Final Report. Report to European Commission

Transcription

1 Cost Benefit Analysis to Support the Impact Assessment Sulphur Content of certain Liquid Fuels Final Report Report to European Commission ED45756 Issue 3 23 rd December 2009

2 Title Customer Cost Benefit Analysis to support the impact assessment accompanying the revision of Directive 1999/32/EC on the sulphur content of certain liquid fuels European Commission Customer reference ENV.C.5/FRA/2006/0071 Service Request No. ENV C5/GK/em/Ares (08) Confidentiality, copyright reproduction and This Report has been prepared in response to Service Request ENV C5/GK/em/Ares (08) Under framework contract No ENV.C.5/FRA/2006/0071 between the Commission and AEA Technology. File reference AEA/ED45756/Issue 3 AEA Gemini Building Harwell IBC Didcot OX110QR UK Telephone: +44 (0) AEA Technology plc is certificated to ISO9001 and ISO14001 Authors Name Peter Bosch, Peter Coenen (TNO), Erik Fridell, Stefan Åström (IVL), Tom Palmer (AEA Technology), Mike Holland (EMRC) Approved by Name Mike Woodfield (AEA) Signature Date 23 rd December 2009 ii

3 iii

4 Table of contents 1 Summary 1 2 Introduction 5 3 Methodology Overview of methods 7 4 Scenario Description Description of the Baseline Scenario Description of the Policy Scenarios 10 5 Dispersion results; concentrations and exceedance of critical loads Introduction SEA SEA2 (applies only to 2015) SEA SEA4 (2020 only) SEA5 (2020 only) Land based A Land based B ASEA1-BSEA1-ASEA4-BSEA Benefits assessment Background Methodology Results 22 7 Cost analysis of the new IMO marine fuel and engine standards and a potential extension of emission control areas Cost calculation method description Abatement Technology Description Unit abatement costs Scenario abatement costs Results 43 8 Comparing Costs and Benefits Preliminary comparison of costs and benefits Detailed uncertainty assessment 53 iv

5 8.3 Summary of the CBA results 58 9 Emission abatement equipment Summary Introduction Abatement of sulphur oxides A comparison between reducing SO 2 emissions with low sulphur fuel or scrubbers Emission criteria for EGC Wash-water criteria Environmental assessments Environmental impact related to the IMO wash water criteria Monitoring Conclusions and suggestions Use of distillate fuels by ships at berth Introduction Outline of the problem Types of ships where the risks may be relevant Relevant number of ships The nature of the risks Technical solutions and costs Conclusions Requirements with regard to maximum sulphur content of heating oil for condensation boilers Objective Introduction to oil fired condensation boilers Heating oil quality and availability Construction of condensation boilers Condensation boilers and sulphur content of heating oil Conclusions References: Appendix 1 - Emission factors used Appendix 2 - Emission Figures by Scenario and Area Baseline Appendix 3 - Relevant geographical representations of the dispersion calculations Appendix 4 - Results per EU member states 125 v

6 17 Appendix 5 - Health Benefit Diagrams Appendix 6 - Methodology for cost calculations 160 vi

7 1 Summary This report documents the Cost Benefit Analysis to Support the Impact Assessment accompanying the revision of Directive 1999/32/EC on the Sulphur Content of certain Liquid Fuels. Maritime activities constitute a significant fraction of anthropogenic emissions of air pollutants in Europe. International shipping in the EU is expected to increase in the coming decade and there is a risk that this will offset the effects of the current emission reduction legislation for maritime activities and the future emissions will surpass the target levels established by the European Commission in its Thematic Strategy on Air Pollution for land based sources. A number of ongoing developments will lead to reductions in emissions of sulphur oxides (SO x ) and nitrogen oxides (NO x ). In this report the effect of these developments is determined in order to inform the Commission s review of Directive 1999/32/EC relating to a reduction in the sulphur content of certain liquid fuels. This study has examined the effects of emission reduction measures (for the years 2015 and 2020) in maritime activities by: 1. Analysing the benefits of: the new IMO marine fuel and engine standards (SO 2 and NO x reduction respectively), and the potential extensions of emission control areas. The benefits of different scenarios have been evaluated against a revised NECD 2008 baseline scenario. The study did not take into account the effects of the current economic crisis (assuming this will only be temporary). The benefits were quantified with respect to a baseline scenario for emissions, air quality (concentrations) as well as their impact on human health. 2. Calculating the costs of emission reductions that result from the new IMO marine fuel and engine standards and the potential extension of emission control areas. 3. Assessing the effects, costs, risks and technical requirements that are associated with the use of emission abatement equipment and steam boilers by ships as well as the use of condensation boilers for heating purposes. Benefit analysis The following scenarios were investigated and were compared to a Baseline scenario (revised NECD 2008 baseline): Designation of the Baltic Sea, North Sea and the English Channel SECAs ( and ) Inclusion of Mediterranean as a SECA (2015) Inclusion of the Black Sea as a SECA (2015 and 2020) Reduction of land based sources 4 by reduction of the S-content in gasoil to 0.05 % (2015 and 2020) Reduction of land based sources by reduction of the S-content in gasoil to % (2015 and 2020) For each scenario the emissions of relevant air pollutants were calculated, and the effects on ambient air quality were quantified using state of the art modelling tools. These concentrations were used to determine the health benefits of the emission reduction using the methods developed in the CAFE and EC4MACS studies for EC DG-ENV. Cost analysis The cost calculations for NO X and SO 2 abatement from shipping in this report contain two parts. The first consists of calculating unit costs (per PJ) for each abatement option under consideration, the second the calculation of scenario specific costs based on the penetration of the low-no X and low-so 2 1 In all 2015 marine scenario s Tier I & Tier II for NOx reduction is assumed 2 In all 2020 marine scenario s Tier I - III for NOx reduction is assumed simulating the designation of the Baltic Sea, North Sea, the English Channel and the Mediterranean as NECA 3 For 2020 also scenario s are elaborated with a reduced S content (0,5 %) in marine fuels used outside the SECA 4 Only for heating in households and agriculture AEA 1

8 technology options. The calculations allow for scenario comparison and comparison of alternative technologies to reach the required standards. For each scenario, costs are calculated, specifying high and low costs for NO X and SO 2 abatement. For SO 2 abatement, separate cost calculations are presented for scrubber costs and low S-fuel costs Comparison of costs and benefits Annual costs and benefits (low and high estimates) for the different scenarios are given in the following tables - taking the results directly from the analysis. The low bound for costs is the lower estimate for fitting a scrubber to reduce SO 2 emissions together with the costs of NO x control. The upper bound for costs is for the high estimate for using low sulphur fuels to reduce SO 2 emissions and the costs of NO x control. The range for benefits is based on use of either the value of a life year (VOLY, giving the lower bound) or the value of statistical life (VSL, giving the upper bound) to value mortality ( billions) Baltic Sea, North Sea and the English Channel as SECA Inclusion of Mediterranean in the SECA Inclusion of the Black Sea in the SECA Costs low high Benefits low high ( billions) Baltic Sea, North Sea and the English Channel as SECA Inclusion of Mediterranean and the Black Sea in the SECA Baltic Sea, North Sea and the English Channel as SECA Reduced S- content outside the SECA Inclusion of Mediterranean and the Black Sea in the SECA Reduced S- content outside the SECA Costs low high Benefits low high In all cases there is no overlap between the estimates for costs and benefits as shown above, indicating that for all scenarios benefits would exceed costs.. Further analysis was undertaken to assess the effect of further uncertainties on the balance of costs and benefits. This found that the probability of benefit exceeding cost can be described as very likely / virtually certain, using terminology developed by the Intergovernmental Panel on Climate Change (IPCC). With one slight exception (the last 2020 scenario: All European seas Included in the SECA and reduced S-content outside the SECA ) the probability of benefits exceeding costs can be described as likely, very likely or virtually certain even for worst case assumptions on cost. It is to be remembered that the probabilities calculated here are quantified only against benefits for improved health. The inclusion of impacts to materials (including cultural heritage) and ecosystems, in particular, would add to the benefits and further increase the already high probabilities. Taking these issues together it is concluded that more detailed assessment of the limitations of the analysis finds that the initial conclusion, that benefits of the proposed actions will exceed costs, is indeed robust. AEA 2

9 Comparison of the benefits of the land based scenarios against the maritime scenario (Baltic Sea, North Sea and the English Channel as SECAs) shows that the additional health benefit from lowering the S content in fuel on land amounts to 0.4 to 0.8 billions (depending on the S content) in For 2020 the additional benefits are calculated at 0.6 billions. These figures are only small compared to the benefits of the maritime scenario (which amount to billions) and therefore do not constitute an alternative to the maritime emission reduction. Technical criteria for exhaust abatement equipment Exhaust abatement equipment for the reduction of emissions of sulphur dioxide from ships is an alternative to reduced S content in marine fuels. The only technology available at present is scrubbing. Scrubber units can be of two kinds, seawater scrubbers and closed scrubbers, or a combination of two. The use of scrubbers to clean the exhaust from marine engines using residual oil could in principle lead to high concentrations of a number of harmful compounds in the marine environment around the ships. The present situation is somewhat incongruous in the sense that washwater limits for scrubbers are set by IMO while there still are no limits for the gaseous exhaust (except for SO 2 and NO X ). It is recommended that criteria are developed that promote the abatement of PM. At least the levels of PM reduction should be established to allow for voluntary or market based reductions of PM emissions. Further, it is recommended, when possible, to keep the EU regulations in line with the IMO guidelines. The present IMO washwater criteria contain limits for PAH, ph, turbidity and nitrates. Of these all except the nitrate criterion are intended for continuous monitoring onboard the ship. A major factor when formulating the criteria is that they should be possible to monitor online. However, it would also be possible to add more criteria that are to be tested during certification. This could give further assurance that the units are environmentally safe and would not be constrained by the possibility to measure online in a simple way. With the present criteria for ph it seems as if dilution and buffering in the sea will rapidly give acceptable levels. Even in confined areas it is likely that the effect is limited. However, it is recommended that further studies are undertaken for brackish and river water and that the criteria then are reviewed. It is suggested that the IMO guidelines are adopted for the ph criteria but at the same time allowing for revisions as more data becomes available. The criteria for PAH are expressed as phenanthrene equivalents. This is not a well established concept and should be defined in the guidelines. It may be appropriate to use phenanthrene for monitoring purposes but the criteria ought to guarantee low emissions of a range of PAHs and thus it is necessary to establish how the measured value relates to e.g. PAH 16 or THC. The criteria for turbidity may be used to establish that unwanted solid material is not formed in the scrubber unit. However this criterion is not appropriate to monitor PM reduction because there is no relation between particle concentration and turbidity. It is recommended that the concentrations of particulate matter, metals, and PAH are regulated and tested during commission, and, further, that the correlation to turbidity is investigated in order to have continuous measurement available onboard. The present criteria for nitrates could in principle lead to large additional contributions of nitrates to sensitive areas. The best way to deal with this issue is to use abatement technology for NO X reduction. It is recommended to set up new criteria regarding nitrate emissions which prevent increased nitrogen loads to sensitive areas. It seems appropriate to adopt the IMO guidelines regarding monitoring, i.e., allowing for the possibility of not using continuous monitoring of the SO 2 emissions. The reason is that there will be units that can be sufficiently monitored by other means. There is only limited experience from trials with marine scrubbers for SO 2. It is important to follow the trials taking place and make sure that thorough environmental impact assessments are conducted in order to evaluate and develop the washwater criteria. AEA 3

10 Use of distillate fuels by ships at berth From 1 January 2010 ships using ports within the EU will be required to use low-sulphur fuel, maximum 0.1 wt% sulphur, while at berth, with an exception for those making only short stops (Directive 2005/33/EC). This will most likely make it necessary to use distillate fuel while at berth. There are concerns that there are risks associated with switching from residual oil to distillate fuel in boilers. One can state that there are real safety concerns with switching from HFO to MGO in existing boilers. The main safety concern is that a distillate fuel may vaporise in hot boilers with heated pipes. If the boiler system has not been modified, there is in the case of flame failure and improper handling an explosion risk. Although many boilers can be used with MGO, all boilers should be inspected before using MGO at berth. There are technical solutions to solve these issues where fuel piping systems, burners, and flame inspection systems are modified. The control system may have to be modified and new handling routines established. For LNG ships, modifications are required. The costs range from to for small auxiliary boilers, around per ship for tankers and to for LNG tankers. Requirements with regard to maximum sulphur content of heating oil for condensation boilers A typical issue concerning condensation boilers used for space heating is use of the latent heat of the water vapour in the exhaust gases. This increases the fuel efficiency and therefore decreases the CO 2 emission by 6% compared to conventional boilers. Because the flue gases are cooled down below the dew point, condensation will occur on the heat exchanger in the boiler and using heating oil will cause an acidic (and thus corrosive) condensate. It is suggested that this phenomena sets limits for the maximum sulphur content of heating oil to be used in condensation boilers. In this study its is established that heating oils with sulphur content up to the maximum allowed concentration of 1000 mg/kg can be used in condensation boilers. However, to maintain high efficiency, heat exchangers have to be cleaned more frequently than in conventional boilers. Using low sulphur oil will prevent this. Furthermore are there no technical obstacles which prevent the use of low sulphur heating oil in condensation boilers designed for standard heating oil. AEA 4

11 2 Introduction Maritime activities constitute a significant fraction of anthropogenic emissions of air pollutants in Europe. In 2000, SO 2 and NO x emissions from international maritime shipping in Europe amounted to approximately 30 percent of the land based emissions in the EU-25. While legislation is in force to control emissions from international shipping, the expected increase in the volume of ship movements will offset the positive environmental impacts of these measures and will lead to further growth in ship emissions. Under business-as-usual assumptions, by 2020, emissions from maritime activities would come close to the projected baseline emission levels from land based sources and surpass the target levels established by the European Commission in its Thematic Strategy on Air Pollution for land based sources. This anticipated increase in ship emissions will counteract the envisaged benefits of the costly efforts to control the remaining emissions from land based sources in Europe. While at present emissions from ships are responsible for 10 to 20 percent of sulphur deposition in coastal areas, their contribution is expected, by 2020, to have increased to more than 30 percent in large areas in Europe, and up to 50 percent in coastal areas. Analysis by IIASA in 2007 demonstrated that limiting air pollution from shipping reduces the necessity to further control emissions from land based sources and provides important cost savings in achieving air quality targets in Europe. A number of ongoing developments will lead to reductions in emissions of sulphur oxides (SO x ) and nitrogen oxides (NO x ), these include: Annex VI of MARPOL and the Technical Code on Control of Emissions of Nitrogen Oxides from Marine Diesel Engines (the new IMO marine fuel and engine standards) from the International Maritime Organisation (IMO) have been revised and formally adopted by the Marine Environmental Protection Committee (MEPC) of IMO in October These revisions set criteria for a gradual reduction in the maximum permitted sulphur content of marine fuels as well as the maximum permitted emissions of nitrogen oxides from marine diesel engines; Annex VI of the Marine Pollution Convention (MARPOL) defines Sulphur Emission Control Areas (SECAs) where stricter limits on sulphur emissions apply. As of now, two areas are classified as SECAs: the Baltic Sea and the North Sea/English Channel. Article 7 of Directive 1999/32/EC requires the Commission to give consideration to proposals for the addition of SECAs. The possibility that European institutions will designate emission control areas for NOx; The agreement by the European institutions on the Climate and Energy Package. (26 March 2009) The main purpose of this report is to determine the effects of these developments in order to inform the Commission s review of Directive 1999/32/EC relating to a reduction in the sulphur content of certain liquid fuels. This study has examined the effects of emission reduction measures in maritime activities by: Analysing the benefits of the new IMO marine fuel and engine standards as well as the potential extensions of the emission control areas. The benefits of different scenarios have been evaluated against a revised NECD 2008 baseline scenario. All existing and foreseeable Community and national measures were taken into account, as far as these have a direct impact on emissions and impacts on future activity are known. The study did not take into account the effects of the current economic crisis. It has been assumed that the crisis is only temporary (noting apparent recovery of several major Member State economies in the second half of 2009) and that the economy will subsequently evolve as was expected before the crisis. The benefits were quantified with respect to the baseline scenario for emissions, air quality (concentrations) as well as the impact on human health. Calculating the costs of emission reductions that result from the new IMO marine fuel and engine standards and the potential extension of emission control areas. AEA 5

12 Assessing the effects, costs, risks and technical requirements that are associated with the use of emission abatement equipment and steam boilers by ships as well as the use of condensation boilers for heating purposes. AEA 6

13 3 Methodology 3.1 Overview of methods In general terms the cost benefits analysis consists of the different steps illustrated in figure 3.1. Figure 3.1 Flow scheme for the cost benefit analysis Base line & Policy scenarios EMISSIONS tonnes/year DISPERSION Increase in ambient concentrations in all affected regions - Prices different fuel grades - Abatement costs Stock at risk e.g. inventory building materials, population distribution EXPOSURE People, crops, materials IMPACT Using exposure-response curves, e.g. change in building erosion COSTS FOR EMISSION REDUCTION BENEFITS Heath benefits, Damage costs COMPARISON of COSTS & BENEFITS In the first step the baseline scenario is developed. All other policy scenarios will be compared to this baseline. The emissions of SO 2, NO x, PM, CO and NMVOC for these scenarios are calculated and fed into the dispersion model. The scenario descriptions and emission estimates are given in section 4. The most relevant parameters for the dispersion modelling are given in section 5. The assessment of the benefits is elaborated in section 6 and section 7 describes the cost calculations for the emission reduction. In section 8 a comparison is made between costs and benefits based on the results of the two previous sections. AEA 7

14 4 Scenario Description 4.1 Description of the Baseline Scenario The baseline for the study is traceable to the NEC6/C&E package/current policies baseline scenario from IIASA/GAINS, which has been used as baseline for simulations of the NECD plus the proposals for the Climate and Energy Package of the Commission of 23 January It includes the most up to date (Spring 2008) current control policies as laid down in legislation and those presently in the decision phase of European Institutions. It includes: 'Current legislation' in all sectors plus: revision of the IPPC Directive for large stationary combustion sources in industry and power sector (less strict BAT case interpretation); Euro VI on heavy-duty vehicles and Ceilings from the NEC Directive. Assumptions are documented in detail in the NEC6 report, page 18 and 19 (Amman et al, 2008). As the NEC6 scenarios do not fully cover international shipping (they contain national maritime fuel consumption but not international bunkering) a separate baseline has been developed for international shipping. The starting point for the fuel consumption of international shipping is the 2000 fuel consumption - from Cofala et al (page 11). The fuel consumption is given by fuel type (MD/HF), vessel type (cargo/ferry) and regional sea. It is based on the 2000 inventory of emissions produced by ENTEC, which is in itself based on a detailed analysis of four months of vessel movement data from the year Cofala applies an overall 2.5 per cent growth rate for cargo vessels and 3.9 per cent rate for passenger vessels in the baseline whereas, for this study, we adopt the baseline of the Ex-tremis model (Chiffi et al., 2007), which is based on a detailed model of shipping in EU27 countries and is differentiated for marine diesel and heavy fuel oil. The growth factors for bunker fuel consumption from Ex-tremis (for MD and HF) for EU ship movements are used to project the regionally detailed fuel consumption from international shipping in European seas in 2010, 2015 and Over the period up to 2020 the overall growth in HF use is 2.7% per year, while use of MD decreases with 0.5% per year We have also used the outcomes of the Ex-tremis baseline to aproximate the fuel consumption by ship types and engine size. Ex-tremis only covers EU27 countries and misses the traffic over European seas generated by non-eu countries in Europe (Chiffi et al., 2007). Assuming, however, that the developments in transport and developments in the fleet in EU and outside EU will run in parallel, it is possible to apply the distribution of ship types (bulk carrier, tanker, containership, general cargo, etc.), and engine class (gas turbine, high speed diesel, medium speed diesel, slow speed diesel, steam turbine) in each of the years to the regional fuel consumption for international shipping in 2010, 2015 and Emission factors have been added by detailed ship-classification and by regional sea. The baseline is a reference scenario and does not include any policies - it is based on fixed emissions factors and, for instance, does not include the 1997 IMO protocol, introducing Tier I for ships built after 2000 and which took effect in Neither does it include the Tier II regulations that apply for ships built after As a result the emissions reductions in scenario SEA1 compared to the baseline suggest a relatively larger shift than would occur in reality. An overview of emission factors used is provided in Annex 1. It should be noted that the sulphur content of HF in the baseline outside SECAs is 2.70% (according to TFEIP, 2009), which is the current worldwide average. Inside SECAs it is 1.45% representing the present regulation of 1.5%. This is used to be able to illustrate the impact of the agreed limit of 0.1% from 2015 that is used in one of the scenarios (SEA1). The emission factors for particulate matter have been adapted from Cooper and Gustafson (2004) and have been chosen according to the sulphur content of the fuel and the speed of the engines. Table 3 in Annex 1 specifies the emission factors used in the emission calculations e.g. for slow speed diesel engines running on heavy fuel oil the reduction in the PM emission factor resulting from a decrease in sulphur content in HFO from 1.5% to 0.1% is 54%. As it is difficult to model the introduction of scrubbers, we have not introduced a detailed calculation differentiating the part of the sulphur limit that can be reached by using scrubbers and the part that can be reached by using low sulphur fuels. For the calculation of the sulphur emissions the way emissions reductions are realised does not make a difference. It does, however, create uncertainty in the AEA 8

15 calculation of PM emissions. Although there are few measurements to substantiate this, scrubbers reduce PM emissions roughly by half (see also Section 9.4). This reduction is in the same order of magnitude as the reduction in the emission factors for PM, when going from 1.5 to 0.1% S in HFO for the majority of the vessel; consequently the error introduced by not accounting for the share of scrubbers in the PM calculation is limited. The resulting baseline emissions per sea area are given in Table 4.1a. Table 4.1a Baseline emissions international shipping this report, 2000 and Sea area SO 2 NO x NMV OC PMtotal SO 2 NO x NMVO C PMtotal x 1000 tonnes North Sea Black Sea Mediterranean Baltic Sea NE Atlantic Total The 2000 inventory of emissions produced by ENTEC was used to distribute the emissions per sea area to generate the gridded data needed for the dispersion modelling. The national data from the IIASA-GAINS baseline are also gridded according to sector specific distribution patterns. The land based emissions are the same as in the C&E package current policy scenario in Amman et al. (2008). The baseline sea-based emissions are more or less comparable to those in Cofala et al. (2007). Differences, however, are caused by the split in ship types we have introduced by using Ex-tremis, which brings about a split in HF and MD that differs from the split used by Cofala et al. (2007). Further differences are brought about by the larger differentiation in emission factors for NO x, NMVOC, CO and PM which we have applied for different engine types. Cofala et al. (2007) assume lower sulfur and PM emissions in 2020 (See table 4.1a and 4.1b) Table 4.1b Baseline emissions international shipping according to Cofala et al, 2007, 2000 and Sea area SO 2 NO x NMV OC PM2,5 SO 2 NO x NMVO C PM2,5 x 1000 tonnes North Sea nr 67 Black Sea nr 10 Mediterranean nr 198 Baltic Sea nr 28 NE Atlantic nr 91 Total nr 394 Note: nr = not reported. AEA 9

16 4.2 Description of the Policy Scenarios Overview Policy scenarios have been designed to enable the critical comparison of the effects of policy measures that restrict emissions by: The (already decided) designation of the Baltic and the North Sea/English Channel as SECA The designation of the Mediterranean Sea as SECA The designation of the Black Sea as SECA Lowering the maximum sulphur content of maritime fuels outside designated SECA areas to 0.5% Application of the IMO Tier I and Tier II standards Designating the Baltic, the North Sea/English Channel and the Mediterranean as NO x emission control areas (in which IMO Tier III would apply to new engines from 2015)). Additionally, two scenarios for decreasing the emissions from land based sources have been elaborated in which the sulphur content in heating oil for use by households and agriculture has been assumed to be reduced to 0.05 % and 0.001% respectively (see Table 4.2 and 4.3). Table 4.2: Scenario s for 2015 S-content Scenario Year marine fuel outside SECAs (by mass) S-content marine fuel in SECAs (by mass) NO x Tier Standards SECA regions Baseline % 1.45% SEA % 0.10% Tier I & II SEA % 0.10% Tier I & II SEA % 0.10% Tier I & II Baltic Sea North Sea / English Channel Baltic Sea North Sea / English Channel Mediterranean Sea Baltic Sea North Sea / English Channel Mediterranean Sea Black Sea The baseline and SEA1 scenario were also combined with reduced emissions for land based emissions Land based A % S in MD and Gasoil for use by Consumers and Agriculture (Abase and ASEA1) Land based B % S in MD and Gasoil for use by Consumers and Agriculture (Bbase and BSEA1) AEA 10

17 Table 4.3: Scenario s for 2020 Max. S-content Scenario Year(s) marine fuel outside SECAs (by mass) Max. S-content marine fuel in SECAs (by mass) NO x Tier Standards SECA regions Baseline % 1.45% Baseline scenario (Task A1) 1) Baltic Sea SEA % 0.10% Tier I & II,& III North Sea / English Channel SEA % 0.10% Tier I & II,& III SEA % 0.10% Tier I, II & III SEA % 0.10% Tier I, II & III Baltic Sea North Sea / English Channel Mediterranean Sea Black Sea Baltic Sea North Sea / English Channel Baltic Sea North Sea / English Channel Mediterranean Sea Black Sea The SEA4 scenario was also combined with reduced emissions for land based emissions Land based A % S in MD and Gasoil for use in Consumers and Agriculture (ASEA4) Land based B % S in MD and Gasoil for use in Consumers and Agriculture (BSEA4) Note: 1) Simulation of NECA in the Baltic Sea, North Sea/English Channel and the Mediterranean Sea (Tier III standards for ships constructed on or after 1 January 2016) Comparing the scenarios with each other reveals the relative benefits of individual policy measures. For example, the comparison of SEA1 with SEA2 illustrates the impact of designating the Mediterranean Sea as SECA, the comparison of SEA1 with SEA3 illustrates the impact of designating the Mediterranean Sea and the Black Sea as SECA, the comparison of SEA2 with SEA3 illustrates the additional impact of designating the Black Sea as SECA, etc SEA1 The SEA 1 scenario uses the baseline for the land based emissions. For the sea based emissions the maximum S content within SECAs is reduced from the 1.45%, which is the post 2006 standard, to a maximum of 0.1%; the designation of SECAs remain the same as in the baseline scenario - namely the Baltic Sea and the North Sea/English Channel. It is expected that the additional demand for low sulphur fuels for use in the SECAs will lead to a shift towards high sulphur fuel outside SECAs; a rough estimate, based on the worldwide fuel consumption in and outside SECAs whereby the total amount of sulphur is kept constant, results in an increase of the current average sulphur content outside SECAs from 2.7 to 2.9%. The recently revised 5 IMO standards have been used for the NO x emissions; for 2020 we have applied the Tier III standards for ships constructed after 1/1/2016 in the Baltic Sea, the North Sea/English Channel and the Mediterranean Sea areas. The age distribution of the population of ships has been estimated based on the age distribution of the world merchant fleet by type of vessel as of 1st January 2005 (Source: Chiffi et al 2007, Technical 5 Revised Annex VI of MARPOL and the revised NOx Technical Code adopted in October 2008, these will enter into force 1 July AEA 11

