvement (e.g., in each country), but compromises on the spatial distribution of the environmental benefits. It has been shown that, in such a case, overall abatement costs could be reduced, while keeping the environmental impact indicator (over an aggregate of grid cells) constant Compensation Rules The Fourth Interim Report presented a prototype compensation mechanism and illustrated the implications for ground-level ozone. This compensation mechanism allows the violation of environmental targets at single grid cells or single years as long as this excess is compensated by additional improvements in other years or at other grid cells within the same country. This compensation mechanism was applied only to the gap closure target to allow flexibility of the interim improvements on the way towards the achievement of the long term target. It is not applied to the absolute AOT target, which controls the highest absolute ozone levels in Europe. Earlier discussions with Member States proposed to consider differences in the stock at risk over grid cells and to put more relative emphasis on densely populated areas. In response to this, a population weighting was introduced in the national compensation balances. This means that excess AOT60 must be compensated on a population-adjusted basis, e.g., a small excess of AOT60 in a big city by larger improvements in less populated rural areas. In practice, this population-weighting mechanism assures that for each country the population exposure index of the optimized solution (applying the compensation mechanism) may not exceed the index resulting from the original targets. Discussions after the presentation of the last report addressed a possible inequitable treatment of large and small countries implied by the proposed compensation mechanism. Indeed, a situation could be constructed where two countries experience in a limited area the same ozone exposure. If one of these countries is sufficiently large, it might accept an unlimited violation of the original environmental target and compensate it by additional improvements in other areas with possibly different ozone formation regimes and would therefore experience only little additional demand for further emission reductions. A small country will not have the same compensation potential available and would therefore be forced to more ambitious measures to meet stricter targets at the hot spot 2. The acceptable ozone exposure in the problem area might therefore depend via the ability of a country to compensate violations on the size of countries. It has been argued before that the basic rationale of the compensation mechanism is the trading of cost savings against a modified spatial distribution of environmental improvements 2 The additional measures would not necessarily have to be made in the small country itself. 19

21 (the overall environmental improvement is kept constant for each country). Obviously, the spatial flexibility for re-arranging environmental improvements will usually be higher in large than in small countries. As a way out, one could theoretically consider segmenting larger countries into several zones, or to aggregate the compensation areas of several smaller countries into larger regions, accepting that the overall environmental improvements of individual countries might change. Neither approaches appears practical for political and/or technical (time) reasons. As an alternative approach, the scenarios in this Fifth Interim Report address the main concern of smaller countries, i.e., that the violation potential depends on the size of the country. To overcome this perceived inequity, a (uniform) maximum compensation potential was introduced into the compensation mechanism. This means that environmental targets may only be violated up to a certain amount, which is independent of the country. Experiments showed that such a violation limit was best defined in terms of a uniform minimum gap closure, compared to other relative or absolute measures. In practice, the gap closure optimization with compensation proceeds along the following steps: For each grid cell, a soft target is determined. This soft target is either the AOT60 of the base year (1990) reduced by x percent (for a x percent gap closure) or the AOT60 resulting from the REF scenario, whichever is lower. The AOT60 after the optimization may exceed the soft target in a grid, if the excess AOT60 (weighted by the population in the grid) is fully compensated by overachievements of the soft targets at other grids in the same country (again populationweighted). The AOT60 after the optimization may not exceed, however, (a) the absolute AOT60 target (except in the worst year); this guarantees that the absolute AOT60 target is maintained after the compensation; (b) the AOT60 of the REF scenario. This prohibits a deterioration of the environmental situation compared to the REF (no further measures) case; (c) and it must satisfy a minimum gap closure of y percent (to prevent unlimited compensation). For the AOT60, the country balances (of the excess population exposure indices) extend not only over all grids of a country, but also over all five meteorological years. This means that (a) for the gap closure approach the worst meteorological year is also considered in the optimization, and (b) that excess in some years may be compensated by additional improvements in other years. In addition, a lower cut-off for the AOT60 of 0.4 ppm.hours is introduced 3. This means that the minimum target is set at 0.4 ppm.hours, and that improvements below the level of 0.4 ppm.hours are not allowed to compensate violations at other grids. The major argument for this cut-off is the fact that the regression model used for calculating the AOT60 (see Section 2.5 in Part A) is not valid below this level. Model artifacts should not be allowed to drive the optimization solution, nor should they justify violations of environmental targets at other grid cells. 3 For the calculations of the Third Interim Report, the cut-off level of 1 ppm.hours was related to the 1990 situation. In this report, the cut-off of 0.4 ppm.hours is applied to the actual target level. For a 60 percent gap closure the outcome is identical. 20