18 Annex, Maritime Transport Inventory, Table 2.5). In the absence of better data it has been assumed that the age distribution will not change over the period up to Ships constructed on or after 1 January 1990 but prior to 1 January 2000 (and with a displacement 90 litres per cylinder and rated output 5000 kw) will be required to comply with the NO x emission limits in force today, subject to the availability of approved engine upgrade kit (Tier I). After contacts with marine engine producers MAN and Wärtsila we estimated that about 1/3 of the pre-2000 engines are already in compliance with IMO Tier 1. Thus, for 70% of the ships constructed from January 1990 to January 2000 the baseline emission factors have been used, while for the rest of this age cohort Tier 1 emission factors have been applied SEA2 The SEA2 scenario only differs from SEA1 in the designation of the Mediterranean as a SECA region; it has been calculated for 2015 only SEA3 The SEA3 scenario adds the Black Sea as a SECA in comparison to SEA2. In the scenario run for 2020 it is additionally assumed that the Baltic Sea, the North Sea/English Channel and the Mediterranean will be designated as NO x Emission Control Areas, in which Tier III standards apply SEA4 For the SEA4 scenario, which has been calculated for 2020 only, the sulphur content of fuels used outside SECAs has been reduced to 0.5%. For the rest SEA4 is comparable to the 2020 run of SEA1, with only the Baltic and the North Sea/English Channel designated as SECAs and the Baltic, the North Sea/English Channel and the Mediterranean Sea designated as NO x Emission Control Areas SEA5 The SEA5 scenario is the most advanced scenario: The sulphur content of fuels used outside SECAs has been reduced to 0.5%; all European seas (except the remaining part of the North Eastern Atlantic Ocean) have been designated as SECAs allowing a maximum sulphur content of 0.1% only. The Baltic Sea, the North Sea/English Channel and the Mediterranean Sea have been designated as NO x Emission Control Areas, implying IMO Tier III standards Land based A Apart from regulating the quality of fuels used by maritime activities, Directive 1999/32/EC on the reduction of, also covers medium distillates and gasoil used in small scale combustion on land. Combustion in industry falls under the large combustion plants directive (Directive 94/66/EC). The current limit for gasoil in the fuel quality directive is a sulphur content of 0,10 % (since 1 January 2008). Scenario Land based A assumes a further reduction of the sulphur content used of gasoil by domestic heating and agriculture to 0.05%. As the baseline scenario includes emissions from all European countries, it is necessary to mention that this reduction only has been applied for the EU27 countries only. The LandA scenario has been combined with the baseline sea emissions (ABASE) and the SEA1 scenario (ASEA1) Land based B The Land based B scenario takes the reductions a step further, limiting the sulphur content of gasoil by domestic heating and agriculture in EU27 to 0.001%. The Land based B scenario has been combined with the baseline sea emissions (Bbase) and the most stringent sea scenario SEA4 (BSEA4). 6 This results in a conservative estimate of the NOx emissions AEA 12

19 4.2.9 SO 2 emissions in the different scenarios The emission figures for SO 2 in the different areas and scenarios as calculated in this study are summarised in the Tables 4.4.and 4.5. Appendix 2 gives the detailed emission figures for all scenarios all areas and all pollutants. Table 4.4 Sulphur dioxide emissions, , all scenarios Baltic Sea Black Sea Remaining NE Atlantic ocean x 1000 tonnes baseline SEA SEA SEA SEA SEA Mediterranean Sea North Sea X 1000 tonnes baseline SEA SEA SEA SEA SEA Note: In SEA1 and SEA2 the sulphur emissions in non-secas are higher than in the baseline because it is assumed that the demand for low sulphur fuels in SECAs will push up the average sulphur content in other seas. Table 4.5 Sulphur dioxide emissions land based sources, baseline and scenarios x 1000 tonnes Baseline Land based A Land based B Note: exclusive of emissions related to national maritime fuel consumption (e.g. fisheries), which are included in the GAINS-totals Source: GAINS Europe, NEC6, C&E package current policy AEA 13

20 5 Dispersion results; concentrations and exceedance of critical loads 5.1 Introduction The LOTOS-EUROS model was used to compute (from the emission data as presented in the previous section) the atmospheric dispersion of ship emissions, and their effect on the European distribution of sulphur- and nitrogen deposition, and on ground-level ozone and particulate matter concentrations. The domain of the LOTOS-EUROS model (Schaap et al., 2005) runs from 10 West till 60 East and from 35 North till 70 North; it covers the whole of Europe (Iceland excluded) with a horizontal resolution of 0.5 latitude x 0.25 longitude (~30x30 km). In vertical direction it extends to 2-3 km above the surface resolved in four layers: a planetary boundary layer with a fixed surface layer of 25 metre height, and two reservoir layers on top of the planetary boundary layer. The background concentrations (from other parts of the world) are provided by a global model, TM5. Transport is assumed to be by advection and diffusion and is based on analysed meteorological fields. The dispersion calculation uses1997 meteorology - to stay in line with two preceding studies in 2006 and 2007 concerning the trade of emissions of NO x and SO 2 ; in addition 1997 was put forward by TNO (and supported by a.o. EMEP) as a year which represents an average atmosphere over Europe. Dry deposition is described via a multiple serial and parallel resistance approach (surface-to-air, laminar and turbulent resistance) and wet deposition is described by scavenging rates and precipitation data. The chemistry is based on a modified CBM-IV scheme in combination with a thermodynamic equilibrium module and it describes the chemistry for ozone, acidification, eutrophication and aerosols. Anthropogenic emissions of 7 species (NO x, SO 2, NMVOC, CH 4, PM 10, PM 2.5 and NH 3 ) are based on Gains Europe (NEC report Nr. 6). Emissions of the remaining component CO come from TNO s own emission database. Furthermore, the model accounts for natural emissions from trees (isoprene, terpene), and from soil (NO x ). The standard output (on an hourly basis) 7 is generated as a longitude-latitude grid which was subsequently converted into the EMEP grid. Ecosystem burden is expressed in terms of the sulphur and nitrogen deposition and the average accumulated exceedence (AAE) for acidification and eutrophication. The model also delivers: concentrations of primary PM 10 and PM 2.5 and secondary species as sulphate, nitrate and ammonium; concentrations of ground-level ozone are expressed in the four ozone metrics - AOT40c, AOT40f, SOMO35 and annual mean. The results of this modelling exercise will be discussed in the next paragraphs. Emphasis is laid on the comparison between the baseline results and the specific scenarios and the difference between the scenarios and changes between 2025 and The most relevant results in terms of geographic distribution over Europe for the exceedence of critical loads for eutrophication and acidification, the distribution of ozone and PM 10 are presented in maps in 7 The model has been used in several model inter-comparison studies and it shows a good performance (Roemer et al., 2003; van Loon et al., 2004). It was used in earlier studies on SO 2 and NOx emission trading (Roemer et al., 2006, van Harmelen et al 2007). During the first EU study the model output for 2010 (meteorology: 1997) was compared with the results from the EMEP model (See also appendix 3). LOTOS-EUROS showed somewhat (~20%) lower sulphur deposition, mainly caused by differences in the wet deposition. In a current study it appears that the difference in total sulphur deposition over Germany between LOTOS-EUROS and EMEP is very small. The nitrogen deposition is somewhat (+30%) higher in LOTOS-EUROS than in EMEP. The PM 10 concentrations in LOTOS-EUROS are about a factor of two lower than the measured values. Underestimation of PM 10 values is a common problem for nearly all large scale models. The problem occurs with the primary PM10 concentrations (the part of PM 10 that is emitted directly). Quite a good comparison with measurements is seen for secondary PM 10 (sulphate, nitrate). AEA 14

21 appendix 3. Appendix 4 includes tables of the dispersion calculations aggregated to the different EU member states. 5.2 SEA1 The main feature of this scenario is the lowering of the sulphur content in the Baltic and the North Sea/English Channel SECAs, and the implementation of IMO Tier I and II for These measures result in a reduction of exceedance (level and area) of the critical load for acidification in all vulnerable areas in Europe. In coastal areas (up to ca 100 km from the coast) of the Baltic and the North Sea the exceedances have been reduced considerably (by up to 100%) whereas the reduction for inland areas is10-20%. In 2020, compared with 2015, acid deposition has further decreased, due to lower NO x emissions, leading to a larger coastal area that is protected. In inland areas, the gains compared to the 2015 baseline are more limited. As an effect of implementing IMO Tier I and II, the NO x emissions decrease in all European seas, resulting in a similar decrease of eutrophication all over Europe (0-10%). Larger fractional reductions are apparent in those coastal zones of the Mediterranean where shipping is a major source of NO x and exceedances of critical loads are already low - the absolute reduction, however, is the same as elsewhere in Europe. For 2020 additional reductions are expected as a result of implementing IMO Tier III in the Baltic, the North Sea and the Mediterranean. Both in 2015 and 2020, the concentrations of PM 10 are reduced by 20-30% (in the coastal areas around the North Sea, English Channel and Baltic Sea). PM 10 concentrations in the northern part of Europe decrease by up to 10 % but in the Mediterranean region PM concentrations rise as a result of increasing, uncontrolled, sulphur emissions. The increase in sulphur emissions (as sulphate) outweighs the simultaneous decrease in NO x (as nitrate). All round the North Sea and the English Channel the concentrations of ozone show an increase, which is the result of the decreasing NO x emissions due to the introduction of IMO Tier I and II in a NO x rich environment. Similar trends in modelled ozone response to a limited decrease of NO x emissions are seen in urban environments and other NO x rich conditions over land. In the Mediterranean, where NO x concentrations are much lower, the reduction in NO x emissions leads to lower ozone concentrations - as expressed by SOMO35 and AOT40f. By 2020 the effect of introducing IMO Tier III can be clearly seen (Annex 2) in the Mediterranean Sea and the Baltic but less so in the North Sea because of the higher ambient NO x concentration. AEA 15

22 5.3 SEA2 (applies only to 2015) SEA2 differs from SEA1 in that the Mediterranean Sea too is designated as a SECA. By 2015 this addition results in an extra reduction of the exceedance of eutrophication by 10% and acidification by 20%. In the Mediterranean area the reductions to eutrophication can be locally larger than 20 % because maritime NO x emissions here are a larger determining factor of ambient concentrations of NO x. The reduction of PM 10 concentrations in the Mediterranean area is also considerable; on shipping routes in the southern part of the Mediterranean Sea reductions up to 70% are seen whereas the concentrations over the rest of the Mediterranean Sea decrease by 30 40% and in coastal areas up to 20 %. As a result of the significant reduction of SO 2 emissions, compared to SEA 1, the ozone concentrations in the whole Mediterranean area (land and sea) increase slightly (see difference between the following maps) as the result of heterogeneous chemistry in which sulphur reduction causes a slight increase of nitrogen oxides through the formation of additional N 2 O 5. AEA 16

23 5.4 SEA SEA3, by comparison with SEA2, assumes that the Black Sea is also designated as a SECA. Although this results in a relatively minor additional emission reduction in Europe as a whole, and a negligible effect in Western Europe states, the reductions in exceedance of critical loads for acidification and eutrophication in areas close to the Black Sea (Romania and Bulgaria) is significant (up to 30% in some grid cells). Reduction of PM 10 is only significant over the sea and apart from a small reduction in PM 10 concentrations in Bulgaria it has a very small effect over the land. The ozone effect is also small and resembles that of SEA2 on the Mediterranean area; the Black Sea, however, is much smaller and the rise in ozone concentration more limited The SEA3 scenario in 2020 includes emission reductions (as compared to SEA1) in both the Mediterranean Sea and the Black Sea. The effects are quite comparable to the effects of the two separate reductions combined as discussed under SEA2 and SEA3 (2015). 5.5 SEA4 (2020 only) The most important difference in SEA4 from the preceding scenarios is the lowering of the sulphur content of fuels used outside SECAs to 0.5%. In SEA4 the non-seca areas are the Mediterranean Sea, the Black Sea and the Remaining North-Atlantic. The measures result in a large reduction of the sulphur emissions in these areas as compared to SEA1 scenario. For acidification the effect of SEA4 is mainly seen in the UK, which is heavily influenced by the emissions in remaining North-Atlantic. Emission reduction in the Mediterranean Sea and the Black Sea has little effect in reducing the exceedance of acidification due to the location and distance of sensitive ecosystems. For eutrophication the benefits of SEA4 (as compared to SEA1) are relatively small since NO x emissions stay the same. PM 10 concentrations are lowered as a result of the PM 10 emission reductions following the reduction in sulphur emissions. The largest changes are seen in the coastal areas of the three seas involved. Due to the marked reduction in sulphur emissions there is a slight increase in the ozone (NOTE: this phenomenon was described under SEA2). AEA 17

24 5.6 SEA5 (2020 only) In SEA5, apart from the lower sulphur content outside SECAs, the Baltic, the North Sea/English Channel, the Mediterranean and the Black Sea have been assumed to be SECAs, resulting in a further reduction of sulphur and PM emissions. Of all the sea-based scenarios SEA5 yields the least exceedance of acidification and eutrophication and the lowest PM 10 concentrations. The ozone related parameters are slightly increased but still less than with the baseline ozone. 5.7 Land based A 2015 The emission reductions of the land-based measures are relatively small compared to the reductions achieved by most of the marine-based measures. The land-based measures apply only to one of the source categories of sulphur (domestic heating and heating in agriculture). In quite a few countries this source contributes very little but in others it is larger. Scenario A results in ~2-3% less sulphur deposition and a small decrease of the exceedance of acidification, the effect on levels of eutrophication is really very small, less than 0.5%; the PM 10 concentrations drop about 1% (as a result of less sulphate formation and less direct PM emissions), and the effects are negligible for ozone related parameters. 5.8 Land based B 2015 The land based measures taken under Scenario B are relatively stringent; the reduction in the sulphur deposition is about twice that envisaged under Scenario A. Scenario B leads to ~4-6% less sulphur deposition, which results in a rather small decrease of the exceedance of acidification. For eutrophication the effects are less than 0.5%; PM 10 concentrations drop by about 2% (less sulphate and less direct PM emissions), and the effects are negligible for ozone related parameters. 5.9 ASEA1-BSEA1-ASEA4-BSEA4 The land based scenarios A and B are combined with the SEA1 scenarios (2015 and 2020) and with the SEA4 scenario (2020). The effects of these scenarios compared to the SEA1 and SEA4 scenarios (without additional land based measures) are as small as the effects of the A and B scenarios in relation to the baseline (as discussed in the previous section). AEA 18

25 6 Benefits assessment The objective of this stage of the study is to quantify the health benefits of the scenarios described above in terms of reduced impact, and then to monetise these benefits in order that they may be compared directly with costs. The focus here is on health impacts because these dominate the quantified (and monetised) benefits assessment. In the event that benefits do not exceed costs it would be appropriate to expand the analysis to bring in other effects such as those on building materials and crops. It would also be appropriate to consider in a qualitative way how un-monetisable impacts, such as those on ecosystems and cultural heritage, would be likely to influence the balance of costs and benefits. 6.1 Background The benefits assessment model used in this study is ALPHA2 the Atmospheric Long-range Pollution Health and environment Assessment model, version 2 (Holland et al, 2008). The original version of ALPHA was developed at AEA Technology in the 1990s, drawing extensively on the ExternE research programme, and was used to inform development of the EC s Acidification Strategy, the Ozone Directive, the National Emission Ceilings Directive and the Gothenburg Protocol to the UN/ECE Convention on Long Range Transboundary Air Pollution. The input data for the ALPHA2 programme are the outcomes of the dispersion modelling reported in Chapter Methodology The CAFE CBA methodology was described in three volumes (Holland et al, 2005a, 2005b; Hurley et al, 2005) 8 and can be traced back to the beginning of the EC DG Research ExternE Programme that started in 1991 and has continued more or less continually since. Further to this, the methodology used here was the subject of intense consultation in 2003 and 2004 with stakeholders from the European Union Member States, academic institutes, environment agencies, industry and non-governmental organisations. It was also subject to formal peer review by senior experts in the USA and Europe (the peer review report is available at the above website). Further work to refine the model is being carried out under the LIFE+ Programme s EC4MACS Project 9 and the model as used here includes these developments. The quantification of impacts and their subsequent monetisation follows the impact pathway approach as shown in Figure 6.1. It is in essence a logical progression from emission, through dispersion and exposure to quantification of impacts and their valuation AEA 19

26 Figure 6.1. Impact Pathway Approach, illustrated with the example of the effects of emissions of NO x and VOCs on ozone concentrations and crop yield Impacts are generally calculated as: Impact = Pollution x Stock at risk x Response function Pollution may be expressed in terms of concentration or deposition. The term stock at risk relates to the amount of sensitive material (people, ecosystems, materials, etc.) present in the modelled domain. For the health impact assessment account is taken of the distribution of population and of effects on demographics within the population, such as children, the elderly, or those of working age. Incidence rates considered representative of the rate of occurrence of different health conditions across Europe are used to modify the stock at risk for each type of impact quantified. These incidence rates ensure that impacts are referenced to actual health data and hence should not be of an unrealistic magnitude. The health impacts quantified are listed in Table 6.1 below. For the CAFE-CBA two sets of response functions were identified, those for which evidence was considered most robust which were grouped as the core set and a group for which quantification was considered less robust, referred to as the sensitivity set. In this study we start by applying only the core set of functions, reserving the sensitivity set for the situation in which it could be found that core benefits do not exceed costs. Valuation is performed as follows: Economic damage = Impact x Unit value of impact Unit values seek to describe the full economic effect of the impacts that they are linked with. For health impacts, for example, which dominate the analysis, this will include elements associated with the costs of health care, lost productivity amongst workers and aversion to premature death or ill health. A more complete account of the methods used for health impact quantification is provided in Volume 2 of the CAFE methodology (Hurley et al, 2005), though the valuation data used here have been revised for the current study drawing on information from the EC4MACS work (Table 6.1). The methodological issue that has received most debate over the years is the approach that should be taken to valuation of mortality, whether it should be performed relative to the loss of life expectancy using the value of a lie year (VOLY) or the number of deaths that can be linked to pollution exposure using the value of AEA 20

27 statistical life (VSL). Whilst the study team has a strong preference for use of the VOLY, we have used both VOLY and VSL estimates to demonstrate sensitivity on this issue. There are two main differences between the values shown and those used in the CAFE-CBA work. First, values have been increased in line with prices for the period 2000 to 2005, reflecting the different price years used in this work compared to CAFE. Second, the value of a life year (VOLY) has been reduced from 52,000 to 40,000 reflecting the results of the recent NEEDS study for EC DG Research. The result of the NEEDS work may be preferred for several reasons, including: Use of a larger total sample Consideration of a larger number of countries (9) compared to the 3 used for the earlier estimate Inclusion of countries from the new member states (Poland, Czech Republic and Hungary) as well as from the original EU15 A further factor is that the mean estimate from the new work was found to be reasonably stable (not influenced by a small number of very large individual estimates of willingness to pay) and so could be adopted as the best estimate of VOLY. In the earlier NewExt study (the basis of the previous estimate of 52,000 per VOLY), in contrast, the mean was not stable, and so two estimates of the VOLY, representing the mean and the median of the study population, were adopted. Use of a single estimate clearly simplifies the analysis. A problem in basing the VOLY purely on the results of the NEEDS project is that it pays no attention at all to the results of the NewExt work. A final decision on the VOLY to be recommended under EC4MACS has yet to be made, though it is currently likely that it will seek to account for both the NewExt and NEEDS results. Basing the present analysis only on the NEEDS result leads to a lower estimate of benefits, making it less likely that benefits would exceed costs than if a higher VOLY estimate were applied. For the purposes of the current analysis we consider this conservative approach to be appropriate, partly as it makes very little difference to the comparison of costs and benefits in Chapter 7: a prolonged discussion here of whether or not a higher value should be adopted for the VOLY would therefore add nothing to the present study. It is not recommended that the 40k estimate is adopted more widely without further discussion. The CAFE-CBA also adopted two estimates of the VSL, again reflecting the median and the mean of the study population. However, on reflection, this was considered unnecessary: whilst the NewExt mean for VSL was considered unstable, reference to the wider European literature on characterisation of the VSL suggests that a best estimate in the region of 1 million is appropriate. Table 6.1. List of health impacts quantified with unit values applied in this analysis. Impact / population group Pollutant Unit value ( ) Acute Mortality (All ages) O 3 40,000 Respiratory Hospital Admissions (65yr +) O 3 2,364 Minor Restricted Activity Days (MRADs 15-64yr) O 3 42 Respiratory medication use (adults 20yr +) O 3 1 Chronic Mortality (30yr +) - as life years lost PM ,000 Chronic Mortality (30yr +) as premature deaths PM 2.5 1,130,000 Infant Mortality (0-1yr) PM 2.5 1,695,000 Chronic Bronchitis (27yr +) PM ,000 Respiratory Hospital Admissions (All ages) PM 2.5 2,364 Cardiac Hospital Admissions (All ages) PM 2.5 2,364 Restricted Activity Days (RADs 15-64yr) PM Respiratory medication use (children 5-14yr) PM Respiratory medication use (adults 20yr +) PM Lower Respiratory Symptom days (children 5-14yr) PM Lower Respiratory Symptom among adults (15yr +) with chronic symptoms PM Uncertainties are of course present in all inputs to the analysis. Particularly important are those relating to the response functions used and the valuation data. These are accounted for later in the analysis AEA 21

28 when benefits of the scenarios are compared against the costs: uncertainties are only important to the extent that they would change the conclusions drawn from the comparison of costs and benefits. 6.3 Results Scenarios are defined in Table 6.2 drawing on information in preceding chapters. Impacts in terms of the change in incidence of ill health, loss of longevity, etc. are shown Table 6.3 for 2015 and in Table 6.4 for Estimates of monetised morbidity and mortality impacts for the EU27 are presented in Tables 6.5 and 6.6. Total health impacts are presented as a range, using the four different estimates of total annual damage. The health impacts of air pollution are dominated by PM mortality, although PM related morbidity is also significant. The most important effects (in monetised terms) for PM related morbidity are restricted activity days and cases of chronic bronchitis. Table 6.2. Scenario-run codes Run Name Description Year BASE_2015 Base 2015 SEA1_2015 Sea 1 Scenario 2015 SEA2_2015 Sea 2 Scenario 2015 SEA3_2015 Sea 3 Scenario 2015 ABASE1_2015 Land A Scenarios Base 2015 BBASE1_2015 Land B Scenarios Base 2015 AS1_2015 Land A Scenario and Sea BS1_2015 Land B Scenario and Sea BASE_2020 Base 2020 SEA1_2020 Sea 1 Scenario 2020 SEA3_2020 Sea 3 Scenario 2020 SEA4_2020 Sea 4 Scenario 2020 SEA5_2020 Sea 5 Scenario 2020 AS1_2020 Land A Scenario and Sea BS1_2020 Land B Scenario and Sea AS4_2020 Land A Scenario and Sea BS4_2020 Land B Scenario and Sea AEA 22

29 Table 6.3. Estimated annual health impacts in 2015 due to air pollution in the EU27 End point Population at risk Impact Pollutant base_2015 Sea1_2015 Sea2_2015 Sea3_2015 abase1_2015 as1_2015 bbase1_2015 bs1_2015 Acute Mortality (All ages) All Premature deaths O 3 22,368 22,047 22,153 22,153 22,366 22,051 22,363 22,054 Respiratory Hospital Admissions (65yr +) Elderly Cases O 3 19,078 18,801 18,891 18,891 19,075 18,803 19,073 18,806 Minor Restricted Activity Days (MRADs 15-64yr) Adults Days O 3 48,756,183 48,054,241 48,283,256 48,283,081 48,749,473 48,061,760 48,742,322 48,069,976 Respiratory medication use (adults 20yr +) Adults Days O 3 8,920,808 8,791,789 8,834,138 8,834,103 8,919,611 8,793,172 8,918,283 8,794,675 Chronic Mortality (All ages) All Life years lost PM 2.5 2,153,682 2,032,021 1,979,544 1,977,795 2,147,341 2,025,435 2,141,099 2,018,649 Chronic Mortality (30yr +) Infants Premature deaths PM , , , , , , , ,491 Infant Mortality (0-1yr) Adults Premature deaths PM Chronic Bronchitis (27yr +) All Cases PM ,552 99,769 97,096 97, ,244 99, ,941 99,122 Respiratory Hospital Admissions (All ages) All Cases PM ,800 33,777 32,905 32,876 35,694 33,668 35,590 33,555 Cardiac Hospital Admissions (All ages) Adults Cases PM ,079 20,832 20,294 20,276 22,014 20,764 21,950 20,695 Restricted Activity Days (RADs 15-64yr) Children Days PM ,942, ,937, ,227, ,058, ,373, ,346, ,814, ,738,495 Respiratory medication use (children 5-14yr) Adults Days PM 2.5 1,813,697 1,702,016 1,660,289 1,659,437 1,808,029 1,696,144 1,802,545 1,690,067 Respiratory medication use (adults 20yr +) Children Days PM ,371,218 16,406,775 15,976,148 15,961,648 17,320,399 16,353,956 17,270,283 16,299,718 LRS symptom days (children 5-14yr) Adults Days PM ,377,920 76,544,732 74,661,618 74,603,414 81,132,661 76,290,976 80,893,458 76,026,855 LRS among adults (15yr +) with chronic symptoms All Days PM ,956, ,208, ,928, ,785, ,446, ,678, ,943, ,132,904 Note: Results are rounded to the nearest 100. AEA 23

30 Table 6.4. Estimated annual health impacts in 2020 due to air pollution in the EU27 End point Acute Mortality (All ages) Population at risk All Impact Pollutant base_2020 Sea1_2020 Sea3_2020 Sea4_2020 Sea5_2020 as1_2020 as4_2020 bs1_2020 bs4_2020 Premature deaths O 3 22,244 21,417 21,539 21,567 21,598 21,418 21,569 21,420 21,571 Respiratory Hospital Admissions (65yr +) Elderly Cases O 3 21,059 20,272 20,387 20,413 20,442 20,273 20,415 20,275 20,416 Minor Restricted Activity Days (MRADs 15-64yr) Adults Days O 3 45,487,097 43,798,565 44,045,892 44,109,754 44,173,452 43,800,727 44,112,882 43,803,714 44,116,542 Respiratory medication use (adults 20yr +) Adults Days O 3 8,722,895 8,398,086 8,445,792 8,457,617 8,469,909 8,398,511 8,458,240 8,399,083 8,458,951 Chronic Mortality (All ages) All Life years lost PM 2.5 2,006,114 1,844,775 1,784,037 1,766,272 1,756,678 1,839,089 1,760,318 1,835,106 1,756,075 Chronic Mortality (30yr +) Infant Mortality (0-1yr) Infants Adults Premature deaths Premature deaths PM , , , , , , , , ,697 PM Chronic Bronchitis (27yr +) All Cases PM ,238 93,329 90,158 89,302 88,801 93,046 89,006 92,848 88,795 Respiratory Hospital Admissions (All ages) All Cases PM ,347 30,665 29,655 29,360 29,200 30,570 29,261 30,504 29,190 Cardiac Hospital Admissions (All ages) Adults Cases PM ,566 18,912 18,289 18,107 18,009 18,854 18,046 18,813 18,003 Restricted Activity Days (RADs 15-64yr) Children Days PM ,608, ,476, ,154, ,558, ,718, ,982, ,042, ,636, ,673,511 Respiratory medication use (children 5-14yr) Adults Days PM 2.5 1,558,598 1,422,846 1,379,585 1,362,086 1,355,244 1,418,167 1,357,160 1,414,908 1,353,645 Respiratory medication use (adults 20yr +) Children Days PM ,466,061 15,161,445 14,654,192 14,512,750 14,432,597 15,115,158 14,464,269 15,082,675 14,429,751 LRS symptom days (children 5-14yr) Adults Days PM ,418,108 64,527,840 62,543,893 61,835,829 61,523,051 64,323,547 61,621,863 64,181,827 61,468,607 LRS among adults (15yr +) with chronic symptoms All Days PM ,596, ,535, ,515, ,094, ,301, ,073, ,610, ,749, ,266,489 Note: Results are rounded to the nearest 100. AEA 24

31 Table 6.5. Monetised annual health impacts under each scenario in 2015 in the EU27 ( billion) End Point Poll BASE_2015 SEA1_2015 SEA2_2015 SEA 3_2015 AS1_2015 BS1_2015 Acute Mortality (All ages) O Respiratory Hospital Admissions (65yr +) O Minor Restricted Activity Days (MRADs 15-64yr) O Respiratory medication use (adults 20yr +) O Total O3 Mortality Chronic Mortality (30 yr+) VOLY valuation PM Chronic Mortality (30yr +) VSL valuation PM Infant Mortality (0-1yr) PM Chronic Bronchitis (27yr +) PM Respiratory Hospital Admissions (All ages) PM Cardiac Hospital Admissions (All ages) PM Restricted Activity Days (RADs 15-64yr) PM Respiratory medication use (children 5-14yr) PM Respiratory medication use (adults 20yr +) PM LRS symptom days (children 5-14yr) PM LRS among adults (15yr +) with chronic symptoms PM Total PM damage (low, VOLY valuation) Total PM damage high, VSL valuation) Total PM+O3 damage high, VSL valuation) Total PM+O3 damage (low, VOLY valuation) AEA 25

32 Table Monetised annual health impacts under each scenario in 2020 in the EU27 ( billion) End Point Poll base_2020 SEA1_2020 SEA 3_2020 SEA 4_2020 SEA 5_2020 AS1_2020 AS4_2020 BS1_2020 BS4_2020 Acute Mortality (All ages) O Respiratory Hospital Admissions (65yr +) O Minor Restricted Activity Days (MRADs 15-64yr) Respiratory medication use (adults 20yr +) O O Total O3 Mortality Chronic Mortality (30 yr+) VOLY valuation PM Chronic Mortality (30yr +) VSL valuation PM Infant Mortality (0-1yr) PM Chronic Bronchitis (27yr +) PM Respiratory Hospital Admissions (All ages) PM Cardiac Hospital Admissions (All ages) PM Restricted Activity Days (RADs 15-64yr) PM Respiratory medication use (children 5-14yr) Respiratory medication use (adults 20yr +) PM PM LRS symptom days (children 5-14yr) PM LRS among adults (15yr +) with chronic symptoms PM Total PM damage (low, VOLY valuation) Total PM damage high, VSL valuation) Total PM+O3 damage high, VSL valuation) Total PM+O3 damage (low, VOLY valuation) AEA 26

33 Table 6.7. Monetised annual health benefits compared to the relevant baseline under each scenario in 2015 in the EU27 ( billion) End point Pollutant SEA1_2015 SEA2_2015 SEA3_2015 AS1_2015 BS1_2015 Acute Mortality (All ages) O Respiratory Hospital Admissions (65yr +) O Minor Restricted Activity Days (MRADs 15-64yr) O Respiratory medication use (adults 20yr +) O Total ozone benefit Chronic Mortality (30 yr+) VOLY valuation* PM Chronic Mortality (30yr +) VSL valuation* PM Infant Mortality (0-1yr) PM Chronic Bronchitis (27yr +) PM Respiratory Hospital Admissions (All ages) PM Cardiac Hospital Admissions (All ages) PM Restricted Activity Days (RADs 15-64yr) PM Respiratory medication use (children 5-14yr) PM Respiratory medication use (adults 20yr +) PM LRS symptom days (children 5-14yr) PM LRS among adults (15yr +) with chronic symptoms PM Total PM benefit (low, VOLY valuation) Total PM benefit (high, VSL valuation) Total PM+O3 benefit (low, VOLY valuation) Total PM+O3 benefit (high, VSL valuation) AEA 27

34 Table 6.8. Monetised annual health benefits compared to the relevant baseline under each scenario in 2020 in the EU27 ( billion) End point Pollutant Sea1_2020 Sea3_2020 Sea4_2020 Sea5_2020 as1_2020 as4_2020 bs1_2020 bs4_2020 Acute Mortality (All ages) O Respiratory Hospital Admissions (65yr +) O Minor Restricted Activity Days (MRADs 15-64yr) O Respiratory medication use (adults 20yr +) O Total ozone benefit Chronic Mortality (30 yr+) VOLY valuation* PM Chronic Mortality (30yr +) VSL valuation* PM Infant Mortality (0-1yr) PM Chronic Bronchitis (27yr +) PM Respiratory Hospital Admissions (All ages) PM Cardiac Hospital Admissions (All ages) PM Restricted Activity Days (RADs 15-64yr) PM Respiratory medication use (children 5-14yr) PM Respiratory medication use (adults 20yr +) PM LRS symptom days (children 5-14yr) PM LRS among adults (15yr +) with chronic symptoms PM Total PM benefit (low, VOLY valuation) Total PM benefit (high, VSL valuation) Total PM+O3 benefit (low, VOLY valuation) Total PM+O3 benefit (high, VSL valuation) * Alternative estimates, not to be combined AEA 28

35 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 In addition to the impacts by scenario, a key consideration is the location of the benefits that arise from each scenario. The following tables demonstrate the health benefits in Euros millions (high, VSL valuation) by Member State against the baseline case in 2015 and These results are roughly twice as large as the equivalent results based on use of the VOLY for mortality valuation. The main purpose of these tables is to demonstrate the way that benefits are distributed across Member States. An overview of the impact by scenario is provided below: SEA 1 The impacts under SEA1 are primarily felt by those Member States that are closely located to the North Sea/English Channel, with the greatest impacts being experienced in the United Kingdom, Germany, France and Netherlands. This is in part driven by the SECA and the high population densities of the Member States. There are also significant health benefits around Baltic Sea, the other SECA region. SEA 2 Under the SEA2 scenario the Mediterranean Sea also becomes a SECA region. This impact is seen immediately with significant health benefits being seen in Italy, Greece and Spain. France also experiences higher health benefits from the new SECA. SEA 3 Under the SEA3 scenario, the Black Sea now becomes a SECA. The result of the scenario increases the health benefits in Bulgaria, Romania, and Greece. However, the impact under SEA 3 is minimal to the change occurring under SEA1 and SEA2 AS1 and BS1 The AS1 scenario includes SEA 1 and the land based scenario (0.05% S). The difference against the SEA1 scenario is a 400m increase, compared to the changes in sea scenario only. This results in slightly higher health benefits in Germany, France, Italy, Netherlands, Belgium, Poland and the United Kingdom. The BS1 scenario includes SEA 1 and the more restrictive land based scenario (0.001% S). The overall impact is almost double (approximately 800m increase) that of the impact of AS1 against SEA1 (approximately 400m increase). The impact occurs primarily in the same countries as identified under AS1. AS4 and BS4 (2020 only) The AS4 scenario includes SEA 4 and the land based scenario (0.05% S). The difference against the SEA1 scenario is again minimal 400m increase, compared to the changes in sea scenario only. This results in slightly higher health benefits in Germany, France, Italy, Netherlands, Belgium, Poland and the United Kingdom. The BS4 scenario includes SEA 4 and the more restrictive land based scenario (0.001% S). The overall impact is almost is an increase of approximately 300m against the AS4 scenario. The impact occurs primarily in the same countries as identified under AS4 AEA 29

36 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Table 6.9 Member State Net Benefits by Scenario ( Million) - (high, VSL valuation) Member State SEA1_2015 SEA2_2015 SEA3_2015 AS1_2015 BS1_2015 Austria Belgium ,007 Bulgaria Cyprus Czech Republic Denmark Estonia Finland France 1,827 2,372 2,370 1,972 2,147 Germany 3,605 3,782 3,786 3,870 4,145 Greece -6 1,133 1, Hungary Ireland Italy 501 3,714 3, Latvia Lithuania Luxembourg Malta Netherlands 1,774 1,800 1,798 1,829 1,881 Poland 1,074 1,222 1,226 1,125 1,186 Portugal Romania Slovakia Slovenia Spain 92 1,956 1, Sweden United Kingdom 4,500 4,530 4,536 4,586 4,702 Grand Total 16,480 24,345 24,624 17,399 18,337 AEA 30

37 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Table 6.10 Member State Net Benefits by Scenario ( Million) 2020, (high, VSL valuation) Member State SEA1_20 20 SEA3_20 20 SEA4_20 20 SEA5_20 20 AS1_20 20 AS4_20 20 BS1_20 20 BS4_20 20 Austria Belgium 1,197 1,220 1,290 1,291 1,247 1,341 1,282 1,378 Bulgaria Cyprus Czech Republic Denmark Estonia Finland France 2,486 3,132 3,395 3,491 2,621 3,543 2,707 3,647 Germany 4,914 5,121 5,468 5,512 5,138 5,701 5,298 5,848 Greece 22 1,394 1,193 1, , ,223 Hungary Ireland Italy 1,035 4,732 4,270 4,843 1,122 4,362 1,195 4,438 Latvia Lithuania Luxembo urg Malta Netherlan ds 2,378 2,408 2,540 2,545 2,427 2,592 2,456 2,623 Poland 1,421 1,594 1,647 1,676 1,472 1,698 1,502 1,739 Portugal Romania Slovakia Slovenia Spain 299 2,480 2,881 3, , ,938 Sweden United Kingdom Grand Total 5,860 5,909 7,032 7,038 5,951 7,110 6,017 7,188 22,456 31,871 34,178 35,662 23,267 35,030 23,837 35,630 Results by Member State are shown graphically by Member State in Appendix 2. AEA 31

38 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Benefits outside of the EU27 In addition to the EU27 Member States the ALPHA model also includes the following 4 additional countries: Croatia Norway Switzerland Turkey The impact of the additional countries differs by scenario. However the additional 4 countries if included in the analysis are able to contribute an additional 1.3billion in health benefits under the SEA3 Scenario 2020, equating to an additional 9% benefit (low, VOLY valuation). The reason for the additional health benefits under SEA3 in particular is the result of the additional SECA regions; Mediterranean and Black Seas. This is combined with the high population densities that exist in Turkey that result in a net benefit of approximately 1.27bn under this scenario. The health impacts in Croatia, Norway and Switzerland are minor and are each comparable to the results demonstrated by Slovakia. There is only a minor change in the results by scenario as a result of having relatively small populations. The table below demonstrates the additional impact for countries outside the European Union. The results in this table can be compared against Table Table 6.11 Member State Net Benefits by Scenario ( Million) 2020, (high, VSL valuation) Member State SEA1_2020 SEA3_2020 SEA4_2020 SEA5_2020 AS1_2020 AS4_2020 BS1_2020 BS4_2020 Croatia Norway Switzerland Turkey 141 2,171 1,882 2, , ,904 Total 358 2,490 2,214 2, , ,262 Total EU27 22,456 31,871 34,178 35,662 23,267 35,030 23,837 35,630 Total EU ,814 34,361 36,392 38,193 23,661 37,270 24,250 37,892 Difference 1.6% 7.8% 6.5% 7.1% 1.7% 6.4% 1.7% 6.3% AEA 32

39 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Summary A summary of the net health benefits, against the baselines are provided below. These figures of course omit benefits arising from effects on other receptors including ecosystems, agriculture and buildings including cultural heritage. Table The range of total health benefits relative to baseline for the EU27, billions Run Name Net Health Benefit Range (2015) Net Health Benefit Range (2020) SEA1 8 to to 23 SEA2 11 to 24 - SEA3 11 to to 32 SEA4-15 to 34 SEA5-16 to 36 AS1 9 to to 23 BS1 9 to to 35 AS4-11 to 24 BS4-16 to 36 Note: Low = VOLY valuation, High = VSL Valuation AEA 33

40 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 7 Cost analysis of the new IMO marine fuel and engine standards and a potential extension of emission control areas This section of the report estimates the cost impact of the new IMO marine fuel and engine standards and potential extension of the Sulphur Emission Control Areas. Furthermore, the costs for stricter control of NO x emissions from shipping are explored. The scenarios developed are aimed at illustrating the impact of varying the stringency of emission controls in the sea regions under consideration. In the cost analysis, the scenarios represent altering either: the SO 2 content of fuels (thereby changing the fuel costs) or introducing SO 2 abatement technology (scrubbers), as well as the introduction of NO X control technologies classified into Tier I, II and III emission standards. 7.1 Cost calculation method description The cost calculations for NO X and SO 2 abatement from shipping in this report have two parts. The first part consists of calculating unit costs for the abatement option under consideration. In the second part scenario specific costs are calculated - based on the penetration of low-no X and low- SO 2 technology options. The cost calculations are presented in low and high cost settings. In the calculations, technologies that may, but are not certain to, reach the emission reduction levels stipulated by the Tier standards are considered. The calculations also allow for scenario comparison between closed loop scrubbers, open scrubbers and low sulphur fuels. The unit costs are calculated per petajoule (PJ), corresponding to the heat of combustion of the fuel used, for each fuel and engine category. The unit costs are based on the costs for installation, maintenance, operation and fuel penalties for using the equipment. These costs take into consideration the relevant technology-specific cost parameters. Furthermore, the technologies and associated costs are split into the NO X emission standard categories Tier I, Tier II and Tier III, as well as between retrofitted and new scrubbers. In the scenario where low sulphur fuel is used instead of scrubbers, the costs for SO 2 emission abatement are calculated as the difference in costs between the high and low fuel qualities. The costs are subject to scenariospecific estimates on hours of use for the technologies. In this report scenario-specific operating hours are of specific relevance for sulphur scrubbing options - because the regions considered as SECA varies between scenarios. For NO X, the Tier III technologies are for all relevant scenarios implemented in the Baltic, Mediterranean, and North Seas. For each scenario, costs are calculated, specifying high and low costs estimates for NO X and SO 2 abatement. For SO 2 abatement, separate cost calculations are presented for scrubber costs and low S-fuel costs. All costs are given in Adjustments of literature values and expert estimate values to inflation and currency exchange rates have been made when necessary. In the central analysis, a 4 % investment interest rate has been used. In obtaining the costs the following economic assumptions have been made: The dynamics of supply / demand on price have not been considered. The actual price of fuels or emission abatement equipment is dependent on the future market price for the equipment. The costs estimates given in this study represent the costs that are found today, and do not adjust for the future market prices. AEA 34

41 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 The economies of scale are unknown and so have not been considered The prices given in this report are all based on fairly small production volumes of abatement equipment. Given that emission standards become more strict by 2016 and 2020, the market volume for abatement technologies will increase, this may lead to lower unit prices due to learning and economies of scale. This effect is not taken into account in this report. No price effects on demand for transport or transport modal shifts have been considered. This report does not consider any modal shifts away from marine shipping as a consequence of increased shipping prices following costs for introducing emission abatement. 7.2 Abatement Technology Description This section briefly describes the technical options for implementing regulations for SO 2 emissions to meet the standards corresponding to 0.1 and 0.5% S in the fuel. Further, different options to reach Tier II and Tier III NO X emissions are described. The costs for the different abatement options are discussed below. Sulphur oxide abatement Low sulphur fuel The emission of sulphur dioxide is directly proportional to the sulphur content in the fuel, the sulphur is oxidised in the combustion chamber and the mass of emitted SO 2 is twice that of the sulphur (S) in the fuel used. Thus, a reduction in the S-content in the fuel can, in principle, meet the lower emissions sought. Most of the high sulphur fuel (with a sulphur content of 1-3.5%) used in ships today is heavy fuel oil or residual fuel. The fuel currently available with 0.1% S is typically marine gasoil; it is likely that the fuel with this sulphur content, which will be required in SECA areas after 2015, will be similar to it. Fuel with a sulphur content of 0.5% can be residual fuel that has been de-sulphurised or marine diesel formulated from a mixture of residual oil and distillates. Purvin & Gertz (2009) have described a number of scenarios in which the quality of the 0.5% S fuel varies; the lowest priced fuel in that study had a high content of heavy oils. Scrubbers An alternative low sulphur fuel is to use an abatement technology which reduces the SO 2 emissions by an equivalent amount. While several technologies might be discussed well established data is only available for wet scrubbing a technique widely used for SO2 reduction from land based combustion plants. The principle of operation is that the exhaust gas is brought into contact with water with a buffered alkalinity such that SO 2 is trapped and converted to sulphate ions. Scrubbers for marine applications can be divided into open (seawater) scrubbers and closed (freshwater) scrubbers. The former utilises the natural alkalinity of sea water for the scrubbing and neutralisation process, the residual SO 2 emission following scrubbing depend on the sulphur content of the fuel, the ph of the seawater and the flows of exhaust and scrubber water. These systems have been shown to achieve a high efficiency of SO 2 emissions reduction. There are certain wash-water criteria which likely can be met in the future (and the sea water may require treatment before being returned to sea). The systems may also be custom designed for specified geographical areas and fuel qualities, i.e., areas with low alkalinity could be excluded and there could be specified maximum sulphur content in the fuel used. The closed scrubbers use freshwater with an added neutralising agent (normally caustic soda); these systems have been demonstrated to meet the emission criteria and the wash-water criteria. In the future it is anticipated that combined systems will be developed where seawater is used in open waters with a strong buffering capacity together with a closed loop operational mode for use in ports, estuaries and seas with brackish water. AEA 35

42 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Nitrogen oxides Marine engine manufacturers will be able to meet the Tier II specifications from 2011 for new engines by improving the control of fuel injection and other engine related modifications. There are several, different, options for meeting the Tier III specifications - from The main approaches are based either on selective catalytic reduction (SCR) or exhaust gas recirculation (EGR) in combination with other measures. SCR has already been demonstrated to reach Tier III emission levels on many ships while the EGR alternatives are still in the development phase. SCR In SCR a reducing agent, normally urea, is mixed with the exhaust gases in the presence of a catalyst. For marine applications the catalyst is usually composed of vanadium and titanium oxides. Under these conditions the urea is decomposed to ammonia which then selectivity reacts with nitrogen oxides to produce nitrogen gas. The SCR systems require a minimum exhaust gas temperature to operate effectively, this is normally around 300 C but the temperature depends on the sulphur content of the fuel. Consequently SCR is less effective at low loads and for twostroke engines; in these cases it may be necessary to put the SCR between the cylinders and the turbo in order to achieve the required temperature. The SCR operational costs include: urea purchase and catalyst replacement; in practice catalysts need to be replaced after a few years due to deactivation - likely to have been caused by: the formation of sulphur compounds, impurities in the urea, or by thermal deactivation. On the other hand the reduction of nitrogen oxides in the SCR is so effective that the engine may be tuned for low fuel consumption with increased engine NO X emissions thus saving some fuel. The SCR system must, however, be designed to avoid slip of ammonia with the exhaust - this can be achieved by means of an oxidation catalyst after the SCR. EGR/engine combinations Alternatives to SCR are being developed to reach Tier III. While engines can be modified through improved fuel injection systems and improved turbo systems these alone will not be enough to reach Tier III and so additional methods are required. The main track seems to be to use exhaust gas recirculation (EGR) in combination with other techniques. EGR means that some of the exhaust is re-circulated into the engine leading to lower cylinder temperatures and reduce formation/emission of NO X. The potential for NO X reduction is likely to be around 50%. The recirculated gas must not contain excessively high concentration of SO 2 or soot. Therefore a low sulphur fuel (around 0.1%) would be needed or alternatively exhaust cleaning integrated with the circulation system - this could be either a scrubber or a particulate filter (if low sulphur fuel is used); EGR may also impose a fuel consumption penalty. In order to reach Tier III EGR may have to be combined with some other technique - in all likelihood direct water injection (DWI), or use of a fuel-water emulsion. These methods may require freshwater - implying a further fuel penalty, there may also be a practical limit on the maximum sulphur content in the fuel. Tier I-III In the cost calculations two alternative means are used to reach Tier III - SCR and EGR in combination with engine modifications and DWI (similar costs are expected with water emulsion instead of DWI). Tier II is assumed to be reached by engine modifications. Retrofit costs for Tier0 engines produced between 1990 and 2000 to reach Tier I standards have also been estimated. Compatibility As mentioned above there are problems with using high sulphur fuel in combination with NO X abatement technologies. EGR requires very low sulphur content in the fuel or an internal scrubber. SCR performance can fall away due to catalyst poisoning with high sulphur fuel and the techniques using water injection or emulsion may also impose limits on the fuel sulphur content. Consequently there are unresolved design issues with combining NO X reduction techniques with scrubbers: if scrubbers are combined with SCR the exhaust gas may need to be reheated after the scrubber to reach the operating temperature of the SCR, if the SCR is put in front of the scrubber it will be exposed to a high sulphur gas. AEA 36

43 Cost benefit analysis to support the impact assessment AEA/ED /Issue Unit abatement costs The abatement costs are presented as costs per petajoule of the heat value for the fuel used (PJ); the costs are also presented as costs per tonne emission reduced - to facilitate comparison with other cost estimates. The costs for the different control standards are dependent on which technology can reach these standards. Abatement costs have been calculated for NO X and SO 2 abatement using previous literature values (Entec 2005, UK Maritime and Coastguard Agency 2009, US-EPA 2009) and information resulting from direct communications with main engine and equipment manufacturers,. Sulphur oxides In order to calculate costs for reducing sulphur emissions the following options have been considered: Table 7.1: SO 2 abatement options SO 2 Comment New Scrubber In combination with high sulphur fuel (2.94 % ) Retrofit Scrubber In combination with high sulphur fuel (2.94 % ) Low sulphur HF, 0.5 % - Low sulphur MD, 0.1 % - The scenarios developed in this study separate costs for scrubbers from costs of low sulphur fuels. The following text presents the costs for fuel shifts and scrubbers respectively. Fuels The costs of fuel shifts depend mainly on the projected future fuel prices for different sulphur categories. The main parameters influencing the costs of fuel relate to sulphur content of crude oils as well as the necessary investments in refinery capacities. The vessels using the fuels are assumed to be subject to the relatively small costs of adapting to the different fuels. The fuel costs used in this report are based on the fuel price projections in Purvin & Gertz (2009). The fuel shift costs, specified as a cost per PJ of heat value in the fuel is given in the table below: AEA 37

44 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Table 7-2: Fuel costs in 2020, adapted from Purvin & Gertz, 2009 and corresponding SO 2 associated abatement cost. Low High Low SO High cost 2 Low SO 2 Option cost Low cost cost abatement abatement [$ / (%S in fuel) [$ / [k tonne] 2005 / PJ] [k 2005 / cost [ / cost [ / tonne] PJ] ton SO 2 ] ton SO 2 ] Fuel shift ( ) Fuel shift (2.94 1) Fuel shift ( ) Fuel shift ( ) Fuel shift ( ) The fuel cost for heavy fuel with 2.94 % sulphur content is projected to be $ 420 / tonne in Purvin & Gertz The fuel costs per tonne are converted to costs per PJ by considering the specific heat value of the fuel and the conversion rate between euro and dollar. Scrubbers The most important parameters deciding the costs for scrubbers are whether the technology would be installed in a new vessel or retrofitted to an existing vessel, and whether an open or closed circuit scrubber system is chosen. The fuel penalty when using a scrubber is considered to be 2% (Krystallion), which is used in the cost calculations. In closed loop systems, costs for maintenance, purchase of NaOH and fresh water are also taken into consideration. For open systems, only maintenance is considered relevant for the cost calculations in addition to the fuel penalty. The cost for disposal of sludge is included in the cost estimates but is relatively small and could decrease even further following construction of new sludge disposal facilities and higher rates of usage of these facilities. For closed scrubber systems, the amount of fresh water and NaOH used depend on the sulphur content of the fuel, and the costs will vary accordingly. The market price for NaOH has varied over the last couple of years. In our central cost estimate, a price for 2005 of 0.5 / litre 50% NaOH is used. This price estimate is high but engine manufacturers estimate NaOH prices will range between / litre, so the price is on the lower boundary of other estimates. In 2007 the price for different grades of NaOH was / litre. The price for urea is also subject to variation, but the engine manufacturer experts are closer in their estimates on urea prices with a price range of / litre. In our central estimate we use an urea price of 0.2 / litre. Cost parameters for the unit costs used in this report are given in Table 7.3. Table 7.3: Central cost parameters Tier operating hours at sea 4000 h/year Scrubber operating hours at sea 4000 h/year Urea price / litre NaOH price / litre AEA 38

45 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Costs for different scrubber solutions are presented in Table 7.6. The costs in the Table are given for medium sized SSD-RO engines running on residual oil fuel with 2.94 % sulphur content as specified above. However, in the scenario calculations, costs for other engine categories are also included. Low and high cost options are presented. These two options represent the costs associated with either using fresh water in combination with NaOH, or sea water in the scrubber. The investment costs (2005) are in the range / kw for new installations and / kw for retrofit installations, which is in line with manufacturer estimates (Wärtsilä 2009, Marine and Energy Consulting 2009). The costs for scrubber operation vary with the required SO 2 content in the exhaust gases - and the different amounts of NaOH that will be needed. The costs are therefore presented as costs for installations corresponding to certain sulphur content in the fuel that would give the same SO 2 emission. For example, Scrubber-0.1 corresponds to emissions from low sulphur fuels with 0.1 % sulphur content. The NaOH consumption needed to reduce sulphur emissions from high sulphur fuel (2.94 %) to a level corresponding to 0.1 % S fuel is estimated as 15 litres per MWh installed engine capacity. If trying to reach a 0.5 % S fuel emission level, it is assumed that 12 litres / MWh are needed. For a 1.5 % S fuel emission level, 6 litres / MWh is assumed. The lifetime of the equipment is estimated to be the same as in the ENTEC reports (ENTEC 2005). Scrubber sludge disposal costs ENTEC(2005) looked at the production of sludge when using a scrubber and the permitted amount of sludge for disposal according to the standard tariff used at the port of Rotterdam. ENTEC (2005) indicated that the amount of sludge produced would be within the current allocations/ tariffs of a port like Rotterdam. Meech (2009) indicates that the costs associated with sludge disposal could be small and Wärtsilä (2009) do not take the costs of sludge disposal specifically into consideration. Nevertheless, using the Gothenburg harbour port tariffs as a reference, based on the Gross Tonnage (GT) of the vessel of maximum 0.4 / GT, and assuming that vessels spend 500 hours between sludge discharge in ports, we have calculated that a maximum cost of sludge disposal could be ~ euro for a small vessel. Table 7.4 Sludge disposal costs for Port of Göteborg Scrubber sulphur sludge disposal cost estimates - based on the Gothenburg port charges in 2009 Vessel GT SEK2009 / 2005 / Number of Annual cost for sludge GT GT calls per year disposal Small 3, ,080 Medium 15, ,398 Large 50, ,326 These costs are much exaggerated since they allocate all sludge disposal costs to the scrubber sludge, which would constitute only one part of total amount of sludge produced at the vessel. Another price estimate for scrubber sludge disposal have been given by CE Delft (2006), who used the contemporary fees at the port of Oslo (which were 0.12 / litre at the time), to calculate the costs for sludge disposal. The study indicated that a large (40 MW installed power) container ship would produce 0.8 tonne of scrubber sludge during 30 operating hours. This corresponds to an hourly cost of 2.9. A medium vessel (15 MW) would produce some 0.4 tonne over 40 operating hours, corresponding to a cost of ~1 / hour. No estimates are given for small vessels. Wärtsilä (2009) specifies that some scrubbers produce 1.3 litres sludge per MWh. By using these numbers, the following sludge disposal costs are calculated for scrubbers, AEA 39