22 The country balances ensure that for each country the population excess exposure indices will be reduced at least by the percentage of the selected gap closure, or phrased differently, that the desired gap closure is achieved for the country population exposure indices rather than for individual grid cells. The mathematical formulation of the optimization problem is provided in Section 2.5 of Part A of this report. 5 Ozone Reduction Scenarios Based on the target setting principles outlined in the preceding Section, the Fourth Interim Report explored the following emission control scenarios: For AOT60: Absolute targets of 3.25, 3.00 and 2.75 ppm.hours, to be achieved in four out of five years; Gap closures of 55%, 60% and 65 percent, and Combinations of these targets. For the AOT40: Absolute targets to the excess AOT40 of 10.5, 10.0 and 9.5 ppm.hours; Gap closures of 30/35/40 percent, respectively, and Combinations of these targets. Combined AOT60 and AOT40: Combinations of the targets listed above. The discussions with the EU Member States following the presentation of the Fourth Inerim Report suggested a wider range of targets to be explored, and raised concerns about the compensation potential of small countries being lower than that of larger countries. The discussions also adopted the concept of addressing targets for health and vegetation protection simultaneously. However, there was a general wish to see the solutions for the individual problems presented separately in order to enable the in-depth analysis of the factors determining the overall outcome of the optimization analysis. Consequently, this Section presents first a range of individual optimization runs for four different healthoriented and three different vegetation-oriented targets, and provides the results for the simultaneous optimization for both targets afterwards. 5.1 Main Input Assumptions The input assumptions are identical for all scenarios presented in this section: the energy development in the EU countries follows the baseline scenario, i.e., projections provided by Member States to the Commission or, if no national submissions were received, the business as usual scenario of DG-XVII. For the non-eu countries, the Official Energy Pathway of the UN/ECE database was used. 21

23 It is furthermore assumed that the current legislation on emission reductions will not be reversed. This means that for the EU country the REF scenario is taken as the upper bound for emissions (i.e., the optimization may not result in higher emissions than produced by the REF scenario), and the cost curves for emission reductions consider only the measures not already taken in the REF scenarios (see also Section 3 in Part A of this report). The same assumption holds for the non-eu countries. However, since Part B of this report focuses primarily on ozone in the EU Member States, no further reductions are assumed for non-eu countries in this Section, which fixes the emissions of these countries to the REF level. ECE-wide ozone scenarios are presented in Section 6.3. A further assumption relates to emissions from ships on the sea, which are kept unchanged at the REF level. It should be mentioned that the calculations presented in this report consider only NO x emissions from marine vessels and ignore, due to a complete lack of data, all VOC emissions from this category. 5.2 Baseline Scenarios Targeted at Ground-level Ozone The main objective of this Fifth Interim Report is to assess possible ozone scenarios which could serve as a basis for the Commission s proposal on an EU Ozone Strategy and for the Ozone Directive. The recent discussion process with Member States led to the consensus that the ultimate scenarios considered for the proposal should simultaneously consider targets for the protection of human health and ecosystems. Such scenarios will be presented in Sections and However, in order to provide transparency about the main driving forces and environmental targets leading to certain emission reduction requirements, the following two sections will present scenarios which address the two problems separately Scenarios Targeted at Health Protection (E1) It was mentioned before that the environmental long-term target for the protection of human health established in the Position Paper of the Commission is the full achievement of the WHO health guidelines value. It was also explained in preceding sections of this report that, for practical modeling purposes, the AOT60 is used as a surrogate indicator for the optimization analysis in this report, and that numerical values of the AOT60 indicator must not be used to derive conclusions about actual health risks. Earlier Interim Reports highlighted the large spatial differences in the severity and in the potential improvement of the ozone exposure in Europe. To achieve internationally balanced reduction efforts, a two-track concept for setting environmental targets was introduced: A uniform absolute AOT60 exposure target for all countries (in order to control ozone in the worst polluted areas), and a gap closure requirement, which forces everywhere an equal relative environmental improvement towards the full achievement of the long-term target. 22