46 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Table 7.5. Sludge disposal costs. Vessel type Sludge production [litres / MWh] Sludge disposal cost / litre [ ] Annual disposal costs in the central analysis [ ] Small ,600 Medium ,300 Large ,300 These costs correspond to % of the total scrubber costs for a new scrubber ( % for retrofit). The costs for sludge disposal are included in the O&M column in the table below. The table presents the scrubber costs associated with a medium sized vessel reaching a sulphur content of the flue gases corresponding to 1.5, 0.5, or 0.1 % S content in fuel. Table 7.6: Scrubber costs used in this study SCRUBBER COSTS New New Retrofit Retrofit New New Retrofit Retrofit New New Retrofit Retrofit Tech Scrubber- 1.5 Scrubber- 1.5 Scrubber- 1.5 Scrubber- 1.5 Scrubber- 0.5 Scrubber- 0.5 Scrubber- 0.5 Scrubber- 0.5 Scrubber- 0.1 Scrubber- 0.1 Scrubber- 0.1 Scrubber- 0.1 Tech spec Investment [k /vessel] Lifetime [year] O&M [k /vessel] Fuel cost [k /vessel] Annual cost [k ] Costs / tonne SO 2 [ ] open scrubber 1, ,035 Closed scrubber - high cost 2, ,680 open scrubber 2, ,840 Closed scrubber - high 4, ,290 open scrubber 1, Closed scrubber - high 2, ,875 open scrubber 2, ,045 Closed scrubber - high 4, ,785 open scrubber 1, Closed scrubber - high 2, ,730 open scrubber 2, Closed scrubber - high 4, ,505 Following the simulated impact of retrofit costs and the impact of NaOH use, the costs for scrubbers vary with a factor of 5 between the least expensive open scrubber option and the most expensive retrofit close circuit scrubber option. This large variation is a consequence of the relatively expensive NaOH costs used in these cost estimates. AEA 40

47 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 NO X The cost for reduction of NO X emissions in the scenarios studied is only applicable as a replacement of Tier II technologies with Tier III technologies. There is currently much research ongoing by engine manufacturers to reach Tier III standards by methods other than SCR. It is likely that these other methods will be associated with comparatively lower costs but it is still uncertain whether they will able to reach Tier III standards. SCR When calculating the costs for SCR, the cost assumed for urea is important. Its price, like that of fuel, is heavily dependent on future supply and demand; recent fluctuations show that the future price is uncertain. One potential advantage of SCR is that the emission reduction efficiency is potentially so large that the engine can be modified for optimal fuel performance - the NO X emissions increase but remain within the capacity of the SCR system. However, this is not considered in this study since there are different findings regarding this issue from different engine manufacturer experts. EGR (Exhaust Gas Recirculation) combinations Lower cost options to SCR feature a combination of engine alterations (such as turbo charging) and EGR and water injection. The latter come in several different forms - Humid Air Motors, HAM; Direct Water Injection DWI; and Water Emulsion WE. The fuel efficiencies of the individual measures act in opposing directions, EGR reduces fuel efficiency while turbo charging may improve it. The net effect of EGR combinations in our study is assumed to imply a 2% fuel penalty. The Tier I and Tier II costs are based on the estimates from ENTEC (2005). In addition, communication with manufacturers has indicated that these costs may be even lower and will only require marginal additions for most engines. The following list presents the emission standards and main technologies considered in the cost calculations for NO X abatement. Table 7.7: NO X control standards considered in this study NOx % Removal Comment Tier I retrofit 5 % NO X For ships constructed before the year 2010 Tier II new 20 % NO X Built after 2011 Tier III new 80 % NO X Built after 2016 for NO X - ECAs For a typical vessel with a medium sized SSD-RO engine using residual oil with a 2.94 % sulphur content, using the main and auxiliary engines at some 4500 hours per year (4000 hours at sea), the following unit costs are obtained in the cost calculations. AEA 41

48 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Table 7.8: Tier cost estimates used in this study Inv Tier [k / Costs vessel] Lifetime [year] O&M [k /vsl] Fuel cost [k /vsl] Annual cost [k ] Costs / tonne NO X [ ] vintage Tech Tech spec Retrofit Tier I Basic IEM New New (0.1% S) New (0.5% S) New (1.45%S) New (2.7% S) New Tier II Tier III Tier III Tier III Tier III Tier III Adv. IEM EGR+WIF EGR+WIF EGR+WIF EGR+WIF SCR In the data set used for cost calculations, Tier I, Tier II and Tier III (SCR) are assumed to have the same cost regardless of sulphur content of the fuel. The annual cost for SCR is, however, dependent on the sulphur content of the fuel - because the sulphur in the exhaust gases negatively affects the performance and lifetime of the catalytic elements of the SCR. The additional costs when using high-sulphur fuels is not completely established as of yet. However, in the scenario calculations performed here, SCR will only be used in combination with either scrubbers or low sulphur fuels. The additional costs associated with high sulphur content in the fuel are therefore not invoked in our calculations. For EGR + WIF solutions, the sulphur content in the fuel will be of most concern because the exhaust gases re-circulated into the engine will be unaffected by the installation of a main scrubber. In these options, an 'internal scrubber' is used which de-sulphurises the exhaust gas that is re-circulated. The costs for this internal scrubber is dependent on the amount of NaOH used which is in turn dependent on the sulphur content of the fuel. This impact on costs is illustrated in the table above by the four different price estimates for EGR + WIF. The costs for Tier III equipment and consumables are within the range of the cost estimates in ENTEC (2005) and by the USEPA (2009). The USEPA study presents significantly higher costs per ton NO X abated with Tier III technology. The reason for this difference lies in the fact that the USEPA study allocates the total cost of the abatement equipment to the relatively small amounts of NO X that will actually be abated in the US coastal zone that is suggested for NO X -ECA. In the scenario cost calculations used in this study are based on the cost per PJ. Having the cost per PJ for each fuel/vessel/tech combination, scenario specific abatement costs can be calculated based on the fuel and technology choices made for each region and year. 7.4 Scenario abatement costs In step two of the cost calculations, scenario-specific fuel use and abatement technology penetration is used to calculate the total costs for NO X and SO 2 abatement associated with the scenarios. The specifics of the scenarios analysed are as follows: AEA 42

49 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 For SO 2 : the maximum sulphur content of marine fuels is 0.10% (by mass) for SECAs and 2.94 % (by mass) outside SECAs. A reduction of the maximum S-content to 0.50% outside SECAs will also be taken into account for some 2020 scenarios. In the baseline, a sulphur content of 1.45 in SECA and 2.6 outside SECA is considered. For NO X : Marine diesel engines with a power output that exceeds 5,000 kw and a per cylinder displacement at or above 90 litres are upgraded as follows: Tier I standards for 30 % of the ships constructed between 1 January 1990 and 1 January 2000; Tier I standards for ships constructed between 1 January 2000 and 1 January 2011; Tier II standards for ships constructed on or after 1 January 2011; and Tier III standards for ships constructed on or after 1 January 2016 and sailing in the Baltic Sea, North Sea/English Channel and/or the Mediterranean Sea (only applicable to the 2020 scenarios) 7.5 Results The following results are presented per scenario in aggregated and disaggregated form. First the aggregated results are presented in tables , followed by region specific results in tables Given the assumptions of the scenarios, it is assumed that scrubbers are not used to reach the 1.45 % S content emission standards, therefore in the scrubber scenarios costs for scrubbers will be associated with fuel cost savings that represent the difference in costs for 1.5 and 2.94 % S fuel for the Baltic Sea and North Sea. Parameters of major importance for the cost results are as follows: Table 7.9: Central parameters for cost estimates, 2020 Central parameters, 2020 Tier operating hours at sea 4000 h / year Scrubber operating hours at sea 4000 h / year Urea price / litre NaOH price / litre NOTE: 4000 operating hours per year corresponds to a bit less than 50 % of the vessels running time within a SECA or NO X Emission Control Area. The baseline scenario The baseline scenario is described in Section 2 and is compared with the other scenarios (SEA1 - SEA5). Table 7.10: Scenario results - total baseline costs [million 2005] BSL Fuel In the baseline scenario, no NO X removal is assumed. The costs for fuel represent the total fuel costs for all sea regions considered in the report. Readers should note that Table 7.10 is the only cost result table that indicates total annual costs. In the following tables estimates all relate to the costs of the baseline scenario. In order to get the total costs of each scenario, the costs of the baseline should be added to the costs presented in the tables below. AEA 43

50 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Table 7.11: Scenario SEA1 - Total annual abatement costs on top of baseline [million 2005] SEA Alternative 1 Scrubber - scrubber costs low high low high SO2 Fuel cost compared to baseline Sum Alt Alt 2: low S fuel cost NOx Tier I - III Following the introduction of Tier III NO X control in the Baltic, Mediterranean and the North sea, the annual costs for NO X emission removal range between 452 and 645million In the scrubber scenario, the introduction of scrubbers in the Baltic and North Sea allows for the use of high sulphur fuels, which are less expensive than the 1.5 % S fuel used in the baseline. Therefore, the costs for high S fuel given in the table are negative. If reaching the SO 2 emission levels by shifting to low sulphur fuels instead of using scrubbers, the costs are, at a minimum, twice as high. Table 7.12: SEA2 - Total annual abatement costs on top of baseline [million 2005] SEA low high low high Alternative 1 Scrubber - scrubber costs SO 2 Fuel cost compared to baseline n.c. n.c. Sum Alt n.c. n.c. Alt 2: low S fuel cost n.c. n.c. NO X Tier I - III n.c. n.c. Note: n.c. not calculated In SEA2, only covering the year 2015, the Mediterranean Sea is considered as a SECA. The fuel costs increase significantly following the large fuel use in the Mediterranean. In the scrubber scenario, the fuel savings are smaller relative to SEA1. The reason is that the marine diesel used in the Mediterranean will contain 0.1% S rather than 0.5% S (as in the base case) which implies extra costs. AEA 44

51 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Table 7.13: SEA3 - Total annual abatement costs on top of baseline [million 2005] SEA low high low high SO 2 Alternative 1 Scrubber - scrubber costs Fuel cost compared to baseline Sum Alt Alt 2: low S fuel cost NO X Tier I - III In SEA3, both the Mediterranean and the Black Sea are considered as SECAs. In addition to this the Mediterranean is considered as a NO X ECA. The low cost estimate on Tier III (EGR+WIF) is lower than in SEA1 following that the Mediterranean is included in the SECA in SEA3, which reduces the amount of NaOH needed for the internal scrubber. Table 7.14: SEA4 - Total annual abatement costs on top of baseline [million 2005] SEA low high low high Alternative 1 Scrubber - scrubber costs SO 2 Fuel cost compared to baseline n.c. n.c Sum Alt 1 n.c. n.c Alt 2: low S fuel cost n.c. n.c NO X Tier I - III n.c. n.c Note: n.c. not calculated In SEA 4, the permitted sulphur content in the areas outside SECA is reduced to 0.5 %. The Mediterranean and the Black Sea are again outside SECA. The Mediterranean Sea is together with the Baltic and North Sea considered as a NO X Emission Control Area. The low cost estimates for Tier III is lower than in SEA1 but higher than SEA3, following the allowed sulphur content of 0.5 % (0.1% in SEA3). AEA 45

52 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Table 7.15: SEA5 - Total annual abatement costs on top of baseline [million 2005] SEA SO 2 Alternative 1 Scrubber - scrubber costs Fuel cost compared to baseline low high low high 2,432 7,328 n.c. n.c Sum Alt 1 n.c. n.c. 2,172 7,067 Alt 2: low S fuel cost n.c. n.c. 10,962 13,402 NO X Tier I - III n.c. n.c Note: n.c. not calculated In SEA5, all sea regions but the North East Atlantic are considered as SECAs (as in SEA3), and the North East Atlantic has a maximum sulphur content of 0.5 %. This results in the highest sulphur removal costs of all scenarios. The low cost estimates on Tier III are again following the sulphur content of the low sulphur fuel scenario for the Mediterranean. The inclusion of the Mediterranean and the Black Sea is not visible in the low scrubber estimates (sea water scrubber) compared with SEA 4, since the costs for this type of scrubber is assumed to be similar for removal to SO 2 emissions corresponding to either 0.1% or 0.5%. Disaggregated into regions, the costs for the different scenarios are as follows: AEA 46

53 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Table 7.16: SEA1 - Regional annual abatement costs on top of baseline [million 2005] SEA1 - specified per region low high low high Baltic SO 2 Alternative 1 Scrubber - scrubber costs Fuel cost compared to baseline Sum Alt Alt 2: low S fuel cost 901 1, ,180 NO X Tier I - III Black SO 2 Alternative 1 Scrubber - scrubber costs Fuel cost compared to baseline Sum Alt Alt 2: low S fuel cost NO X Tier I - III Med SO 2 Alternative 1 Scrubber - scrubber costs Fuel cost compared to baseline Sum Alt Alt 2: low S fuel cost NO X Tier I - III NE Atl SO 2 North SO 2 ALL REGIONS SO 2 Alternative 1 Scrubber - scrubber costs Fuel cost compared to baseline Sum Alt Alt 2: low S fuel cost NO X Tier I - III Alternative 1 Scrubber - scrubber costs Fuel cost compared to baseline 479 1, , Sum Alt , ,271 Alt 2: low S fuel cost 2,123 2,543 2,302 2,778 NO X Tier I - III Alternative 1 Scrubber - scrubber costs 683 2, ,088 Fuel cost compared to baseline Sum Alt , ,810 Alt 2: low S fuel cost 3,024 3,622 3,279 3,957 NO X Tier I - III AEA 47

54 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Table 7.17: SEA2 - Regional annual abatement costs on top of baseline [million 2005] SEA2 - specified per region low high low high Baltic SO 2 Alternative 1 Scrubber - scrubber costs n.c. n.c. Fuel cost compared to baseline n.c. n.c. Sum Alt n.c. n.c. Alt 2: low S fuel cost 901 1,080 n.c. n.c. NO X Tier I - III 6 6 n.c. n.c. Black SO 2 Alternative 1 Scrubber - scrubber costs 0 0 n.c. n.c. Fuel cost compared to baseline 0 0 n.c. n.c. Sum Alt n.c. n.c. Alt 2: low S fuel cost 0 0 n.c. n.c. NO X Tier I - III 2 2 n.c. n.c. Med SO 2 Alternative 1 Scrubber - scrubber costs Fuel cost compared to baseline 1,155 3,387 n.c. n.c n.c. n.c. Sum Alt 1 1,171 3,403 n.c. n.c. Alt 2: low S fuel cost 5,341 6,356 n.c. n.c. NO X Tier I - III n.c. n.c. NE Atl SO 2 Alternative 1 Scrubber - scrubber costs 0 0 n.c. n.c. Fuel cost compared to baseline 0 0 n.c. n.c. Sum Alt n.c. n.c. Alt 2: low S fuel cost 0 0 n.c. n.c. NO X Tier I - III 2 2 n.c. n.c. North SO 2 Alternative 1 Scrubber - scrubber costs 479 1,406 n.c. n.c. Fuel cost compared to baseline n.c. n.c. Sum Alt ,322 n.c. n.c. Alt 2: low S fuel cost 2,123 2,543 n.c. n.c. NO X Tier I - III n.c. n.c. ALL REGIONS SO 2 Alternative 1 Scrubber - scrubber costs 1,838 5,390 n.c. n.c. Fuel cost compared to baseline n.c. n.c. Sum Alt 1 1,734 5,286 n.c. n.c. Alt 2: low S fuel cost 8,365 9,978 n.c. n.c. NO X Tier I - III n.c. n.c. Note: n.c. not calculated AEA 48

55 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Table 7.18: SEA3 - Regional annual abatement costs on top of baseline [million 2005 ] SEA3 specified per region low high low high Baltic SO Alternative 1 Scrubber - 2 scrubber costs Fuel cost compared to baseline Sum Alt Alt 2: low S fuel cost 901 1, ,180 NO X Tier I - III Black SO Alternative 1 Scrubber - 2 scrubber costs Fuel cost compared to baseline Sum Alt Alt 2: low S fuel cost NO X Tier I - III Med SO Alternative 1 Scrubber - 2 scrubber costs 1,155 3,387 1,157 3,535 Fuel cost compared to baseline Sum Alt 1 1,171 3,403 1,173 3,552 Alt 2: low S fuel cost 5,341 6,356 6,042 7,191 NO X Tier I - III NE Atl SO Alternative 1 Scrubber - 2 scrubber costs Fuel cost compared to baseline Sum Alt Alt 2: low S fuel cost NO X Tier I - III North SO Alternative 1 Scrubber - 2 scrubber costs Fuel cost compared to baseline Sum Alt , ,271 Alt 2: low S fuel cost 2,123 2,543 2,302 2,778 NO X Tier I - III ALL REGIONS SO Alternative 1 Scrubber - 2 scrubber costs 1,898 5,566 1,899 5,807 Fuel cost compared to baseline Sum Alt 1 1,794 5,462 1,638 5,547 Alt 2: low S fuel cost 8,642 10,308 9,634 11,521 NO X Tier I - III AEA 49

56 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Table 7.19: SEA4 - Regional annual abatement costs on top of baseline [million 2005 ] SEA4 - specified per region low high low high Baltic SO Alternative 1 Scrubber - 2 scrubber costs n.c. n.c Fuel cost compared to baseline n.c. n.c Sum Alt 1 n.c. n.c Alt 2: low S fuel cost n.c. n.c ,180 NO X Tier I - III n.c. n.c Black SO Alternative 1 Scrubber - 2 scrubber costs n.c. n.c Fuel cost compared to baseline n.c. n.c. 0 0 Sum Alt 1 n.c. n.c Alt 2: low S fuel cost n.c. n.c NO X Tier I - III n.c. n.c. 2 2 Med SO Alternative 1 Scrubber - 2 scrubber costs n.c. n.c Fuel cost compared to baseline n.c. n.c. 0 0 Sum Alt 1 n.c. n.c. 1,157 3,290 Alt 2: low S fuel cost n.c. n.c. 2,875 4,073 NO X Tier I - III n.c. n.c NE Atl SO Alternative 1 Scrubber - 2 scrubber costs n.c. n.c ,520 Fuel cost compared to baseline n.c. n.c. 0 0 Sum Alt 1 n.c. n.c ,520 Alt 2: low S fuel cost n.c. n.c. 1,328 1,881 NO X Tier I - III n.c. n.c North SO Alternative 1 Scrubber - 2 scrubber costs n.c. n.c ,465 Fuel cost compared to baseline n.c. n.c Sum Alt 1 n.c. n.c ,271 Alt 2: low S fuel cost n.c. n.c. 2,302 2,778 NO X Tier I - III n.c. n.c ALL REGIONS SO Alternative 1 Scrubber - 2 scrubber costs n.c. n.c. 2,432 7,069 Fuel cost compared to baseline n.c. n.c Sum Alt 1 n.c. n.c. 2,154 6,792 Alt 2: low S fuel cost n.c. n.c. 7,631 10,123 NO X Tier I - III n.c. n.c Note: n.c. not calculated AEA 50

57 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Table 7.20: SEA5 - Regional annual abatement costs on top of baseline [million 2005] SEA5 - specified per region low high low high Baltic SO Alternative 1 Scrubber - 2 scrubber costs n.c. n.c Fuel cost compared to baseline n.c. n.c Sum Alt 1 n.c. n.c Alt 2: low S fuel cost n.c. n.c ,180 NO X Tier I - III n.c. n.c Black SO Alternative 1 Scrubber - 2 scrubber costs n.c. n.c Fuel cost compared to baseline n.c. n.c. 0 0 Sum Alt 1 n.c. n.c Alt 2: low S fuel cost n.c. n.c NO X Tier I - III n.c. n.c. 2 2 Med SO Alternative 1 Scrubber - 2 scrubber costs n.c. n.c. 1,157 3,535 Fuel cost compared to baseline n.c. n.c Sum Alt 1 n.c. n.c. 1,173 3,552 Alt 2: low S fuel cost n.c. n.c. 6,042 7,191 NO X Tier I - III n.c. n.c NE Atl SO Alternative 1 Scrubber - 2 scrubber costs n.c. n.c ,520 Fuel cost compared to baseline n.c. n.c. 0 0 Sum Alt 1 n.c. n.c ,520 Alt 2: low S fuel cost n.c. n.c. 1,328 1,881 NO X Tier I - III n.c. n.c North SO Alternative 1 Scrubber - 2 scrubber costs n.c. n.c ,465 Fuel cost compared to baseline n.c. n.c Sum Alt 1 n.c. n.c ,271 Alt 2: low S fuel cost n.c. n.c. 2,302 2,778 NO X Tier I - III n.c. n.c ALL REGIONS SO Alternative 1 Scrubber - 2 scrubber costs n.c. n.c. 2,432 7,328 Fuel cost compared to baseline n.c. n.c Sum Alt 1 n.c. n.c. 2,172 7,067 Alt 2: low S fuel cost n.c. n.c. 10,962 13,402 NO X Tier I - III n.c. n.c Note: n.c. not calculated AEA 51

58 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 8 Comparing Costs and Benefits The comparison of costs and benefits is carried out in two parts. First, a straightforward comparison is made of the results from the previous sections. Results from this preliminary comparison are then used to frame a more detailed comparison that takes more thorough account of the uncertainties that are present. 8.1 Preliminary comparison of costs and benefits Annual costs and benefits (low and high estimates) for the different scenarios are given in the following tables, taking the results directly from the preceding analysis. The low bound for costs is the lower estimate for fitting a scrubber to reduce SO 2 emissions together with the costs of NO x control. The upper bound for costs is for the high estimate for using low sulphur fuels to reduce SO 2 emissions and the costs of NO x control. The range for benefits is based on use of either the VOLY (lower bound) or VSL (upper bound) to value mortality. Table 8.1 Summary of costs and quantified benefits for the 2015 scenarios ( billions) SEA1 SEA2 SEA3 Costs low high Benefits low high Table 8.2 Summary of costs and quantified benefits in the EU27 for the 2020 scenarios ( billions) SEA1 SEA3 SEA4 SEA5 Costs low high Benefits low high In all cases there is no overlap between the estimates shown above: in other words that benefits exceed costs for all of the cases shown. However, the ranges shown to this point are simply indicative they do not account for all of the uncertainties that could affect the analysis, particularly on the side of the benefits estimation. Without more detailed analysis of the uncertainties that are present it is therefore unclear how robust the conclusion that benefits exceed costs actually is (even though the results appear conclusive). To illustrate the problem, the surplus of benefit over cost would grow were other benefits, to ecosystems, crops and the built environment to be added in. However, if certain key aspects of the analysis are overestimated, in particular elements relating to the quantification of benefits of reduced mortality using the lower bound (VOLY) approach, benefits would in turn be overestimated. It is therefore clearly appropriate to consider the uncertainties in the analysis in greater detail. This is done in the next section which applies Monte Carlo analysis to examine more closely the effects of different assumptions on the distribution of both costs and benefits. Table 8.3 summarises the uncertainties accounted for and not accounted for in the ranges shown to this point, and how they can be dealt with in the assessment that follows. AEA 52

59 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Table 8.3 Uncertainties included and excluded from the ranges shown in Table 7.2 Element Uncertainty Accounted for? Costs Response to legislation in terms of option applied Yes Variability in operating costs Yes Economies of scale No Price effect on demand for transport or specific transport modes No Price dynamics relating to supply / demand of fuels or abatement equipment No Benefits Method for valuation of mortality Yes Health incidence rates, response functions, valuation data Exclusion of impacts on ecosystems and cultural heritage (would increase benefits) No No, can be assessed using Monte Carlo 8.2 Detailed uncertainty assessment It is important to be clear about the purpose of the more detailed comparison of costs and benefits. The analysis in the preceding section already suggests that it is very likely that costs will be less than benefits. The uncertainty analysis therefore needs to consider how robust this conclusion is and whether there could be any grounds for coming to a different conclusion. As will be shown, some uncertainties can be addressed quantitatively and in a lot of detail, whilst others need to be addressed more qualitatively. In some situations it is reasonable to estimate the possible spread of data by combining extreme outcomes (e.g. adding or multiplying all high estimates together). However, when this involves the combination of a large number of variables, or when it involves variables which have a large range around them, it can provide only weak information about probable values for two reasons: 1. It does not account for the potential of errors in different parameters to cancel out against each other 2. It does not account for the distribution of values within a range, whether they are all equally likely or whether values somewhere towards the centre of the range are more likely than others. The result generated using these extreme outcomes may therefore be one possible outcome, but it may also be extremely unlikely to happen. On this basis the use of extremes to generate ranges is not helpful for informing policy development unless it demonstrates unequivocally that the range for two parameters of interest (e.g. total cost and total benefit) do not overlap. It has been observed that the ranges given in Tables 8.1 and 8.2 do not overlap. However, for the benefits in particular the ranges shown account only partially for the uncertainties that are present, as they show the effect of using a best estimate of the VOLY or a best estimates of the VSL to value mortality. They do not account for uncertainty in these best estimates or in the other parameters used in the quantification of impacts and monetary benefits. This section factors in the uncertainties in the estimation of benefits relating to the population at risk, the incidence of ill health, concentration-response functions and unit values for each of the health impacts that have been quantified. This generates not only a range around the final estimate of benefit, but also the probability distribution within that range. Extreme values may arise, but they can be shown to be of very low probability compared to values closer to the centre of the distribution. The ranges described for costs in the previous chapter account for all available data accessed during the project. These ranges are adopted for comparison with the newly-computed ranges for AEA 53

60 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 benefits. However, we have only limited information to tell us how values are distributed within the cost ranges. It is therefore necessary to make some assumptions on the shape of the distributions. General options are shown in Figure 8.1, illustrating log normal, normal, triangular and uniform distributions (note that the figure is illustrative and that the exact values and probabilities shown are not important to the results of the present study). Other distributions are possible, but are broadly similar to those shown. In cases where there is a reasonable amount of information it may be appropriate to assume a detailed distribution like the lognormal or normal case. In the present situation, however, we have only limited information and so it is appropriate to adopt a simpler assumption with the distribution around cost estimates assumed to be triangular or uniform. Choices made below reflect, in part whether a worst case or average/best case situation is under investigation Lognormal Normal Triangular Uniform Probability Value Figure 8.1 Examples of possible distributions for variables Once distributions have been determined for the input variables the Monte Carlo analysis is runover 10,000 iterations. At each iteration a value is selected at random from within each of the defined probability distributions that have been input to the model. The resulting distributions can be used to assess the extent of overlap in the distribution of cost and the distribution of benefit, telling us the probability that quantified benefits will exceed costs. The analysis is focused on the scenarios for the year 2020 rather than 2015 to account for more complete implementation of the policies under investigation. Describing the uncertainties in cost estimates Three options have been considered for sulphur controls, the use of open or closed loop scrubbers and the use of low sulphur fuels. For each technique a range of costs was derived taking into account various uncertainties (see the preceding chapter), and these ranges were used for the preliminary comparison of costs and benefits above. It should be pointed out that the low range scrubber costs are for sea water scrubbers. It is still unclear under what regulations these will operate (see Section 9). A likely scenario is that scrubbers operate with seawater at open sea and as closed scrubbers or with low-sulphur fuel close to shore. It may also be that seawater scrubbers are not used in sea regions with low alkalinity (i.e., parts of the Baltic Sea and the Black Sea. Therefore, a probable cost for scrubbers could be in the middle or higher part of the range given here for scrubbers. Various techniques were investigated for NO x control, as listed in Table 7.6. Total annual costs for 2020 are shown in Table 8.4 from which the following are apparent: Costs are dominated by controls on sulphur, the costs of controls on nitrogen being an order of magnitude less (with the partial exception of SEA1) The extent of variation in the NO x estimates is small compared to variation in the SO 2 control costs Despite a significant level of variation in costs for the two S control options, there is no overlap in their ranges. AEA 54

61 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Table 8.4 Summary of annual abatement costs given in Chapter 6, million SO2, scrubbers SO2, low sulphur fuel NOx SEA1, ,800 3,280 3, SEA3, ,640 5,550 9, SEA4, ,150 6,790 7,630 10, SEA5, ,170 7,070 10,960 13, These observations imply that the main focus of the uncertainty analysis should be on the variation in SO 2 control costs. They also raise questions about the appropriate shape for the probability distribution to be associated with the costing, with the following variations considered: Best case: Triangular or uniform distributions based on the sum of NO x control costs with the low range for SO 2 control costs (using scrubbers). Average Case: Triangular or uniform distributions across the full ranges shown, which would assume that ship owners opt for a mixed approach to new regulations. A uniform distribution is applicable where all values within a range are considered equally likely, whilst a triangular distribution applies when it is thought most likely that the true value lies towards the centre of the range. Worst Case 1: Triangular or uniform distributions based on the sum of NO x control costs with the high range for SO 2 control costs (using low sulphur fuel). Worst Case 2: A point estimate for costs at the upper end of the range: if this worst case generates a high probability of producing a net benefit there is no need to be concerned about other parts of the cost range as they would simply further increase the probability. For the first case we apply a triangular distribution with the most likely value assumed to be in the centre of the range. For the second case we assume a triangular distribution with the most likely estimate in the middle of the range on the assumption that some ships would use scrubbers and some would use low sulphur fuel. For the third case we assume a uniform distribution, as this seems likely to give a lower probability of attaining a net benefit (remembering that it is intended as a worst case) given the observed relationship between costs and benefits from Tables 8.1 and 8.2. Resulting distributions can be compared for scenario SEA5 in Figure 8.2. The two worst cases are shown to be significantly more pessimistic than the average case. Best case Average case Worst case 1 Worst case 2 Triangular Triang(2.416, (2.42, 4.99, ) 7.56) Triangular Triang(2.4, (2.42, 8.2, 14) 8.2, 14) Uniform Uniform(11.513, (11.5, 13.9) ) Point Uniform(11.513, estimate (13.9) ) Probability density Cheapest Value Most expensive Figure 8.2 Probability distributions used for the sensitivity cases on costs (examples are all for scenario SEA5). AEA 55

62 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Uncertainties in benefit estimates As noted above, account is taken here of uncertainties in the population at risk, the incidence of ill health, concentration-response functions and unit values for each of the health impacts that have been quantified. Ranges and associated probability distributions for each parameter are defined by Holland et al (2009). Bringing these data together is found to generate a log-normal distribution as shown in Figure 8.3 which takes the example of scenario SEA5 in The results in this form provide us with a mean value and confidence intervals the 90% interval being shown which covers a range from 9.23 billion/year to 26.9 billion/year. Using these results it is now possible to compare against the cost ranges shown in Figure 8.2. Probability density Mean=16.6 Mean= % 90% 5% Value ( billion/year) Figure 8.3 Calculated probability distribution for the benefits of Scenario SEA5 in 2020, with mortality valued using the VOLY Comparing costs and benefits accounting for uncertainties Results from the Monte Carlo analysis showing the probability of gaining a net benefit for each SEA scenario accounting for 8 sensitivity cases are shown in Table 8.5 and Figure 8.1 considering the set of uncertainties that have been addressed t this point. The following observations are made: : Using the VSL to value mortality the probability of deriving a net benefit is more than 98% for all SEA scenarios, even for Worst case 1 (the extreme worst case). Using the VOLY for mortality valuation the lowest probability of gaining a net benefit is 65% for the SEA5 scenario (compare Worst Case 2 in Figure 8.2 with the distribution calculated for benefits in Figure 8.3). AEA 56

63 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 For the best case assumptions (lower range for S control costs with full range for N control costs) there is a more than 99% probability of deriving a net benefit even when applying the VOLY. For SEA1 the probability of gaining a net benefit is greater than 98% in all cases. Perhaps most importantly, for the average case scenario there is a probability greater than 94% of gaining a net benefit for all SEA scenarios. The precise probability estimates reported in the table are themselves a function of assumptions made in the uncertainty analysis. What is important therefore is not the precise figures shown, but the overall perspective that when further uncertainties are taken into account, the view that benefits of the proposed actions are very likely to exceed the costs is unchanged. Table 8.5 Probability of deriving a net benefit under various sensitivity cases Worst 1 Worst 2 Average Best VOLY estimates SEA1 98.7% 99.2% 99.8% 100.0% SEA3 66.3% 75.2% 95.6% 99.8% SEA4 79.7% 88.3% 96.9% 99.5% SEA5 65.3% 75.5% 94.8% 99.7% VSL estimates SEA1 99.8% 99.8% 99.9% 100.0% SEA3 98.1% 98.8% 99.7% 99.9% SEA4 99.0% 99.3% 99.7% 99.8% SEA5 98.1% 98.7% 99.6% 99.8% Probability of benefit>cost VOLY estimates SEA1 SEA3 SEA4 SEA5 VSL estimates SEA1 SEA3 SEA4 SEA5 Worst 1 Worst 2 Average Best Figure 8.1 Probability of deriving a net benefit under various sensitivity cases. AEA 57

64 Cost benefit analysis to support the impact assessment AEA/ED /Issue Summary of the CBA results Results presented in this section demonstrate that the probability of benefits exceeding costs for all of the different scenarios investigated is high, in all cases greater than 65%. The IPCC (Houghton et al, 2002) provides some guidance on the interpretation of probabilities, as follows: <1% Extremely unlikely 1 10% Very unlikely 10 33% Unlikely 33 66% Medium likelihood 66 90% Likely 90 99% Very likely >99% Virtually certain For the cost ranges that we consider most reliable from the sensitivity analysis presented above (the average or best cases, rather than the worst cases) the probability of benefit exceeding cost can therefore be described as very likely / virtually certain, irrespective of the approach taken to mortality valuation. With one slight exception (SEA5 with mortality valued using the VOLY) the probability of benefits exceeding costs can be described as likely, very likely or virtually certain even for the worst case assumptions on cost. It is to be remembered that the probabilities calculated here are quantified only against benefits for improved health. The inclusion of impacts to materials (including cultural heritage) and ecosystems, in particular, would add to the benefits and further increase the already high probabilities. Economies of scale, as the use of abatement techniques becomes more widespread have also been excluded from the analysis. These would tend to reduce costs from the levels used here which are based on data from current suppliers. It is more difficult to make a qualitative appraisal of the effect of other uncertainties that have been excluded from the current analysis, for example price effects on demand for transport and the potential for modal shift in response for legislation, or price dynamics of supply/demand for fuels, equipment, etc. Some account has been taken of such variation, for example in the costs of abatement reagents such as NaOH and urea. It is, however, not beyond possibility that there could be even more significant variation in cost than that considered here in the future, and that this would lead to a net cost being incurred. However, from observations made here on the balance of costs and benefits we consider this unlikely for the following reasons: There is a high probability of gaining a net benefit even against the upper range for abatement costs The upper range for costs is about a factor 2 greater than the lower range We consider it likely that actual costs would be in the lower range than the upper range The following should also be considered: Alternative methods for reaching the same goals have been identified. If the costs of one option were to increase significantly it may therefore be appropriate to switch to a different option A conservative best estimate was adopted for the VOLY. The challenge of accounting for all possible uncertainties in this analysis, given the number of inputs and factors that need to be accounted for, is great. It would undoubtedly be possible to make alternative assumptions in some areas to those taken here and this would change the probability of gaining a net benefit. However, as the ranges etc. used here have been derived from discussion with a number of manufacturers (on the cost side) and on several occasions with stakeholders (on the benefits side), the principal conclusion reached, that benefits are very likely to exceed costs, seems robust. AEA 58

65 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 9 Emission abatement equipment 9.1 Summary This section of the report discusses technical criteria for exhaust abatement equipment for the reduction of emissions of sulphur dioxide from ships. The only technology available at present is scrubbing. Scrubber units can be of two kinds, seawater (open) scrubbers and freshwater (closed) scrubbers, or a combination of these. Criteria for the SO 2 to CO 2 ratio for emissions to air have been presented by the IMO. Also criteria for the wash-water have been developed which are discussed in this report. At present there are criteria for wash-water but no regulations regarding emissions to air from shipping (except for sulphur dioxide and nitrogen oxides); this is unfortunate from an environmental point of view and better optimised abatement units could be developed if the regulations covered both the emissions to air and to water. The wash-water criteria developed by IMO covers ph, PAH content, turbidity and nitrate content. While the criterion for ph seems to limit the amount of water that will be affected by low ph during scrubber operation to a reasonable amount specific attention is needed for waters with low alkalinity. The criteria for PAH is set in "phenanthrene equivalents"; it is not clear how this is defined but it seems to mean that the concentration of phenanthrene is monitored. The criterion corresponds to an emission of 2.2 mg/kwh of phenanthrene, this seems to be a high limit and widespread use of scrubbers emitting wash-water with PAH concentrations close to this limit could pose a risk to the environment. The criteria for turbidity can be a way to show that the scrubber is operating correctly but the method has not been shown to be a reliable measure of the concentration of particulate matter emitted with the wash-water. There is also a criterion for nitrate content in the wash-water. A scrubber unit on a ship operating close to this limit could, depending on the sea area, increase the ships' contribution to the nitrate load in the sea. It is suggested that regulations are constructed so that this can be avoided in sensitive areas. The IMO regulations allow two monitoring schemes. In one the SO 2 -emission to air is measured continuously and in the other the scrubber s operational parameters are monitored to check its efficiency of operation; it is suggested that the latter is applied because there will be scrubbers, notably closed scrubbers, where the operation can be monitored without continuously measuring the SO 2 concentration in the exhaust gas. AEA 59

66 Cost benefit analysis to support the impact assessment AEA/ED /Issue Introduction Emission of sulphur dioxide from burning of fossil fuels is still the most significant source of sulphur pollution to the atmosphere. These emissions have historically been the most important source for the acidification of lakes and soils leading to deterioration of soils, collapsing ecosystems and loss of forest production. Sulphur emissions also lead to the formation of particulate matter. High levels of particulate matter are a major health concern in large parts of Europe. In many parts of the world the emission of sulphur has recently decreased due to policy measures such as lower use of coal-burning, cleaning of exhaust gases, sulphur limits on diesel and gasoline for road transportation, etc. One sector that, until recently, has been unaffected by regulations regarding sulphur emissions is shipping. Here it is customary to use residual oil that contains high levels of sulphur in addition to e.g., high levels of aromatics. In principle, all sulphur in the fuel ends up as emission of sulphur oxides (SO X ) after combustion in the engine. Some of this SO 2 is then oxidized further and may then form particles. There is a clear connection between the sulphur content in the fuel and particle emissions (Winnes 2009). Eyring et al. (2005) estimate that the worldwide annual SO 2 emission from shipping is 12.0 Mtonne (almost three times more than for all road traffic). There is now a maximum allowed sulphur level of 4.5% agreed on worldwide. Further, so called SECA areas, (from 2006 the Baltic Sea, and from 2007 also the North Sea and the English Channel) have a limit of 1.5% sulphur. In view of the environmental and health impacts of sulphur emission IMO have decided on future reduction in the sulphur content in marine fuels according to Table 9.1. Table 9.1. Sulphur limits in marine fuels From year Max S-content (wt%) in fuel Area Worldwide SECA 2010 July SECA 2012 Jan Worldwide 2015 Jan SECA 2020 Jan 1* 0.5 Worldwide *may be postponed to 2025 In addition to these regulations there are some regional limits for sulphur in marine fuels. Within the European Union ships at berth and ships using inland waterways must use a fuel with less than 0.1% sulphur from Ferries within the EU must use a fuel with less than 1.5% sulphur. In the waters outside California ships have not been able use heavy fuel oil since July 1, 2009 and the sulphur content is restricted. In Sweden a system of lower fairway dues for shipping using low-sulphur fuel has been in use for a number of years. There are two principle approaches to reducing SO 2 emissions from shipping. The first alternative is to use a low sulphur fuel - either desulphurised residual oil or marine gasoil. The second is to remove the sulphur oxides by exhaust gas cleaning; this allows for the continued use of the relatively less expensive high-sulphur heavy fuel oil together with an SO 2 abatement system. The most practical abatement method is gas scrubbing, a technique where the sulphur oxides react with water and are captured as sulphates. This is an established technique that has been used for many years for land-based power plants etc. Scrubbers are also frequently used on large tankers as a way to produce inert gas that is used while unloading the cargo. The water used for scrubbing can be either re-circulated with a continuous addition of alkali, or alternatively, sea water - where the natural alkalinity of sea water is utilised for the capture of SO 2. The removal efficiency is normally high but depends on the alkalinity of the water. The latter also decides the acidity of the effluent water together with factors like the scrubber water flow in relation to exhaust AEA 60

67 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 flow. The scrubber will also remove a significant part of the particles and a minor part of the nitrogen oxides (NO x ) from the exhaust. The seawater is by various means filtered before disposed into the sea, producing sludge that should be treated on land. The regulations stating the sulphur limits in Table 9.1 allow for the use of abatement technologies, such as scrubbers, as an alternative to low-sulphur fuel. This is stipulated in Regulation 4 of Annex VI to MARPOL 73/78 and in article 4C of 2005/33/EC. The "equivalence clause" states that an alternative method should be at least as efficient in terms of emission reduction as the use of low sulphur fuel would be. Thus, the use of abatement of SO 2 in combination with HFO should be such that the environmental and health impacts are less than what they would have been had low-sulphur fuel (according to Table 9.1) been used. 9.3 Abatement of sulphur oxides Scrubber systems have been used since the 1930s in order to remove SO 2 from flue gases. The basic principle is that SO 2 reacts with substances in a spray of water, which essentially keeps pollutants in the exhaust from being emitted to air by capturing them in the liquid. In many land based systems, SO 2 is treated with limestone to form gypsum but some facilities utilise the natural alkalinity of sea water for sulphur scrubbing. On land, installations of SO X -scrubbers are relatively common because land based industries do not have the same constraints of space and complexity as ship based equipment. Alternatively freshwater, in combination with caustic soda, can be used as the scrubbing agent. The solid particles that end up in the water are allowed to settle in a sludge tank and the wash-water is treated before being disposed into the sea. If this treatment is not carried out, particulates and other pollutants will reach the sea water. The most important chemical reactions that take place during the desulphurisation process are the following. SO 2 (gas) + H 2 O + ½O 2 ==> SO H + HCO H + ==> CO 2 + H 2 O SO 2 absorbed in seawater reacts with oxygen to form sulphate ions and hydrogen ions. Increased concentration of hydrogen ions means increased acidity and decreased ph. Bicarbonate ions (HCO 3 - ) from the seawater react with hydrogen ions and thereby neutralise the acidity and raise the ph. SO 2 consumes oxygen to form sulphate and SO 2 concentration therefore corresponds to a certain Chemical Oxygen Demand (COD). In addition to the capture of SO 2 some NO x and CO 2 may be removed from the exhaust gas in the scrubber. For NO x it is mainly the NO 2 fraction that is captured and it will be emitted with the wash-water as nitrates. While some CO 2 may be captured and emitted with the wash-water this gives no net capture of CO 2 because the balance between CO 2 in sea and air will be reestablished by CO 2 emission from the sea. The alkalinity of the water is thus a key factor for the acidity of the effluent water. Furthermore, the efficiency of the scrubber process increases with higher alkalinity. There are two principle designs of marine scrubbers: seawater or open scrubbers and fresh water or closed loop scrubbers. In a marine seawater scrubber the flue gas either passes through a spray of seawater or is bubbled through seawater. The alkalinity of seawater is used to capture the sulphur and neutralise the wash-water. The sulphur is discharged into the sea with the treated wash-water which is acidic. In a freshwater scrubber the alkalinity is maintained through the addition of caustic soda and a bleed off of the scrubber water is emitted to the sea after treatment. There are also the possibility to combine the two types so that, for example, closed loop operation can be used in waters with low alkalinity and seawater operation at open sea. A number of trials have been undertaken with seawater and freshwater scrubbers onboard ships (see Skema 2009). AEA 61

68 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Three principle functions can be identified. 1. scrubbing of the flue gas where SO 2, particles and other constitutes are trapped in the seawater. 2. Filtering the seawater and dilution before disposal overboard. This leaves a sludge that is kept onboard for disposal at shore. 3. Demisting and heating of the exhaust gas after the scrubber unit in order to prevent a mist with low ph. 9.4 A comparison between reducing SO 2 emissions with low sulphur fuel or scrubbers The intention of the equivalent clause described above is to allow for abatement technologies to be used in combination with heavy fuel oil which may be economically advantageous. At the same time, the environmental impact when using the abatement technology should be less or equal to what would be the case if low-sulphur fuel were used. The writing specifically emphasizes emissions of sulphur dioxide but there are also other factors to consider. It can be useful to consider impacts which may be expected from the use of scrubbers in combination with high-sulphur heavy fuel oil and compare to the impact from use of low sulphur fuel oil. The abatement of SO 2 emissions required to be equivalent to certain sulphur content in fuel has been calculated by IMO and is expressed as a ratio between SO 2 and CO 2 concentrations in the exhaust gas. Thus, the SO 2 emissions for a ship using scrubbers, when fulfilling these criteria, will be lower or equal to what it would be with the stipulated fuel sulphur content. However, in many cases scrubber equipment would be able to reduce the SO 2 emissions significantly further, for example when the ship is used outside SECA areas. Since there is a running cost for a scrubber system, due to power consumption in pumps and, for closed loop systems, consumption of NaOH, it is likely that the scrubbers will be used so that the regulations are just fulfilled. Consequently there is potential for further reduction in SO 2 emissions that may be accessed through policy initiatives. In the production of low-sulphur fuel the sulphur from the refineries can be used for other products. There is a fuel (and thus CO 2 ) penalty in producing low sulphur fuel the magnitude of which depends on a number of factors in the refinery design and the crude oil that is used. For ship-based scrubbers the sulphur is emitted to the sea as sulphate ions; these will not constitute any significant addition to the sulphates already present. However, the acidity of the effluent may be an issue see above, its neutralisation is achieved either using the natural alkalinity of the sea water (open scrubbers) or by an alkaline chemical (closed scrubbers). In both cases there will be a net CO 2 emission, either during the production of NaOH or from the sea when carbonate and bicarbonate ions neutralise the sulphuric acid. The emission of particulates will be significantly reduced if the sulphur content in the fuel is low because there will be less sulphate-based particulate matter produced. If at the same time there is a switch from residual oil to distillate fuel then the PM emission will be reduced even further since the high concentrations of polyaromatics and ashin the residual oils result in high emissions of particles. A scrubber unit, as well as removing SO 2, has the potential of capturing a large fraction of the PM that would otherwise be emitted to the air; this approach reduces emission of soot particles and a range of metals and PAH. If the scrubber wash-water were not treated, however, the particulate matter would be discharged into the sea which is common for for scrubbers used to produce inert gas for some ships. The level of wash-water contamination for sulphur abatement scrubbers would be sufficiently high that it should be treated before being discharged to the sea in order to avoid risk to the local marine environment. Treating the effluent will produce sludge that must be taken ashore for safe disposal. However, since the regulations focus on using scrubbers for sulphur abatement only, they are not well adapted to utilize the potential of scrubbers as PM abatement systems; this is because a scrubber system that has not been designed, specifically to abate exhaust gas particulate can more easily comply with the wash-water criteria - less PM will be present in the wash-water and less treatment is needed. A potential solution would be to combine wash-water criteria and PM exhaust emission criteria in order to require that the scrubber units abate PM from the exhaust gas and treat the wash-water effectively. The two main alternatives to comply with the sulphur emission regulations in a SECA AEA 62

69 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 area after 2015 are to use marine gasoil with 0.1% sulphur or to use scrubbers; the latter option then allows the use of high sulphur residual oil. The first alternative is likely to give a reduction in PM emissions of about 90% (on mass basis) relative to residual oil with 1.5% sulphur. For outside SECA areas after 2020 heavy fuel oil with a sulphur content of 0.5% is likely to be used. This will give a reduction of around 65% relative to heavy fuel oil with around 3.5% sulphur. The reduction of PM emissions for the scrubbers that have been tested for marine applications has varied between 0 and 85%. The scrubbers were designed for SO 2 abatement and while a high PM abatement efficiency may be possible it is not a trivial matter to measure the PM trapping efficiency -when the temperature and humidity varies greatly between the measurements. In summary, the use of marine gas oil with low sulphur content will significantly reduce the emission of PM compared with the situation today. This will reduce the deposition to the seas as well as the contribution to the concentration of PM in the air over land. The use of scrubbers can also reduce the emissions of PM depending on the design of the unit. The emissions to water would be higher if the effluents were not treated. With effective treatment sludge will be produced and disposed at land. At present there are no real incentives to make scrubbers an effective abatement technology for PM. Regarding nitrogen oxides, scrubbers are likely to capture the NO 2 -fraction in the exhaust and may be designed to capture NO as well. Doing so may cause an emission of nitrates into the sea with the scrubber water which is unwanted in areas at risk of eutrophication. NO x emitted with the exhaust gas is distributed over sea and over land so a scrubber that captures NO x may redistribute the N deposition between the land and the sea, it would also decrease the contribution to particle formation from nitrates. Another difference between the two alternative methods of reducing SO 2 emissions - scrubbers and low sulphur fuel - is that ships would carry different fuels. In the case of an accident or with diffuse oil spills there could be either HFO or MGO present, each with somewhat different properties and different harmful effects, likely more severe in the case of HFO. Furthermore, a greater and more efficient use of scrubbers will produce sludge, for ports to handle, that is potentially toxic. 9.5 Emission criteria for EGC The emissions to air of SO 2 can be monitored as the SO 2 to CO 2 ratio from which corresponding fuel-sulphur content can be obtained. Since there may be capture of CO 2 in the scrubber, or release of CO 2 from the chemical reactions involved in the buffering, the CO 2 concentration should be measured before the scrubber unit. All sulphur compounds emitted to air should be measured and aggregated care should be taken to ensure that the more highly oxidised sulphur compounds are accounted for. The SO 2 to CO 2 ratios corresponding to certain fuel-sulphur levels have been calculated by IMO (see Table 9.2). No other emissions to air are regulated in the IMO guidelines for sulphur abatement technologies. Table 9.2. SO 2/CO 2 ratio corresponding to certain fuel-sulphur levels Fuel-sulphur content (wt %) SO 2 (ppm)/co 2 (%) ratio in exhaust Wash-water criteria The IMO Wash-water criteria are stipulated in section 10 of Annex 5 to MARPOL 73/78. We refer here to the preliminary guidelines dated July 16, The values that are to be monitored should include ph, PAH, turbidity and temperature. There are also limits on the nitrate content in AEA 63

70 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 the discharged wash-water. The text in this version of the guidelines implies that the wash-water criteria apply everywhere. However, in earlier versions the criteria were only applied to harbours, estuaries and ports. It is uncertain, at present, if separate limits will be developed for operation at open sea. ph ph is a measure of the acidic or basic (alkaline) nature of a solution. The concentration of the hydrogen ion [H + ] in a solution determines the ph. Ocean water usually has an excellent buffering system, with interaction of carbon dioxide and water, and ph generally varies between 7.5 and 8.5. Neutral water has a ph of 7 while acidic substances are less than 7 (down to 1, which is highly acidic) and alkaline substances are more than 7 (up to 14, which is highly alkaline). Anything either highly acid or alkaline would severely affect marine life but the oceans are usually very stable with regards to ph. According to USEPA the criteria for ph for effluents emitted to marine waters should be , but no more than 0.2 units outside of normally occurring range. Alkalinity is a measure of the buffering capacity of a solution and molecules, e.g. carbonates and bicarbonates, take up hydrogen from a solution. In the IMO wash-water criteria regarding ph it is stated that the wash-water should comply with one of the following: The discharge wash-water should have a ph of no less than 6.5 measured at the ship s overboard discharge with the exception that during manoeuvring and transit, the maximum difference between inlet and outlet of 2 ph units is allowed measured at the ship s inlet and overboard discharge. During commissioning of the unit(s) after installation, the discharged wash-water plume should be measured externally from the ship (at rest in harbour) and the discharge ph at the ship s overboard ph monitoring point will be recorded when the plume at 4 metres from the discharge point equals or is above ph 6.5. The discharged ph to achieve a minimum ph units of 6.5 will become the overboard ph discharge limit recorded. PAH Polycyclic aromatic hydrocarbons (PAH) are the largest known group of carcinogenic substances and include many individual chemical substances containing two or more condensed aromatic rings. The biological effect of PAH is linked to the planar structure of the molecule and its ability to affect DNA in the cell nucleus. PAH are formed when coal, oil, or other hydrocarbons are heated under anoxic conditions, i.e. with insufficient oxygen, and are the products of incomplete combustion. PAH are not usually present as individual compounds but occur in mixtures. PAH are fat soluble and might be bio-accumulative (in fatty tissues). In the aquatic environments, PAH are usually bound to particles which transport to sediment. Many PAH compounds accumulate in invertebrate organisms. Fish eggs and fry exposed to PAH have been found to suffer mutation, bleeding, heart conditions, reduced growth and increased mortality. A group of 16 PAHs are usually measured and analysed, but in the IMO wash-water criteria, PAH phe or phenanthrene equivalence is used. Phenanthrene is a member of the PAH group and is insoluble in water. PAH is sometimes used as an indicator of the total emissions of hydrocarbons. The IMO wash-water criteria regarding PAH: The maximum PAH concentration in the wash-water should not be greater than 50 µg/l PAH phe above the inlet water PAH concentration. The PAH concentration should be measured downstream of the water treatment equipment. AEA 64

71 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 The 50 µg/l limit is normalized for a wash water flow rate through the EGC unit of 45 tonne/mwh. The limit has to be adjusted upward for lower wash water flow rates per MWh to get a limit of 2.25 mg/kwh of PAH phe. For a 15-minute period in any 12-hour period, the PAH concentration may exceed the limit described above by up to 100%. Turbidity Turbidity is a measure of the amount of particles in the water, or rather the cloudiness caused by suspended solids. Turbidity in open water can be caused by several different reasons, e.g. phytoplankton or high levels of sediment in the water. In lakes and shallow areas the turbidity can decrease the amount of light reaching the lower depths and thereby affect submerged plants and in the end also species that are dependent on the amount of plants. Turbidity is measured in FTU (Formazin Turbidity Unit) or FNU (Formazin Nephelometric Units). The IMO wash-water criteria regarding turbidity: The maximum continuous turbidity in wash-water should not be greater than 25 FNU or 25 NTU above the inlet water turbidity. Nitrates Nitrogen oxides, i.e. the sum of NO and NO 2, are produced during combustion at high temperatures. The oxides can contribute to acid rain and other environmental effects. Nitrate (NO 3 - ) is the most highly oxidised form of nitrogen and excess nitrate concentrations in aquatic systems can lead to algae blooms and eutrophication. The nutrient concentration in seawater usually decreases during springtime and all processes in the nitrogen cycle are seasonally dependent. When oxygen is present, ammonia can be oxidised to nitrate (via nitrite) in a process called nitrification. The IMO Wash water criteria regarding nitrates: Nitrates should be monitored. The wash-water treatment system should prevent the discharge of nitrates beyond that associated with a 12% removal of NO x from the exhaust, or beyond 60 mg/l normalized for wash-water discharge rate of 45 tonnes/mwh whichever is greater. 9.7 Environmental assessments Behrends and Liebezeit (2003) presented an environmental impact analysis of sea water scrubbers. Laboratory tests with mixtures of seawater and sulphuric acid solution at ph 4, to mimic scrubber wash-water, were performed where the resulting ph was studied. A mixture of 1:10 in seawater was found to give a ph increase of around 0.1. This is compared to the EPA limit of 0.2 that is considered safe. For brackish water a higher change in ph was observed initially. In another report by Hufnagl et al. (2005) sampling was carried out of scrubber water and water from ports where a trial with a seawater scrubber onboard the Pride of Kent was undertaken. A decrease in ph of maximum 2 in the effluent was observed together with increased nitrate content. No impacts on seawater or increased toxicity were reported. Buhaug et al. (2006) discussed wash-water criteria for scrubbers and performed an environmental analysis. They suggested using a PNEC (Predicted No Effect Concentration) value for ph of 0.2. A set of criteria are suggested for three different levels of safety. The criteria are for THC and a set of metals. They also propose that no limits are required for ph and COD. The criteria are given in mass per volume of effluent water under the assumption that 45 kg water/kwh is used in the scrubber. They also suggest that the PNEC value for total hydrocarbons (THC) is set to 40 µg/l. Values recalculated to mass per engine output are given in Table 9.3. AEA 65

72 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Buhaug et al. use the concept of near field, outside which no effect concentration should be reached. This is set to 50 m from the stern for a moving ship and to 20 m for quay side manoeuvring. Table 9.3. Wash-water criteria after Buhaug et al. in mg/kwh Component safety level 1 safety level 2 safety level 3 PNEC THC Cu Ni V Pb Karle and Turner (2007) analysed the amount of water needed for scrubbing of SO 2 as well as the dilution of this water required for reaching a ph of 6.5 for different water types and temperatures. They showed that the amount of water needed to capture the SO 2 is significantly higher in river water than at open sea and with brackish water in between. They also showed that the wash-water with a ph of 6.5 will quickly be neutralised to reach a delta ph of 0.2 and that both the capture of SO 2 and the buffering (i.e. the dilution needed) are temperature dependent processes. GESAMP have commented on the wash water criteria in MEPC 59/4/19. No objections were raised to the choice of parameters to be monitored but they recommended that dissolved oxygen should be measured in addition. GESAMP also recommended that a generalised marine environmental risk assessment should be developed in the future when it is possible that many ships use scrubbers. According to GESAMP, the wash-water quality criteria appear to adequately reflect the hazards posed by wash-water discharge to the local marine environment. ph reflects the acidifying properties of dissolved SO x and its potential for acute effects on biota. PAH represents any oily discharge, NO x reflects any tendency towards eutrophication and turbidity is intended to safeguard against particulate matter (heavy metals and ash). GESAMP stated that an environmental benefit in reducing SO x emissions to the atmosphere only can be achieved if the discharge do not create further risks to the aquatic environment. One remark by GESAMP was that IMO should consider the potential contribution to ocean acidification of the large scale application of SO 2 capture from ships. The reference value for a safe change in ph used in the studies is 0.2 which originally comes from EPA. It has been found that wash water emissions from scrubbers is quickly diluted and buffered to reach this level, both in measurements and modelling. No clear data is given for how large zone around a ship is needed to reach this level (in ports or at sea) although the amount of dilution needed can be obtained from Karle and Turner. The PAH criteria is intended as a measure of oil and other hydrocarbons emitted with the wash water - this is the parameter where Buhaug et al. obtained emission values that are high and close to the PNEC; it is not clear from these reports how a PAH phe concentration is related to the emission of THC. Hufnagl et al. report that phenanthrene is the dominating PAH in the wash water and thus the best candidate for analysis. Turbidity is not discussed except that GESAMP regards it as a way to monitor particulate matter. Elevated nitrate levels were observed by Hufnagl et al. and they actually note that up to 20% of the NO x may be captured and released as nitrates. In addition GESAMP suggests that COD should be monitored and Buhaug et al. suggest that emissions of Ni, V, Cu, Pb and THC should be regulated. 9.8 Environmental impact related to the IMO wash water criteria The potential impact of aqueous discharges to the marine environment require information on both the potential toxic effects on representative marine organisms to different concentration levels of the discharged water in sea water, and the sensitivity of the marine environment in the region of concern. This should, furthermore, be related to the discharge conditions and the AEA 66

73 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 resulting rate of dilution and spatial extent of the plume formed by the discharge after disposal to the sea. There is a risk for locally high concentrations of a number of harmful substances if there are no adequate wash-water criteria in force. ph Hufnagel et al. have analyzed the environmental impact of a seawater scrubber for two harbours, Calais and Dover. They took samples from the seawater scrubber system and the samples partially showed values lower than ph 3. The buffering system increased the ph. Experiments showed that a mixture containing 40 % of seawater acidified to a ph of 4 (comparable to what happens in a seawater scrubber) only changed ph about 0.2 units. The lowest ph measured in the overboard discharge was 6.2. The US EPA, as mentioned earlier, has passed a guideline concerning the introduction of acids which states that outside the initial mixing zone the ph change is not allowed to be more than 0.2 units. Under the circumstances stated, the ph would not reach a value that is harmful to the marine environment. This is dependent on the dilution, volume of water and the specific area of concern. Buhaug et al. (2006) stated that discharges from a single ship are rapidly diluted when the ship is moving, but expected to be much slower when the ship is not moving. Karle and Turner calculated the amount of water needed for scrubbing SO 2 and dilution to first obtain ph of 6.5 and then a delta ph of 0.2. In addition they looked at further dilution to reach a maximum of 1% reduction in oxygen concentration. They considered different water types and temperatures for a 12 MW engine using a fuel with 3% sulphur. Adding all these factors gives an indication of the amount of seawater affected by a scrubber. Table 9.4 gives an indication of these amounts when scrubbing to a SO 2 emission corresponding to 0.1% fuel-s. The nominal flow rate of scrubbing water used by MEPC is around 45 tonne/mwh giving about 540 tonne per hour for a 12 MW ship. This should primarily be compared with the water needed for scrubbing in Table 9.4 Table 9.4 Approximate amount of seawater affected by a scrubber for a 12 MW engine using 3% S- fuel and with a seawater temperature of 10 C for scrubbing to a SO 2 emission corresponding to 0.1%S fuel. After Karle and Turner (2007). Open sea Baltic Proper River freshwater Water for scrubbing tonne/h Water for dilution to ph tonne/h Factor for dilution to delta ph Sum tonne/h The results in table 9.4 show that almost 7 times more water for dilution is needed when scrubbing takes place in fresh water, compared to open sea. As an example we can look at the port of Hamburg. In ktonne of CO 2 and 1.9 ktonne of SO 2 were emitted corresponding to a fuel consumption of about 36.3 ktonne with a sulphur content of 2.6% (Whall et al. 2002). If all of these ships were to use open scrubbers around 21 tonnes/s of water would be needed for the scrubbing and buffering, this can be compared with the flow in the Elbe, which is around 720 tonne/s. PAH The criteria for hydrocarbon emissions with the effluent water is given as PAH in phenanthrene equivalents. The reason seems to be that measuring PAH is a means to monitor hydrocarbons and phenanthrene was found to be the most abundant PAH in the analysis by Hufnagl et al. (2005). However, the concept "phenanthrene equivalents" is not well established and needs to be explained or replaced. Assuming that it simply means that the phenanthrene concentration is measured, the criteria correspond to about 2.2 mg/kwh of PAH phe. Cooper et al. (1996) analysed AEA 67

74 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 PAH in the exhaust from two ships and found phenanthrene levels in the gas to be 71 and 250 µg/kwh respectively. The lower value refers to an engine using HFO. Thus, the wash-water criterion seems to be set rather high - considering that the scrubber water should be treated before the samples for the PAH analysis are taken. A 12 MW engine could, with the suggested limit, emit around 26 g phenanthrene per hour and in addition unknown amounts of other PAHs. According to Buhaug et al. (2006) OSPAR have developed Ecotoxicological Assessment Criteria (EAC) for a large range of chemicals, including PAHs. This was done to make the evaluation of monitoring data on toxic substances in the marine environment easier. The EAC can be applied as PNEC (Predicted No Effect Concentration) values. The EAC-water for phenanthrene is proposed to (µg/l) (provisional value). The PNEC value is the maximum permissible concentration that can be sustained over time. The Predicted Environmental Concentration (PEC) equals the proposed wash-water criteria multiplied by the dilution (if inlet concentration is assumed to be zero). Since the wash-water criterion is set to 50 µg/l PAH phe (for a scrubber water flow of 45 l/kwh) the wash-water must be diluted times to reach this PNEC concentration. Furthermore, the background concentration should also be considered, which may lead to further limits on the emissions. In Directive 2008/105/EC EQS (Environmental Quality Standards) have been set for priority substances. The EQS are not applicable for PAH as a group but for identified substances e.g. for Benzo(a)pyrene, the sum of Benzo(b)fluoranthene and Benzo(k)fluoranthene, and the the sum of Benzo(g,h,i)perylene and Indeno(1,2,3-cd)pyrene; the Directive does not include EQS for phenanthrene. It is still to be established, furthermore, that the PAH criteria used will ensure that the levels of total hydrocarbon compound (THC) emissions are at safe levels. The criteria for THC developed by Buhaug et al. are 4500, 450 and 45 µg/l for the three "Safety Margins" that they use. This is under the assumption of a wash-water flow of 44 tonne/mwh. These levels would correspond to 200, 20 and 2 mg/kwh. Thus the IMO criteria which applies only to phenanthrene is set to about the same as the THC value by Buhaug et al. for safety margin 3. If we look at the port of Hamburg again, the amount of PAH phe emitted were all ships to use scrubbers and allowing the scrubber water to be at the criteria limit would be around 400 kg phenanthrene per year. Turbidity Turbidity is chosen as a means to measure the suspended solids in the water and because it a simple method for on-line analysis. Turbidity can be a useful indicator of the status of the scrubber, i.e. that suspended particles are not being formed in the scrubbing process or the wash water treatment. However, turbidity is not a direct method of determining the number of exhaust particles that end up in the wash water - there is no direct correlation between turbidity and particle concentration. Further, turbidity values are strongly dependent on the size of particles in the water and smaller particles are likely to have significantly less influence the measured turbidity than larger ones. Studies are required to correlate the turbidity measurements with concentrations of particulate matter, ash, metals etc in the wash water before this criteria can be considered an indication of particulate levels in the wash-water. These uncertainties make it hard to make an environmental assessment of the turbidity criteria. Nitrates The criteria for nitrates are motivated by the increased contribution to eutrophication that will be the result of increased emissions of nitrates with the effluent water. Since scrubbers are likely to capture the NO 2 in the exhaust the 12% maximum capture requirement of NO x will normally be met. Hufnagl et al. (2005), however, report a capture fraction of up to 20%. To analyse what the 12 % criteria could mean it is of interest to look at the Baltic Sea, which is heavily trafficked, it is a SECA area and it is highly affected by eutrophication and large efforts are made to decrease the flow of nutrients into the Baltic. The total emission of NO x from shipping in the Baltic according to a recent Helcom study (Jalkanen 2009) is around 400 ktonne per year at present. EMEP AEA 68

75 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 modelling (Benedictow 2009) has calculated how much of the emissions from a certain area is deposited in another area, from this work it can be concluded that about 13% of the NO x emitted from shipping in the Baltic Sea is deposited in the Baltic Sea - the rest is deposited over land or (to a smaller extent) in other sea regions. Consequently ~ 107 ktonne per year is deposited in the Baltic Sea from shipping in the Baltic Sea calculated as nitrates. If all of these ships were to use scrubbers and all of these scrubbers captured 12% of the NO x in the exhaust and released this as nitrates in the scrubber water the total deposition would increase to 193 ktonne, i.e., an increase of about 86 ktonne per year. This is, of course, an unrealistically high number considering that not all ships will have scrubbers and that the capture fraction will, in practice, be lower. Further, a certain fraction of the nitrates deposited over land would return to the Baltic Sea with rivers etc.; it does give, however, a maximum possible contribution increase of nitrates. Further, of the total load of nitrogen to the Baltic Sea 25% comes from deposition and the rest via rivers and direct waterborne discharges. Thus the hypothetical maximum increase in load of nitrates, if all ships were to use scrubber, would be about 2%. COD, the chemical oxygen demand, is a measure of the theoretical oxygen consumption of a water sample, i.e. the amount of oxygen needed to break down the organic matter in it. According to Buhaug et al. there are no environmental standards for COD. Tolerance to reduced dissolved oxygen varies between organisms, life stages and environmental conditions. While Buhaug et al. concluded that COD was not a matter of concern under the circumstances of their study. Karle and Turner used a criterion that the COD should be less than 1% of the available free oxygen, which they found to be met when their criteria for ph were met. 9.9 Monitoring In directive 2005/33/EC, article 4c it is stated that ships using abatement technology to achieve low SO 2 emissions must be fitted with continuous monitoring equipment for SO 2 emissions in order to establish that the emissions at all times are lower or equal to that which would have been obtained by using fuel with the stipulated sulphur content. In the "revised guidelines for exhaust gas cleaning systems" from MEPC dated 16 July 2009 there are two permissible schemes, Scheme A (unit certification with parameter and emission checks) and Scheme B (continuous monitoring with parameter checks). In Scheme A the unit should be certified to meet the limit values for SO 2 /CO 2 concentration in the exhaust. The fuel that may be used, including sulphur level, must be specified together with other operational parameters such as gas flow, water flow, and salinity of sea water or fresh water elements necessary to provide adequate neutralisation agents. A SO X emission compliance certification must be obtained from the administration. There must also be a certification procedure established for use during surveys. The system must automatically record that the abatement equipment is in use by monitoring wash water pressure and flow rate at a frequency of at least Hz. Further, daily spot checks of the SO 2 /CO 2 ratio are "recommended". If the manufacturer is unable to provide assurance that the unit will meet the certified value this way, Scheme B is recommended. In Scheme B compliance is demonstrated by continuous monitoring. Thus SO 2 and CO 2 concentrations are to be measured after the abatement equipment with a frequency of at least Hz. In this scheme parameters such as exhaust gas flow rate and water flow rate are to be checked daily. There should be a technical manual where it is stated, for example, the minimum and maximum water flow rate and minimum inlet water alkalinity and other parameters for the operation of the unit. The manual should be approved by the administration. The procedure for emission testing and data recording is described and the same for both schemes. The recording equipment must be tamper proof. For both schemes the ship must have a SO X emission compliance plan showing how the equipment should be used to comply and how this is demonstrated. AEA 69

76 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 In addition, both schemes will require continuous monitoring of the wash-water criteria discussed above, i.e., ph, PAH concentration, turbidity and temperature. Thus, in Scheme A the function of the abatement equipment is certified by continuously recording operational parameters such as water flow and pressure as well as wash-water parameters. In Scheme B the compliance is verified through direct measurements of the SO 2 emission together with the wash-water parameters. Only Scheme B complies with the wording in Directive 2005/33/EC. The advantage with continuous monitoring of the SO 2 /CO 2 ratio is that compliance with regulations is directly monitored at all times. There are also some disadvantages: continuous monitoring of exhaust from engines can be expensive and technically problematic, this is discussed by CIMAC in a report from 2005 and by Johnsen et al, the measurement systems may also be affected by the high temperatures and the high particle load in the exhaust. Further, vibrations, variation in ambient temperature, spray of liquids and fluctuation in voltage may influence the measurements and the monitoring systems may then require much maintenance and calibration in order to function properly. Continuous monitoring can be of great value especially if the conditions that may influence the performance of the abatement equipment are likely to change. For sea water scrubbers this may be when moving in areas with seawater of varying alkalinity. In waters with low alkalinity more seawater is required to meet the required SO 2 emissions and further, more dilution will be required to meet the wash water limit for ph. For closed loop scrubbers the function is not dependent on the characteristics of the surrounding seawater since the operation of the equipment involves fresh water and an added neutralising agent. The requirements in the MEPC guidelines make it possible to use Scheme A for closed scrubbers, it may also be suitable for some sea water scrubbers. During certification the feasibility of using Scheme A should also be investigated, i.e., by showing that the monitoring of the system parameters gives evidence that the SO 2 emission and wash-water requirements are met. There may well be cases where a scrubber is not certified, for example, for use in rivers or brackish water systems and then the functionality can be expected to be more stable. If a system, during certification, is found to have too low a capacity under certain conditions, such as alkalinity, sea water temperature etc, it is possible to certify the unit using Scheme B. Since it is possible that there will be a number of systems that can be monitored satisfactorily using Scheme A it is recommended that the IMO regulations regarding monitoring are incorporated also in the EU regulations. With the use of either scheme it is of great importance that the SO 2 emissions and wash-water criteria are controlled. This might be achieved via regular certification measurements required by administrative provisions. Further surveillance of the SO 2 emissions can be made by authorities using remote sensing of the SO 2 /CO 2 ratio in the exhaust plume from the land or from airplanes, there is an ongoing program for the development of such systems at the JRC. Mellqvist et al. (2008) have demonstrated such a method with surveillance from airplanes. The uncertainty in such a measurement is expected to be around 10% and thus scrubber units that are defective or turned off should be detectable. One problem may be that scrubbers can capture some CO 2 and thus the SO 2 /CO 2 in the exhaust may show values that over-represent the actual SO 2 emissions. This problem could be circumvented by listing the scrubber units' CO 2 capture capacity during certification Conclusions and suggestions The use of scrubbers to clean the exhaust from marine engines using residual oil could in principle lead to high concentrations of a number of harmful compounds in the marine environment around the ships. The use of scrubbers in combination with high sulphur residual oil AEA 70

77 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 may be an economically attractive option in SECA areas and also worldwide from 2020 and so, as a consequence, a large number of ships would use scrubbers; consequently it is of great importance that measures are taken to assure that the marine environment is protected. There should be mechanisms to prevent the use of scrubbers in certain areas if there is a risk to marine ecosystems. The technology allows for the use of closed systems - making it possible to limit the wash-water discharged in areas with limited water circulation or especially sensitive environmental conditions. Scrubbing systems can be used to abate particulate emissions. The present situation is somewhat incongruous in the sense that wash-water limits for scrubbers have been set up while there are only limits for the gaseous emissions of SO 2 and NO x a situation that does not reflect current concerns related to the health impacts of PM. It is recommended that criteria are developed that promote the abatement of PM with, as a minimum, limits set to allow either voluntary or market based reductions of PM emissions; it is suggested, furthermore, that where/when possible, EU regulations are kept in line with the IMO guidelines. The wash-water criteria include limits for PAH, ph, turbidity and nitrates. Of these all except the nitrate criterion are intended for continuous monitoring onboard the ship. So, a major factor to be considered when formulating the criteria is that they should be suitable for online monitoring. However, it would also be possible to add additional criteria that are to be tested during certification. This could give further assurance that the units are environmentally safe and would not have to be constrained by the need to measure online. The wash-water criteria apply everywhere but it is possible that additional, separate, criteria can be developed for open sea. With the present criteria for ph it would appear that the dilution in the sea and natural buffering will rapidly give acceptable levels - even in confined areas. It is recommended, however, that further studies are undertaken for brackish and river water and that the criteria are then reviewed accordingly. It is suggested that the IMO guidelines are adopted for the ph criteria while, at the same time, allowing for revision as more data becomes available. If the criteria for open sea were relaxed then more acidic scrubber water may be emitted to the sea - however, the same volume of water would be affected because less water would be used for dilution within the ship. The PAH criteria in the IMO guidelines are expressed as phenanthrene equivalents. This is not a well established concept and should be further defined in the guidelines. The criteria correspond to an emission of phenanthrene of 2.2 mg/kwh which could, in principle, lead to large emissions of PAH from marine scrubber units. The criterion was probably chosen in order to be easy to measure with continuous monitoring. While it may be appropriate to use phenanthrene for monitoring purposes the criteria ought, nevertheless, to ensure a low discharge of a range of PAHs so how the measured value relate to the PAH 16, for example, or its correlation to THC needs to be established during certification; this correlation may differ for different scrubber systems and fuels. It is suggested that the PAH criteria that are used are in line with Directive 2008/105/EC. The criteria for THC suggested by Buhaug et al. could be used for commissioning. It is yet to be established that PAH phe is an adequate measure of the concentration of oil or THC in the wash-water. The criteria for turbidity may be suitable for establishing that unwanted solid material is not formed in the scrubber unit but it is unlikely to be appropriate as a measure of the proportion of particulate matter from the exhaust gas that captured and discharged with the effluent water. There is no simple relation between particle concentration and turbidity and the turbidity measure is strongly dependent on particle size. It is suggested that: the concentrations of particulate matter, metals, and PAH are regulated and tested during commissioning, and that the correlation to turbidity is investigated in order to allow continuous measurement onboard. Pending the delivery of data from on-going trials it is suggested that the turbidity criteria from IMO are used, alternatively the criteria suggested by Buhaug et al. could be applied. The criteria for nitrates could in principle lead to large additional discharges of nitrates to sensitive areas. The best way to deal with this issue would be to use exhaust gas NO x abatement technology. Taking the Baltic example described above, an engine with an emission factor of 17 AEA 71

78 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 g NO x /kwh should reduce its emission to about 9.4 g/kwh (if 12% of the NO x is emitted as nitrates with the scrubber water) in order not to increase the nitrate load when using a scrubber. It is recommended that provisions are made when setting criteria for nitrate emissions so that increased loads can be prevented in sensitive areas. Since there is limited data on the performance of marine scrubbers for SO 2, it is important to follow on-going trials in order to ensure that thorough environmental impact assessments are conducted when evaluating and developing wash-water criteria. A larger introduction of scrubbers could lead, potentially, to high concentrations of e.g. PAH if the units actually emit PAH in concentrations close to the criterion limit. There is also the opportunity of studying the impact of inert gas scrubbers on the marine environment; these are frequently used though smaller than exhaust gas scrubbers made to comply with sulphur emission regulations. It is also important to obtain more data on the properties and amounts of the sludge produced. This will be needed in order to develop facilities for handling of the sludge by ports. It seems appropriate to adopt the IMO guidelines regarding monitoring, i.e., allowing for the possibility of using alternatives to continuous monitoring of the SO 2 emissions on the basis that there will be units that can be adequately monitored by other means. AEA 72

79 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 10 Use of distillate fuels by ships at berth 10.1 Introduction From 1 January 2010 ships using ports within the EU will be required to use low-sulphur fuel, maximum 0.1 wt% sulphur, while at berth, with an exception for those making only short stops (Directive 2005/33/EC). The regulations apply also to ships that are anchored. The fuel switch should be done as soon as possible after arrival and as late as possible before departure. This regulation applies to all engines and to all boilers, including auxiliary boilers. These regulations will most likely make it necessary to use distillate fuel while at berth. There are concerns that there are risks associated with switching from residual oil to distillate fuel in boilers. This report has been drafted under the project Service contract for a Cost Benefit Analysis to support the Impact Assessment sulphur content of certain liquid fuels, which was undertaken for the Commission by a team including TNO, IVL and AEA under an AEA-led framework contract. It tries to clarify the risks associated with the problem outlined above and aims to address the following questions: Identify the type of ships where these risks may be relevant. Estimate the number of such ships in 2010 and Describe the nature of these risks. Indicate technical solutions and assess the costs of these. The study has been undertaken through literature searches and searches on the internet as well as through interviewing representatives of industry and relevant authorities. We have had contact with manufacturers of boilers and accessories, representatives of ship-owners, port authorities, and maritime administrations. There are only a few reports that address the problem Outline of the problem Boilers that are constructed for the use of Heavy Fuel Oil (HFO) cannot, in general, be used with Marine Gas Oil (MGO) without modifications. A switch from HFO to MGO in a boiler that is designed for HFO use could lead to operational problems with potential flame failure and an increased risk of a boiler explosion. To avoid these problems, boilers may have to be modified before a switch from HFO to MGO can take place in a safe way. The modifications needed must be assessed for each boiler individually. The exact number of boilers that will need modification, as well as the costs and time needed for these modifications, are difficult to assess because the figures are not well known by the ship-owners. Boilers are typically used at berth and tankers, in particular, use large boilers for heating the cargo and for steam-driven cargo pumps. A number of LNG-carriers use boilers for propulsion where boil-off cargo is used as fuel in combination with HFO. These boilers, which are also used at berth for cargo transfer, need to be modified before they can be used with MGO. In order to comply with Directive 2005/33/EC, ship-owners must assess what boilers need to be modified, to what extent, and have these modifications done before the end of Some shipowners have made the assessments and taken steps to do the necessary modifications but many ships still need to be investigated. The process seems to have been started during 2008 and, in most cases, activity intensified in The reason that the industry has not acted earlier seems to be multifold. Some ship-owners anticipated that the EU directive would be brought in line with the IMO directives of Marpol Annex 6 and that the 0.1 sulphur-limit would be implemented in as it will be for SECA areas. It was also expected that the sulphur directive would be reviewed AEA 73

80 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 before 2010, and some ship-owners seem to have anticipated, until relatively recently, that HFO with 0.1% sulphur would be made available Types of ships where the risks may be relevant Boilers are used on ships either for propulsion or as auxiliary boilers. Boilers for propulsion are used in combination with steam turbines and, although once abundant, they are today relatively uncommon because diesel engines have a higher thermal efficiency and are more fuel efficient. An application where boilers are used for propulsion is for ships that transport LNG, the reason being that some of the cargo boils off during transport and using it as fuel for the ship recovers a part of the cargo that would otherwise be lost. For reasons of thermal efficiency, these boilers operate at high pressure ~60 bar. Some newer ships instead use reliquefaction plants to collect the boil-off gas. Boil-off gas (BOG) is sometimes used in combination with a liquid fuel and, depending on power demand and operational mode, the final fuel mix can be either 100% BOG or a combination of oil and BOG. The reasons for the dual fuel arrangements lie in safety and the need for reliability for propulsion. For example, when manoeuvring - or for other purposes that demand a rapid response in boiler power, the gaseous fuel alone may not be sufficient. In such situations, dual fuel mode is used where oil and BOG is burnt simultaneously. A change in power can then be taken up more slowly by using gas. Further, in cases when the gas supply to the boilers is unintentionally reduced or interrupted, the system can use a fuel-oil boost up to ensure continuity in power. There is also a limited turn down ratio for gas burners and so oil is used at low loads too. The boiler powered LNG ships used today are designed for a combination of BOG and HFO. Typically, a ship that would use 150 tonnes HFO per day if operated on HFO alone uses around 10 tonnes of HFO in dual fuel mode. The boilers use small amounts of MGO during flashing from cold when furnace temperatures are low (normally after refit) but cannot use MGO during normal operation without modifications (see below). Auxiliary boilers may be divided into: 1) small boilers used on ships to produce steam that can be used to heat residual oil, to produce hot water and for the purpose of heating; and 2) larger boilers used on tankers for heating cargo and to drive steam turbine pumps. Auxiliary boilers are used on practically all ocean going ships. These boilers can be designed to be used with either HFO or distillate fuel. Some boilers can use both HFO and MGO. Typically, boilers are used with HFO on ships that use this fuel for propulsion. The type of fuel used in auxiliary boilers has not been investigated in detail since the fuel consumption is normally assumed to be a minor part compared to main and auxiliary engines; it is difficult, therefore, to establish the exact distribution of boilers that use only MGO, only HFO or a combination. The vast majority of boilers, however, use HFO and this is certainly true for the larger auxiliary boilers on tankers Relevant number of ships The need for an assessment of the potential modifications to existing boilers to manage the switch from HFO to MGO applies, in principle, to all ships by It is today not clear how many of these will need modifications. Work by the leading manufacturer of boilers, Aalborg industries (Aalborg 2009) and a study by the California Air Resource board (CARB 2008) indicates that most boilers can at least temporarily use MGO without problems. However, both studies, as well as a report by DNV (DNV 2005), state that MGO should not be used in a boiler designed for HFO before it has been inspected. Further, a large number of boilers will need modifications before they can be used safely with MGO. Although a number of ship-owners have assessed and when necessary modified their boilers, it seems unlikely that all boilers will be able to run on MGO by the 1st of January According to OCIMF at least of chemical and oil tankers that enter EU ports use steam for cargo pumps or cargo heating. Other ships may use diesel engines or electrical motors AEA 74

81 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 to drive the pumps, and some use electrical heating. These boilers are more likely to need modifications since they use relatively large amounts of fuel, which affects valves and fuel pumps. A number of tank owners will have difficulty making these modifications before the end of the year, although the exact number is difficult to say. Modifications on all ships will, in all likelihood, not be complete until late The boilers on tankers are often used during the discharge of the cargo. Intertanko estimated that in 2006 there were port calls of tankers in the EU. With a fuel use per discharge of tonnes, this amounts in the EU to a total fuel use of around ktonnes per year. There are 262 LNG ships with steam turbine propulsion, of which most service EU ports. Thirty two ships use diesel propulsion while 19 are dual fuel diesel electric. A number of LNG ships are several decades old but the production of LNG tankers has increased a lot recently. About 43% of the LNG ships were delivered between 2005 and The trend for new ships seems to be that diesel engines in combination with reliquefaction plants are preferred to steam propulsion. There is also a development towards dual fuel diesel engines that can burn fuel oil (distillate or HFO) and gas. LNG tankers are refitted about every 2.5 years - or every 5 years for newer ships. If modifications needed to comply with the sulphur directive are carried out during refit it will be approximately 2014 before all ships are modified. The ships could also be taken out of service for the modifications, which could be complete within approximately one - two years. The latter alternative then implies extra costs when the ships are out of service for modifications. There are approximately 690 deliveries by LNG tankers per year to the EU. At each port visit, around 50 tonnes of HFO is used adding up to about 35 ktonnes per year. By 2020 all ships will have had the time needed to make the necessary adjustments on their boilers The nature of the risks There have been reports of boiler explosions due to defects or improper operations. This can occur if, for some reason, the flame goes out, a high pressure of fuel gas builds up in the burner, and the control system is malfunctioning or disconnected. In such a case, the furnace must be purged before ignition. Failure to do this may cause an explosion, as for example reported by the British Department of Transport (1983). There are flame monitoring and control systems that will automatically purge the furnace so that during normal operation the risks for explosions are small. There are concerns that these risks may be higher when switching over from residual oil to distillate fuel. This is noted by the IMO subcommittee on bulk liquids (2007), which states: "The routine switching between marine distillate and Heavy Fuel Oil (HFO) has safety implications for existing boilers. The boiler combustion systems may require to be modified to prevent the risk of boiler furnace explosions in distillate operation". The risks are also noted in a recent US Coast Guard maritime safety alert (2009) in relation to regulations for California which apply from 1 July These regulations stipulate that Marine Gas Oil (MGO) or Marine Distillate Oil (MDO) must be used in all engines including boilers (with the exception of boilers for propulsion). The notice points out that "proper procedures, training, and maintenance are essential for vessels to safely switch between heavy/intermediate fuel oils and marine distillates". The notice is about all engines and auxiliary boilers and suggests that each system should be assessed for modifications and that training and routines should be established. The reason for the special concerns regarding a fuel switch from HFO to distillates is that the pipes and other parts are heated when using residual oil. When the distillate enters the furnace there are concerns that it will fail to ignite, causing a build-up of gases that in turn is an explosion risk. When distillate fuel is present in hot piping vapour locks may cause irregular fuel flows to the burners (this may also be an issue for diesel engines). This irregular fuel and vapour flows/bursts towards burners can cause flame extinction; the lighter the fuel, the easier the evaporation and AEA 75

82 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 the larger the risk for an air/fuel mix which is potentially unsafe. There are concerns that the crew on many ships may not be familiar with managing these situations safely. The boilers normally run on MGO during start up; however, at this point the system has not yet been heated. After the boiler pressure is up, the pipes can be emptied of MGO and then heated and used for HFO. MGO is thus not introduced to the boiler through heated pipes at this stage. In normal ship installations, "heat tracing" of pipes (by steam or electric heating) can be turned on or off. In this kind of installation, proper operation may be sufficient to avoid the risk of an explosion. In many cases, fuel switching is automatically managed by the ships hardware and software (and there is no difference between fuel flows to engines and boilers in this respect) - this may be the situation in modern ships that have been delivered during the recent years. When fuel switching cannot be done automatically, however, the ship s officers must follow very precise operational instructions. Besides concerns about explosions inside the boilers, there are other issues that need to be addressed before a boiler designed for HFO can be used with MGO. These, reported by DNV (DNV 2005) and Aalborg industries (Aalborg 2009), include: Fuel pumps and valves may need to be replaced due to the different viscosity of the fuels. There is an increased risk of wear and pump breakdown if the oil has a higher viscosity and lower lubricity than what the pump is designed for. There may be increased smoke emissions because the amount of fuel injected in the burner is set for HFO which has a higher density and lower calorific value than MGO. On certain types of burners (rotary cup) coke deposits may build up when using distillate fuel. In steam atomising burners the MGO may be heated - causing vaporisation. Other concerns raised, by the shipping industry, about the regulations from 1 January 2010 related to: 1) problems with the supply of fuel with 0.1% sulphur for ships at berth; 2) issues with ships not having enough fuel tanks to store the different kinds of fuels that they will need; and 3) concerns that there is not enough time to modify existing boilers and train crews. The objective of the present study is not to investigate possible problems with the supply of MGO. However, according to a study by Entec (2002) c ktonnes of fuel was used for manoeuvring and in port in the EU in the year The main portion of this fuel will likely have to be MGO from 2010 and, considering the increase in shipping, a rough estimate is that around ktonnes of MGO is needed. For the LNG ships with steam propulsion, the issues are essentially similar to those described above - with some additional concerns. Since the boilers are used for propulsion, it is essential that the fuel supply is not interrupted. As described above, there must be a reliable supply of fuel oil in addition to the BOG. The fuel oil systems in these ships are designed for HFO. They should not be used with distillate fuel for a number of reasons: 1) There would be an increased risk of failure in fuel pumps and valves. 2) The distillate fuel could be unintentionally heated in pumps or in the burner leading to the mentioned problems with evaporation. 3) Most of the burners used have fuel injectors of the concentric type where the atomising steam is in between the tubes for fuel oil where distillate fuel may be heated. Other burners have parallel tubes for steam and fuel oil. Here the heating of the distillate fuel is less of a concern but temperature gradients when cold fuel oil is used could cause distortion of the tubings used for oil supply. AEA 76

83 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 4) The burner management system and flame supervision/monitoring systems are designed for HFO Technical solutions and costs For the safety and operational reasons discussed above, all boilers designed for the use of HFO should be inspected before they are used with distillate fuels. Many boilers may be used temporarily with gas oil without modifications, but for others modifications may be needed on fuel pumps, burners or pipes, as well as flame supervision and adjustments in air/fuel ratio, fuel flow and post-purging sequence. These modifications are further described in Aalborg (2009) and DNV (2005). Modifications will have to be approved by a Classification Society. The solutions for LNG tankers with boiler propulsion are discussed separately at the end of this section. The risk of a boiler explosion, as discussed above, is associated with the vaporisation of distillate fuel in combination with flame failure and improper purging. To avoid the heating of MGO in the piping system, heaters should be by-passed or the tracing of fuel pipes turned off. The fuel system should have time to cool to around 60 C before MGO is introduced. The pipe system may have to be converted to contain two parallel feeds to the burners. The fuel may also be unintentionally heated in fuel pumps that then have to be adjusted or replaced. There are potential problems with gassing in steam atomising burners, which may have to be modified. Protective heat shields should be present on rotary cup burners to avoid deposit build-up and unintended ignition after flame failure. The flame supervision system may have to be modified to include another flame scanner - due to the different properties of HFO and MGO flames. The control system may have to be modified to implement automatic post-purging sequence. This will ascertain whether a potentially explosive gas mixture in the furnace has been purged in a safe way. Fuel pumps may have to be modified or replaced to be able to tolerate MGO, with lower viscosity and lubricating properties than HFO, for a longer period. In some cases, it may be sufficient to modify the control system. Fuel pressure may have to be adjusted and/or burner nozzles replaced To avoid smoking, the air/fuel ratio may also have to be adjusted. The costs associated with these measures will vary depending on the modifications needed. The exact distribution of costs for the existing boilers is not known. Boiler suppliers estimate costs in the range of to per boiler when modifications are needed. For tankers with large boilers, costs around per ship have been quoted. There will also be costs for approval and certification. For LNG ships with steam turbine propulsion, the boil-off from the cargo is used in combination with HFO. For the reasons described above, these ships should be modified before they are used with MGO. The boiler management system (BMS) settings will have to be adjusted, new flame scanners may have to be installed, and a separate fuel supply system for MGO will be needed. The majority of burners will have to be modified to avoid vaporisation. Further, fuel change-over procedures have to be investigated. Some ships may also need larger tanks for MGO in order to accommodate enough fuel for one port stay. These modifications and actions would make it possible to use MGO rather than HFO at berth and thus comply with the directive. There are still some concerns that there may be an increased number of flame failures and that the main boiler manufacturers lack experience in burning MGO during normal operation. After the modifications, however, these incidents would be handled in a safe way by the BMS, although there may be interruptions and operational problems. The costs for these measures have been estimated by SIGTTO. The costs are difficult to estimate and figures ranging from to per ship have been cited with a median of around 600 AEA 77

84 Cost benefit analysis to support the impact assessment AEA/ED /Issue To this should be added costs for taking the ships out of service if the modifications are done outside a regular refit. An alternative solution would be to use 100% BOG while at berth. There are, however, concerns about the stability of the boiler power at low loads. It is unclear if this can be implemented on all ships and modifications to the control system and BMS would be needed. Further, there are situations where 100% BOG cannot be used. This applies for relighting after a flame failure and in cases where there is no LNG cargo, as when ships enter/leave refit/dry dock. Therefore, this alternative does not comply in full with the EU directive. No cost estimates are available for this alternative because the extent of modifications needed has not yet been investigated. This alternative may require substantial modifications and may be more expensive than the previous alternative. The problems with stability and non-compliance make this alternative less attractive for the industry. Also, some ship-owners do not see this as a solution to comply with the SECA rules coming in 2015 because 100% BOG will be unsuitable during some operational modes. A third possibility which seems more attractive to ship-owners is to minimise the use of HFO and use as much BOG as possible. The dual fuel combination of low sulphur HFO and BOG would then be used in such a way that the emissions of SO 2 are lower than what they would have been if 0.1% fuel oil were used at berth. If the scheme were set up in such a way that BOG together with minimum amounts of low sulphur HFO is used for the complete port stay (including manoeuvring) the resulting SO 2 emissions could be lowered further. This option would allow for fuel boost with HFO in the case of flame failure. This option requires fewer modifications on most ships and thus lower modification costs. It may not be suitable for older ships. Whether or not this solution is in compliance with the EU regulations depends on whether the equivalence principle outlined is regarded to be in line with paragraph 4c. In addtion, the ships would still need to use HFO after refit. This option would be likely to require a monitoring system with either control of fuel use and sulphur levels in the fuel or the installation of flue gas monitoring systems to monitor the SO 2 emissions. The cost of the latter is in the order of per ship. An option of using exhaust gas scrubbing would not available by the end of the year because the technology has not yet been tested for this application Conclusions There are genuine safety risks associated with switching from HFO to MGO in existing boilers that are supported by evidence gained through contacts with the industry and authorities. There are, however, technical and operational ways to circumvent these risks. Although many boilers can be used with MGO, all boilers should be inspected. The exact number of boilers that need to be modified is uncertain at this point. There are 262 LNG ships with steam propulsion that need to be modified. There are around 2500 tankers that use boilers for steam driven pumps and/or cargo heating - and most are likely to need to be modified. The main safety concern is that a distillate fuel may vaporise in hot boilers with heated pipes. If the boiler system has not been modified, there is in the case of flame failure and improper handling a risk of an explosion. There are technical solutions to solve these issues where fuel piping systems, burners, and flame inspection systems are modified. The control system may have to be modified and new handling routines established. For LNG ships, modifications are required. The costs range from to for small auxiliary boilers, around per ship for tankers and to for LNG tankers. AEA 78

85 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 11 Requirements with regard to maximum sulphur content of heating oil for condensation boilers 11.1 Objective The purpose of this section of the report is to make an inventory of the requirements with regard to the maximum sulphur content of heating oil in condensation boilers for heating purposes. Where relevant, the requirements for different categories of condensation boilers are given Introduction to oil fired condensation boilers Until recent years there were only household heating boilers based on constant temperature and low temperature systems. All boilers used the same standard heating oil. With the introduction of condensing boilers a distinction was introduced between boilers that can use only low sulphur heating oils and other boilers that can use the different heating oil available on the market. Condensation technology is an efficient way to use the heating value of burned natural gas and oil. In low temperature boilers condensation of flue gas water vapour on the heat exchange surfaces must be avoided because of corrosion. Condensation technology recovers the latent heat of the water vapour in the flue gases by condensation. Condensation of water vapour occurs when flue gases are cooled down to temperatures below dew point of the gases. The dew point temperature of flue gases from oil fired boilers is about 47 C. The flue gas temperature from condensation boilers is lower than of conventional boilers resulting in a better use of sensible heat in the flue gases as well. In this way heat efficiency of condensation boilers is about 6 % higher (and CO2 emissions about 6 % less) than for conventional oil fired boilers at nominal capacity. Another advantage of condensation boiler technology is that the water temperature in the heating system can be regulated according to outside temperature - whereas conventional boilers provide a constant water temperature. Especially in spring and autumn when heating demand is much lower than nominal capacity, the efficiency of condensing boilers is much higher than 6% compared to conventional boilers because the firing is modulated [1] Heating oil quality and availability According to Directive 1999/32/EC the content of sulphur in heating oils to be used in EU Members States from 1st of January 2008 should not exceed 0.1% by mass (=1000 mg/kg). Since July 2000 the permitted sulphur content is, at maximum, 0.2%. The sulphur content of the diesel oil used for vehicle fuel has been reduced from 50 mg/kg to >10 mg/kg from 1st of January In this chapter sulphur content is given in mg/kg oil, in the literature sulphur content is sometimes given in ppm (parts per million) while mg/kg is meant. In Germany the availability of low sulphur oil (< 50 mg/kg), originally meant for vehicles, has facilitated the introduction and use of condensation boilers. There is also an intermediate quality oil - sulphur reduced heating oil ( mg/kg) - available on the market which can be used in the EU countries to reduce emissions of sulphur dioxide to the environment: The use of gas oil for heating has been in general decline for many years, inland consumption fell by about 22% over the period 1995 to 2007, most of this was in the residential commercial sector and is the result of switching to natural gas. Overall a continuing decrease in heating oil use is expected as the result of further natural gas substitution and improved equipment efficiency (i.e. replacement of old heating systems) but may be sustained in some markets i.e. housing developments outside cities (beyond the gas distribution system) [6] AEA 79

86 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 According to a Concawe study [7] the weighted average sulphur content in the gas oil (automotive diesel fuel, heating oil, industrial gas oil and marine diesel fuel) used in the EU has fallen from 1100 mg/kg in 1998 to 900 mg/kg in Average weighted sulphur content of gas oil varies with European region; in 2002 average content was highest (1150 mg/kg) in the Mediterranean region and lowest (470 mg/kg) in the former EFTA countries (Norway, Sweden, Finland, Austria and Switzerland) compared to Northwest Europe and the Atlantic Construction of condensation boilers In most conventional low temperature boilers hot flue gases flow horizontally through the heat changer and condensation must be avoided. In condensation boilers hot flue gases flow downwards through the heat exchanger, water is allowed to condense, this also drains downwards providing a self cleaning effect that prevents the accumulation of salts on the heat exchange surfaces. Depending on the amount of sulphur in the oil burned, and the acids produced as a result, the water can be acidic, with a ph between 1.5 and 3.5 [3]; standard heating oil produces a condensate that can be very corrosive unless suitable construction materials have been used. Extensive experience has shown that heat exchangers made of stainless steel are most suitable but specific aluminium alloys can be an acceptable alternative. A variety of stainless steel alloys are available (with e.g. chromium, nickel, molybdenum, titanium) that are durable and which do not require additional surface treatment Condensation boilers and sulphur content of heating oil Traditional heating oil (EL quality) contains up to 2000 mg/kg sulphur, its flue gas condensate contains significant amounts of sulphurous and sulphuric acids and, if produced in large amounts, should be neutralized before discharge to the sewer system; neutralisation is not required if low sulphur oil is used. Neither can standard heating oil be used in all condensation boilers on the market - some producers only allow the use of low sulphur heating oil. Condensation boilers for standard heating oil are specifically constructed for this duty and tend to have lower heat efficiencies [4]. Originally producers of condensation boilers in Germany were reluctant to specify the use of low sulphur oil but concerns were allayed following extensive tests [5]. A distinction should be made, irrespective of whether or not heating oil of differing sulphur content can be used, between three categories of condensation boilers: - air-to-water condensation boilers o boilers in which flue gas cooling and condensation both take place on the heat exchange surface integrated in the boiler; o boilers that cool flue gases to the dew point using an internal heat exchanger and where condensation then occurs on separate heat exchanger. - air-to-air condensation boilers o boilers in which the condensation of the flue gas system is used to preheat incoming combustion air. Air-to-water condensation boilers in which flue gas cooling and condensation takes place on a heat exchange surface which is integrated in the boiler This category of condensation boilers has been on the market for a long time. The heat exchangers are exposed to the high temperatures of the flame and are susceptible to deposition of the substances resulting from burning of high sulphur oils. Heat exchangers in this category have to be made of corrosion resistant material like stainless steel. According to [3] also some specific formulated alloys of aluminium can be used. To prevent deposition and maintain high heat efficiency, if cleaned only once a year, these condensation boilers ideally need low sulphur heating oil (< 50 mg/kg) [1] but can tolerate sulphur reduced heating oil (< 500 mg/kg) [2]. The use of high sulphur heating oil (1000 mg/kg) in this type of condensation boilers results in faster deposition and decrease of heat efficiency but this can be avoided by cleaning the heat exchanger surface at least twice a year. AEA 80

87 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Air-to-water condensation boilers that cool flue gases to the dew point using an internal heat exchanger and where condensation occurs on separate heat exchanger. High sulphur heating oil ( mg/kg sulphur) can be used in this category of condensation boilers because condensation does not take place in the first (internal) heat exchanger [1; 2]. Combustion residues (containing reaction products of sulphur) deposit on the heat exchanger in the combustion chamber but because there is no condensation at this point there is no corrosion of the surfaces. Hence the heat exchanger can be made of standard steel. In the second (add-on) heat exchanger, where the condensation occurs, is made of stainless steel and there is little deposition. Because high or medium sulphur heating oil may be used for this category of condensation boilers the condensate can be aggressive and neutralisation of the condensate may be required before discharge - depending on heating capacity of the boiler and wastewater quality regulations. Air-to-air condensation boilers in which condensation takes place in the flue gas system to preheat incoming combustion air. In these systems the internal heat exchanger and water flow are designed in such a way that condensation does not take place. Flue gases reach the second heat exchanger at temperatures far above dew point (about 100 C). Condensation in this second heat exchanger is then governed by control of flue gas and combustion air flow. Heating efficiency of this category of condensation boilers is lower than for the air-to-water condensation boilers. As a summary the suitability of different types of heating oil in different categories of condensation boilers is given in the table below. Table 11.1 Suitability of heating oils in different categories of condensation boilers Boiler integrated heat exchanger Add-on heat exchanger High sulphur heating oil Problematic, high deposits on (1000 mg/kg) heat exchange surface Allowed, moderate deposits Sulphur reduced heating oil ( mg/kg) Allowed, moderate deposits Allowed, minor deposits Low sulphur heating oil (< 50 mg/kg) Allowed, minor deposits Allowed, no deposits 11.6 Conclusions Heating oils with sulphur content up to the maximum allowed concentration of 1000 mg/kg can be used in condensation boilers. Deposits on heat exchanger surfaces caused by use of high sulphur heating oils are more problematic in condensation boilers with integrated heat exchanger than in boilers with an add-on condensation heat exchanger. To maintain high efficiency integrated heat exchangers have to be cleaned more frequently than add-on heat exchangers and more frequently than once a year as is usual. Efficiency will be kept at a high level for both configurations if low sulphur oil is used. There are no technical obstacles which prevent the use of low sulphur heating oil in condensation boilers designed for standard heating oil. AEA 81

88 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 12 References: Section 4 Amman et al, National Emission Ceilings for 2020 based on the 2008 Climate and Energy Package. NEC Scenario Analysis report, nr 6. IIASA. Chiffi, Fiorello, Schrooten, De Vlieger, EX-TREMIS - Exploring non road Transport Emissions in Europe --Annex to Final Report. IPTS - Institute for Prospective Technological Studies, DG-JRC, Seville, Spain ( Cofala et al., Analysis of Policy Measures to Reduce Ship Emissions in the Context of the Revision of the National Emission Ceilings Directive. IIASA/MET.NO/ENTEC UK Ltd. Roemer M. and others (2003). Ozone trends according to ten dispersion models. Eurotrac-2 ISS, GSF-National Research Centre for Environment and Health, Munich, Germany. Roemer M., T. van Harmelen and P. Builtjes (2006). Establishment of optimal control areas for acidification, eutrophication and ground level ozone. TNO report 2006-A-R0251/B, TNO, Utrecht, The Netherlands. Schaap M., M. Roemer, F. Sauter, G. Boesen, R. Timmermans and P. Builtjes (2005). LOTOS- EUROS: documentation. TNO report B&O-A R2005/297, TNO, Utrecht, The Netherlands. Van Harmelen T., M. Roemer, J. Pesik and G. Boersen (2007). Analysis of the potential costs savings and environmental and health implications of emission trading for sulphur dioxide and nitrogen dioxide. TNO report 2007-A-R1256/B, TNO, Utrecht, The Netherlands. Van Loon M.,M. Roemer and P. Builtjes (2004). Model intercomparison in the framework of the review of the Unified EMEP model. TNO report R2004/282, TNO, Utrecht, The Netherlands. Section 7 Purvin&Gertz, 2009, Avis M.J., Birch C.H., IMPACTS ON THE EU REFINING INDUSTRY & MARKETS OF IMO SPECIFICATION CHANGES & OTHER MEASURES TO REDUCE THE SULPHUR CONTENT OF CERTAIN FUELS Entec, 2005, Service contract on ship emissions: Assignment, Abatement and Market-based instruments UK-Maritime and Coastgard Agency, 2009, Impact assessment of merchant Shipping (prevention of Air pollution from ships (Amendment)) Regulations 2009 US-EPA, 2009, Proposal to Designate an Emission Control Area for Nitrogen Oxides, Sulfur Oxides and Particulate Matter - Technical Support Document - Chapter 5 Costs, Assessment and Standards Division Office of Transportation and Air Quality U.S. Environmental Protection Agency Krystallion, Sea Water Scrubbing Does it Contribute to increased Global CO2 Emissions AEA 82

89 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Wärtsilä, 2009, General description of Wärtsilä scrubber system Marine and Energy Consulting limited, 2009, presentation at 2nd Bunker Fuel Management, 16th & 17th September, London, UK Delft Greenhouse Gas Emissions for Shipping and Implementation Guidance for the Marine Fuel Sulphur Directive Chiffi, Fiorello, Schrooten, De Vlieger, EX-TREMIS - Exploring non road Transport Emissions in Europe --Annex to Final Report. IPTS - Institute for Prospective Technological Studies, DG-JRC, Seville, Spain. Section 9 Winnes, H and Fridell, E, Particle Emissions from Ships: Dependence on Fuel Type, J. Air & Waste Manage. Assoc. 59:1391 (2009). V. Eyring, H. W. Köhler, J. van Aardenne and A. Lauer, Emissions from international shipping: 1. The last 50 years. J. Geophys. Res. 110, D17305 (2005). Skema 2009: Impact Study of the future requirements of Annex VI of the MARPOL Convention on Short Sea Shipping, Skema MEPC 59/WP.10, 16 July 2009 US EPA, Quality criteria for water. United States Environmental Protection Agency, Washington D.C , The Red book, July Polycyclic aromatic hydrocarbons (PAH), KEMI Swedish Chemicals Agency, Phenanthrene CAS Number , United States Environmental Protection Agency Billiard, S. M., Meyer, J. N., Wassenberg, D. M., Hodson, P. V., Di Giulio, R. T Nonadditive effects of PAHs on early vertebrate development: Mechanisms and implications for risk assessment. Toxicol. Sci. 105: Behrends, B., & Liebezeit, G., 2003, Reducing SO X and NO X Emissions from Ships by a Seawater Scrubber. Research Centre Terramare, Wilhelmshaven, Germany. Hufnagel, M., Liebezeit, G., & Behrends, B., Effects of Sea Water Scrubbing, Final report. Buhaug, Ø., Fløgstad, H., Bakke T., MARULS WP3: Wash-water Criteria for seawater exhaust gas-so X scrubbers. MARINTEK report, for Norwegian Marine Technology. Karle, I-M. and Turner, D., Seawater Scrubbing - reduction of SO X emissions from ship exhausts. Department of Chemistry, Göteborg University, ISBN: MEPC 59/4/19. Marine Environment Protection Committee, 59 th session Agenda item 4. Prevention of air pollution from ships. Advice by GESAMP on the interim criteria for discharge of wash-water from Exhaust Gas Cleaning Systems for removal of sulphur-oxides. Whall et al., Quantification of emissions from ships associated with ship movements between ports in the European Community, Entec AEA 83

90 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Cooper, D. A., Peterson, K., and Simpson, D., 1996, Hydrocarbon, PAH and PCB emissions from ferries: a case study in the Skagerak-Kattegatt-Öresund region, Athmospheric Environment 30, MEPC 56/INF.5, ANNEX 1, Page 22 Directive 2008/105/EC of the European Parliament and of the Council of 16 December 2008 on environmental quality standards in the field of water policy, amending and subsequently repealing Council Directives 82/176/EEC, 83/513/EEC, 84/156/EEC, 84/491/EEC, 86/280/EEC and amending Directive 2000/60/EC of the European Parliament and of the Council. Jalkanen, J. -P., Brink A., Kalli, J., Pettersson, H., Kukkonen, J. and Stipa, T., 2009, A modelling system for the exhaust, emissions of marine traffic and its application in the Baltic Sea area, Atmos. Chem. Phys. Discuss., 9, Benedictow et al., Transboundary Acidification,Eutrophication and Ground Level Ozone in Europe in 2007, EMEP Status Report. Johnsen et al., (2006) Part C of Delft report Greenhouse Gas Emissions for Shipping and Implementation Guidance for the Marine Fuel Sulphur Directive, DNV. Mellqvist, J, Berg, N. and Ohlsson, D., Remote surveillance Remote surveillance of the sulfur content and NO x emissions of ships, Presentation at Haqcc Section 10 Aalborg Aalborg Industries. Aalborg Solutions No 12 January British Department of Transport Explosion in boiler furnaces, Merchant Shipping Notice No. M. 1083, August CARB California Environmental Protection Agency Air Resource Board, Fuel sulfur and other operational requirements for ocean-going vessels within California waters and 24 nautical miles of the California baseline, June DNV DNV, Regulations for the prevention of air pollution from ships, technical and operational implications. February Entec C Whall et al., Quantification of emissions from ships associated with ship movements between ports in the European Community, Entec, European Commission, IMO subcommittee on bulk liquids, Review of MARPOL VI and the NO x technical code December, US coast guard Maritime Safety Alert, Avoiding propulsion loss from fuel switching: American Petroleum Institute Technical Considerations. May Section 11 Condensation technology Condensation technology for energy saving and a clean environment.documentation Viessmann ( [1] AEA 84

91 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Rogatty, W.: Modern oil-heating technology. IKZ-Haustechnik, Volume 8/2004, page 60 [2]. Brennwerttechnik. : [3] Frequently asked questions and answers in relation to the common declaration of the National Government and the Mineral oil business on January 15th Bundesministerium für Umwelt, Naturschutz und Reaktorsicherheit. ( [4] Geräteeignung und Verfügbarkeit. Institut für wirtschaftliche Oelheizung e.v. [5] Impacts on the EU refining industry & markets of IMO specification changes & other measures to reduce the sulphur content of certain fuels. Purvin & Gertz Inc., 30 June 2009 [6] Post, L.: Sulphur dioxide emissions from oil refineries and combustion of oil products in Western Europe and Hungary (2000). CONCAWE report no. 2/07, March 2007 [7] Appendix 1 Cooper and Gustafsson, 2004 D. Cooper and T. Gustafsson, Methodology for Calculating Emissions from Ships. 1. Update of Emission Factors Report Series for SMED and SMED&SLU, SMHI, Norrköping, Sweden (2004) ISSN: IPCC 1995, Greenhouse gas inventory Reference manual. Intergovernmental Panel on Climate Change. Lloyds s Register, Marine Exhaust Emissions Research Programme. Lloyds s Register Engineering Services, London. TFEIP, EMEP/EEA Emission Inventory Guidebook, 2008 version, Task Force for Emission Inventories and Projections (TFEIP). Available via AEA 85

92 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 13 Appendix 1 - Emission factors used Table 1: Unabated emission factors used in the baseline scenario kg/ton kg/gj source notes heavy fuel oil SO 2-SECA TFEIP % S, after 2006 in North Sea and Baltic Sea in baseline SO TFEIP ,70% S, pre-2006 NO x Ex-tremis NMVOC Ex-tremis CO IVL and Lloyds PM Ex-tremis diesel oil SO TFEIP % S NO x Ex-tremis NMVOC Ex-tremis CO IVL and Lloyds PM Ex-tremis Note: Since , for seagoing ships on route between member states, the sulphur content of diesel fuel has been lowered to 0.1%. Data were insufficient to make this distinction in the emission calculation. Table 2: Unabated emission factors by fuel and motor speed CO Fuel consumption NO x NMVOC g fuel/kwh kg/ton kg/gj kg/ton kg/gj kg/ton kg/gj GT-MG/DO GT-RO HSD- MG/DO HSD-RO MSD- MG/DO MSD-RO SSD- MG/DO SSD-RO ST-MG/DO ST-RO Note: ] GT-MG/DO: gas turbine - marine diesel oil/marine gasoil GT-RO: gas turbine bunker fuel oil HSD-MG/DO: high speed diesel - marine diesel oil/marine gasoil HSD-RO: high speed diesel bunker fuel oil MSD-MG/DO: medium speed diesel - marine diesel oil/marine gasoil MSD-RO: medium speed diesel bunker fuel oil SSD-MG/DO: slow speed diesel - marine diesel oil/marine gasoil SSD-RO: slow speed diesel bunker fuel oil ST-MG/DO: steam turbines - marine diesel oil/marine gasoil ST-RO: steam turbines bunker fuel oil Source: TFEIP A.3.d Table 3-5, and (for CO) IVL and Lloyds (for Diesel engines) and Cooper and Gustafsson, 2004 for GT and ST. AEA 86

93 Cost benefit analysis to support the impact assessment AEA/ED /Issue 3 Table 3: Unabated PM emission factors by fuel and motor speed and sulphur content of the fuel Sulphur content 0.10% 0.20% 0.50% 1.50% 2.70% kg/ton kg/gj kg/ton kg/gj kg/ton kg/gj kg/ton kg/gj kg/ton kg/gj GT-MG/DO GT-RO HSD- MG/DO HSD-RO MSD- MG/DO MSD-RO SSD- MG/DO SSD-RO ST-MG/DO ST-RO Note: see table 2 for abbreviations; Table 4: SO 2 emission factors in the scenarios Black Mediter- North Remaining Outside SECAs Within SECAs Baltic Sea ranean Sea NE-atl Scenar S content EF S content EF IN IN IN IN IN SECA? io (mass%) (kg/gj) (mass%) (kg/gj) SECA? SECA? SECA? SECA? % % 0.05 Y N N Y N % % 0.05 Y N Y Y N % % 0.05 Y Y Y Y N % % 0.05 Y N N Y N % % 0.05 Y Y Y Y N BASE LINE 2.70% % 0.72 Y N N Y N AEA 87

94 Table 5: NO x emission factors in scenarios Tier I Tier II Tier III kg/gj ships constructed between and ships constructed after for 2020 only: ships constructed after GT-MG/DO GT-RO HSD-MG/DO HSD-RO MSD-MG/DO MSD-RO SSD-MG/DO SSD-RO ST-MG/DO ST-RO Note: see table 2 for abbreviations Table 6: Applicability of the Tiers in the various scenarios Scenario 1 Tier I&II, for 2020 Tier III 2 Tier I& II 3 Tier I&II, for 2020 Tier III 4 Tier I,II & III (2020 only) 5 Tier I,II & III (2020 only) Background on NO x limits to marine diesel engines and Emission Control Areas (ECAs) NO X emissions are regulated under the International Maritime Organization (IMO), in the International Convention for the Prevention of Pollution From Ships, 1973 as modified by the Protocol of 1978 (in short MARPOL 73/78). NOx limits for marine diesel engines are set in Annex VI of MARPOL and the revised NOx Technical Code. In October 2008, the IMO s Marine Environmental Protection Committee (MEPC) adopted both the revised Annex VI of MARPOL and the revised NOx Technical Code. The annex and the code will enter into force 1 July The new revision substantially tightens the NOx and Sulphur limits compared to the existing annex, and also includes requirements governing NOx emissions from ships constructed from 1 January 1990 to 1 January New NO x limits The revised NOx regulations contain a three-tier approach as follows: Tier I (identical to today s limits) For diesel engines installed on ships constructed from 1 January 2000 to 1 January 2011, the allowable NOx emissions are: AEA 88

95 17.0 g/kwh when n is less than 130 rpm 45.0*n(-0.2) g/kwh when n is 130 rpm or more but less than 2,000 rpm 9.8 g/kwh when n is 2,000 rpm or more Tier II For diesel engines installed on ships constructed on or after 1 January 2011, the allowable NOx emissions are: 14.4 g/kwh when n is less than 130 rpm 44.0*n(-0.23) g/kwh when n is 130 rpm or more but less than 2,000 rpm 7.7 g/kwh when n is 2,000 rpm or more Tier III Ships constructed on or after 1 January 2016 will have additional limitations when operating in an Emission Control Area (ECA). No ECAs have yet been designated for NO x emissions, but it is expected that both the Baltic Sea and North Sea will be designated as NO x ECAs well ahead of 1 January The allowable NOx emissions are: 3.4 g/kwh when n is less than 130 rpm 9.0*n(-0.2) g/kwh when n is 130 rpm or more but less than 2,000 rpm 1.96 g/kwh when n is 2,000 rpm or more AEA 89

96 14 Appendix 2 - Emission Figures by Scenario and Area Baseline Emission Figures by Scenario and Area Baseline Baltic Sea Black Sea Remaining NE Atlantic ocean x 1000 tonnes SO NMVOC CO PM NO x Mediterranean Sea North Sea X 1000 tonnes SO NMVOC CO PM NO x SEA1 Baltic Sea Black Sea Remaining NE Atlantic ocean Mediterranean Sea North Sea x 1000 tonnes SO NMVOC CO PM NO x ,018 2,055 1, AEA 90

97 SEA2 Baltic Sea Black Sea Remaining NE Atlantic ocean Mediterranean Sea North Sea x 1000 tonnes SO NMVOC CO PM NO x ,018 2,055 2, SEA3 Baltic Sea Black Sea Remaining NE Atlantic ocean Mediterranean Sea North Sea x 1000 tonnes SO NMVOC CO PM NO x ,018 2,055 1, SEA4 Baltic Sea Black Sea Remaining NE Atlantic ocean Mediterranean Sea North Sea x 1000 tonnes SO NMVOC CO PM NO x ,018 2,055 1, SEA5 Baltic Sea Black Sea Remaining NE Atlantic ocean Mediterranean Sea North Sea x 1000 tonnes SO NMVOC CO PM NO x ,018 2,055 1, AEA 91

98 Emissions land based sources, baseline and scenarios Baseline x 1000 tonnes SO 2 16,726 10,324 9,897 9,734 NMVOC 16,236 13,033 12,374 11,931 PM10 4,726 4,013 3,988 3,988 PM2.5 3,231 2,768 2,716 2,678 NO x 16,582 13,839 12,691 11,312 Note: exclusive of emissions related to national maritime fuel consumption (e.g. fisheries), which are included in the GAINS-totals Source: GAINS Europe, NEC6, C&E package current policy Land based A x 1000 tonnes SO 2 9, NMVOC 12, PM10 3,988 3,988 PM2.5 2,716 2,678 NO x 12,691 11,312 Land based B x 1000 tonnes SO 2 9,806 9,665 NMVOC 12,374 11,931 PM10 3,988 3,988 PM2.5 2,716 2,678 NO x 12,691 11,312 AEA 92

99 15 Appendix 3 - Relevant geographical representations of the dispersion calculations In this appendix the most relevant results of the dispersion calculations are presented. First some background information is given on the relevant background concentrations used in the dispersion calculations and the effect of inter-annual changes due to the meteorology. As states earlier we used the meteorology of 1997 given the fact that this is considered a representative year for the European meteorology. The absolute values of the calculated pollutant concentration are certainly relevant but in this cost benefit analysis the emphasis is on the differences between the results for the different scenarios. Sensitivity of results to inter-annual variability in meteorology and background ozone concentrations. Background ozone and changes therein are important to metrics such as AOT40, SOMO35 etc. The increase of hemispheric ozone is under scientific debate and it is not easy to answer how much more of ozone we can expect in Because the effect of IMO regulations on NO x emissions in the remaining North Atlantic beyond 10 W is not quantified we do not know how this will affect background ozone concentrations. In general terms one can state; with each 1ug/m 3 increase of the background one can expect at the fringe of Europe, as a rough maximum estimate, an increase of 2500 ug/m 3.h in AOT40f, simply being the number of associated hours in the AOT40f definition that will be increased with 1 ug/m 3, plus a few hours that in addition will exceed the threshold. At the interior of Europe this will be less, since part (say 70%) of the extra ozone will be consumed by dry deposition, leaving an increase as a rule of thumb of 800 ug/m 3.h. Suppose the hemispheric increase of ozone between 2010 and 2020 will be 4 ug/m3, then the AOT40f in continental Europe will increase by about 3200 ug/m3.h which is an increase of 10% compared to 2010 levels. Based on Laurila et al., 2004 we find, following their lowest SRES scenario, an increase of AOT40f due to hemispheric changes of 2000 ug/m3.h from 2010 to 2020 on top of an AOT40f in 2010 of about ug/m3.h. Uncertainty estimates in changes in hemispheric ozone are hard to give since the cause of the changes is poorly understood; given this lack of knowledge we estimate an uncertainty of 100%. The inter-annual variability in AOT40f can be as large as 8000 ug/m3.h fluctuation from one year to another, but on average the fluctuation is typical in the other of 3000 ug/m3.h. For continental sites as in Switzerland and Austria the aforementioned fluctuations are 30% and 12% of the long term average of AOT40f in these countries. In the UK of the Netherlands (with much lower longterm baseline AOT40f values) the fluctuations in relative terms are much larger, typically in the order of %. More important are the inter-annual fluctuations of the scenarios. In absolute terms (change in ug/m3.h) the fluctuations will be considerable, but once expressed in terms of fractional changes with respect to the baseline AOT40f the inter-annual fluctuations are much smaller, and more robust against the choice of meteorological year. For other ozone metrics that are derived from larger portions of the ozone distribution, such as SOMO35 (annual sum of daily max eight hour running mean of ozone exceeding 35 ppb) and annual mean the fluctuations from year to year are smaller than for AOT40. AEA 93

100 The atmosphere is never in equilibrium. The lifetime of NO x is short but due its effect on ozone (which has a longer lifetime) still in the order of weeks to months. However, I do not see any transient effects in the range of years due to NO x emission changes. The interannual variation in PM concentrations is typically in the order of 10%. Again, scenario changes expressed with respect to the baseline are less sensitive to interannual variability. Based on the EMEP status report 2005/1 in which the exceedances of critical loads for either acidification or eutrophication in 2002 and 2003 were compared (with 2003 being an extreme year) the relative differences amount to about 20% from one year to another Does the calculation grid size have an effect on the calculated benefits? A smaller grid means a smaller grid for the dispersion calculations and for the ecosystems/population data bases. Assuming that the concentration (deposition) in a 50x50 km grid cell differs only slightly from the averaged concentration (deposition) of all associated 10x10 grid cells, then we can expect differences in the choice of grid cell only if the set 10x10 km concentration data within 50x50 km bears some correlation with the set of 10x10 km ecosystem (population) data. This will be specified for the themes separately. Acidification. / Eutrophication We have no information on the distribution of sensitive ecosystems (with regard to acidification or eutrophication) at scales smaller than 50x50 km. Therefore, the question cannot be addressed. SOMO35 (annual sum of daily max eighth hour running mean of ozone exceeding 35 ppb. ) In this study the SOMO35 values are multiplied with population numbers. In rural areas with low NO x concentrations grid cells within a 50x50 km grid cell will not differ much from the value in a 50x50 km cell. This part won t be affected by working on smaller scales. In urban areas, characterised by high levels of NO x ozone concentrations (SOMO35) vary in general from relatively low values in the city centres (high NO x levels, densely populated) to relatively high values in residential suburban areas (lower NO x, less populated). The effect is to put more weight on the low values, and consequently, a 10x10 km grid in an urban area results in somewhat lower SOMO35*inhabitant s values than a corresponding 50x50 km would. PM10 For PM10 similar considerations apply as for SOMO35, although with a different sign. In rural areas, there is no effect expected from reducing the grid size. In urban areas the primary PM10 emission (assuming it is proportional to population numbers) correlates with populations and, as a result, so does the PM10 concentration (although the effect is smaller due to the damping effect of the smoother sulphate and nitrate fields). The effect is that in urban areas the PM10*inhabitants values are larger for a smaller grid than for a larger grid. AEA 94

101 AEA 95

102 Figure A1: Average Accumulated Exceedence for acidification (eq./ha/yr) for a) Baseline In all areas, not colored green, there is an exceedence of the critical load b) Fractional change (%) of SEA1 compared to base line The map shows clearly the reduction of the exceedence of the critical load all round the North Sea and the Baltic due the designation of these waters as SECA. c) Fractional change (%) of SEA 2 compared to SEA 1 Thus the effect of inclusion of the Mediterranean sea and Black sea in the SECA does add only to a small extend to the improvement gained in scenario SEA1 compared to the baseline. d) Fractional change (%) of SEA 3 compared to SEA 3 Reduction of the S content outside the SECAs does lower the exceedence of the critical load in the UK and IR. See also table 7 in appendix 4 AEA 96

103 AEA 97

104 . Figure A2 Average Accumulated Exceedence for acidification (eq./ha/yr) for a) Baseline In all areas, not colored green, there is an exceedence of the critical load b) Fractional change (%) of Bbase compared to Baseline The map shows the positive effect on the exceedence of the critical load for acidification when lowering the S content of fuel on land to 0.001%. This map indicates the maximum effect of lowering the S content in land based fuel. c) Fractional change (%) of BSEA1 compared to SEA 1 The map shows the same picture as map b), the minor differences are caused by the fact that in these scenario s the North Sea /British channel and the Baltic are already designated as SECA and thus exceedences were already reduced. The general conclusion of these maps is that the effect on the exceedence of the critical load for acidification when lowering the S content of fuel on land is marginal approximately 1.3 to 2.8 % reduction. Because acidification is the parameter effected the most by sulfur emissions this analysis is not performed for the other parameters See also table 7 and 9 in appendix 4 AEA 98

105 AEA 99

106 . Figure A3: Average Accumulated Exceedence for eutrophication (eq./ha/yr) for a) Baseline In all areas, not colored green, there is an exceedence of the critical load b) Fractional change (%) of SEA1 compared to baseline The map shows the reduction of the exceedence of the critical load as a result of the NOx Tier I & II standards in all European seas. The designation of the North Sea and the Baltic des not have an significant impact on the exceedence. Which can be seen in the next map c) Fractional change (%) of SEA2 compared to SEA1 Thus the effect of inclusion of the Mediterranean sea and Black sea in the does not add significant to the improvement gained in scenario SEA1 compared to the baseline. d) Fractional change (%) of SEA3 compared to SEA2 Extension of the SECA with the Black Sea does lower the exceedence of the critical load only to a very small extend. See also table 10 en 12 in appendix 4 AEA 100

107 AEA 101

108 . Figure A4 AOT40f (ppm.h) for a) Baseline b) Difference (ppm.h) of SEA1 compared to baseline The map shows clearly the reduction of the ozone concentration as an result of the reduced NOx emissions in all seas as a result of the Tier standards. The effect in the Mediterranean is enhanced by the longer periods of sunshine. The maps c and d show a similar picture but there are small differences. These are elaborated in table 22 in appendix 4. c) Difference (ppm.h) of SEA2 compared to baseline d) Difference (ppm.h) of SEA3 compared to baseline. See also table 22 and 24 in appendix 4 AEA 102

109 AEA 103

110 ... Figure A5 SOMO35 (ppm.h) for a) Baseline b) Difference (ppm.h) of SEA1 compared to baseline The map shows clearly the reduction of the ozone concentration as an result of the reduced NOx emissions in all seas as a result of the Tier standards. The effect in the Mediterranean is enhanced by the longer periods of sunshine. The maps c and d show a similar picture but there are small differences. These are elaborated in table 16 in appendix 4. c) Difference (ppm.h) of SEA2 compared to baseline d) Difference (ppm.h) of SEA3 compared to baseline. See also table 16 and 18 in appendix 4 AEA 104

111 AEA 105

112 . Figure A6 PM10 (ug/m 3 ) for 2015 a) Baseline b) SEA 1 The map shows the positive effect on PM concentration when designating the North Sea / English Channel and The Baltic as SECA. The reduction of sulfur in fuels is associated with an reduction in PM emissions. In the coastal zones all round the SECA show reduced concentrations of PM c) SEA 2 The map shows the positive effect when designating the Baltic Sea, North Sea / English Channel, the Mediterranean Sea and the Black Sea as SECA. Although not clearly indicated in the map the differences with the former map is that the concentrations in the Mediterranean coastal zones are reduced. See also table 13 and 15 in appendix 4 d) SEA 3 Extension of the SECA with the Black Sea does lower the concentration in this area.. AEA 106

113 AEA 107

114 . Figure A7 PM10 (ug/m 3 ) for 2015 a) Baseline b) Fractional change (%) of SEA1 compared to baseline The next map shows the positive effect on PM concentration when designating the North Sea / English Channel and The Baltic as SECA. The reduction of sulfur in fuels is associated with an reduction in PM emissions. c) Fractional change (%) of SEA2 compared to baseline The map shows the positive effect (% reduction in concentration) when designating the Mediterranean sea and the Black Sea as SECA. d) Fractional change (%) of SEA3 compared to baseline Extension of the SECA with the Black Sea reduces the PM concentration in this area. See also table 13 and 15 in appendix 4 AEA 108

115 AEA 109

116 Figure A8: Average Accumulated Exceedence for acidification (eq./ha/yr) for a) Baseline In all areas, not colored green, there is an exceedence of the critical load b) Fractional change (%) of SEA1 compared to baseline The map shows clearly the reduction of the exceedence of the critical load all round the North Sea and the Baltic due the designation of these waters as SECA. c) Fractional change (%) of SEA 3 compared to SEA 1 Thus the effect of inclusion of the Mediterranean sea and Black sea in the SECA does add only to a small extend to the improvement gained in scenario SEA1 compared to the baseline. d) Fractional change (%) of SEA 4 compared to SEA 3 Reduction of the S content outside the SECAs does lower the exceedence of the critical load in the UK and IR. See also table 8 in appendix 4 AEA 110

117 AEA 111

118 . Figure A9 Average Accumulated Exceedence for acidification (eq./ha/yr) for a) Fractional change (%) of ASEA 1 compared to SEA 1 The map shows the small positive effect on the exceedence of the critical load for acidification when lowering the S content of fuel on land to 0.05%. b) Fractional change (%) of BSEA 1 compared to SEA 1 The map shows the positive effect on the exceedence of the critical load for acidification when lowering the S content of fuel on land to 0.001%. There are only small difference to map a especially in those grid cells with high population densities. c) Fractional change (%) of ASEA 4 compared to SEA 4 The map shows the same picture as map a), the only difference is that in these scenario s the sulphur content outside the SECA is reduced to 0.5 %. d) Fractional change (%) of BSEA 4 compared to SEA 4 The map shows the same picture as map b), the only difference is that in these scenario s the sulphur content outside the SECA is reduced to 0.5 ). The general conclusion of these maps is that the effect on the exceedence of the critical load for acidification when lowering the S content of fuel on land is marginal approximately 1.2 to 2.1 % reduction. Because acidification is the parameter effected the most by sulfur emissions this analysis is not performed for the other parameters See also table 9 in appendix 4 AEA 112

119 AEA 113

120 . Figure A10: Average Accumulated Exceedence for eutrophication (eq./ha/yr) for a) Baseline In all areas, not colored green, there is an exceedence of the critical load b) fractional change (%) of SEA1 compared to baseline The map shows clearly the reduction of the exceedence of the critical load all round the North Sea and the Baltic due the designation of these waters as SECA. Also the exceedence in the Mediterranean countries improves as a result of the NOX Tier standards. c) fractional change (%) of SEA 3 compared to SEA 1 Thus the effect of inclusion of the Mediterranean sea and Black sea in the SECA and designation these waters as NECA does add to the improvement gained in scenario SEA1 compared to the baseline. d) fractional change (%) of SEA 4 compared to SEA 3 Reduction of the S content outside the SECAs does lower the exceedence of the critical load only to a small extend. The remainder reduction in SEA4 stems from the NECA in the Mediterranean and the Black Sea See also table 11 in appendix 4 AEA 114

121 AEA 115

122 . Figure A11 PM10 (ug/m 3 ) for a) Baseline b) SEA 1 The map shows the positive effect on PM concentration when designating the designating the North Sea / English Channel and The Baltic as SECA. The reduction of sulfur in fuels is associated with an reduction in PM emissions. In the coastal zones all round the SECA show reduced concentrations of PM c) SEA 3 The map shows the positive effect when designating the Baltic Sea, North Sea / English Channel, the Mediterranean Sea and the Black Sea as SECA. Although not clearly indicated in the map the differences with the former map is that the concentrations in the Mediterranean coastal zones are reduced. See also table 14 in appendix 4 d) SEA 5 The map shows that reducing the S content outside the SECA has a positive effect on PM10 particularly in the UK and IR (compare to map b). For more details see also table 14 in appendix 4 AEA 116

123 AEA 117

124 . Figure A12 PM10 (ug/m 3 ) for a) Baseline b) fractional change (%) of SEA 1 compared to baseline The next map shows the positive effect on PM concentration when designating the designating the North Sea / English Channel and The Baltic as SECA. The reduction of sulfur in fuels is associated with an reduction in PM emissions.. c) fractional change (%) of SEA 3 compared to baseline The map shows the positive effect (% reduction in concentration) when designating the Mediterranean sea and the Black Sea as SECA and the Baltic Sea, North Sea / English Channel and the Mediterranean Sea as NECA. d) fractional change (%) of SEA 4 compared to baseline The map shows that reducing the S content outside the SECA (Mediterranean Sea and Black Sea are in this scenario not SECA) also has a positive effect on PM10 particularly in the UK and IR (compare to map b). See also table 14 in appendix 4 AEA 118

125 AEA 119

126 . Figure A13 AOT40F (ppm.h) for a) Difference (ppm.h) of ASEA 1 compared to SEA 1 ; b) difference (ppm.h) of BSEA 1 compared to SEA 1 The map shows clearly the reduction of the ozone concentration as an result of the reduced NOx emissions in all seas as a result of the Tier standards. c) difference (ppm.h) of ASEA 4 compared to SEA 4 Thus the effect of inclusion of the Mediterranean Sea and Black sea in the SECA does not add to the improvement gained in scenario SEA1 compared to the baseline. d) difference (ppm.h) of BSEA 4 compared to SEA 4 Reduction of the S content outside the SECAs does not add to a reduction of ozone compared to SEA1. SEA1 is comparable to SEA4. Al the maps show little differences as one could expect because the differences shown in the maps are the result of lowering the S content in fuel used on land. Because the differences are that small please see also 24 and 27 in appendix 4 AEA 120

127 AEA 121

128 . Figure A14 AOT40F (ppm.h) for a) Baseline b) Difference (ppm.h) of SEA1 compared to baseline The map shows clearly the reduction of the ozone concentration as an result of the reduced NOx emissions in all seas as a result of the Tier standards. c) Difference (ppm.h) of SEA3 compared to baseline Thus the effect of inclusion of the Mediterranean Sea and Black sea in the SECA does not add to the improvement gained in scenario SEA1 compared to the baseline. d) Difference (ppm.h) of SEA4 compared to baseline Reduction of the S content outside the SECAs does not add to a reduction of ozone compared to SEA1. SEA1 is comparable to SEA4. See also table 23 in appendix 4 AEA 122

129 AEA 123

130 . Figure A15 SOMO35 (ppm.h) for a) Baseline b) Difference (ppm.h) of SEA 1 compared to baseline The map shows clearly the reduction of the ozone concentration as an result of the reduced NOx emissions in all seas as a result of the Tier standards. c) Difference (ppm.h) of SEA 3 compared to baseline Thus the effect of inclusion of the Mediterranean sea and Black sea in the SECA does not add to the improvement gained in scenario SEA1 compared to the baseline. d) Difference (ppm.h) of SEA 4 compared to baseline Reduction of the S content outside the SECAs does not add to a reduction of ozone compared to SEA1. SEA1 is comparable to SEA4. See also table 17 in appendix 4 AEA 124

131 16 Appendix 4 - Results per EU member states In this appendix the results of the dispersion modelling are aggregated to the level of the individual countries. In order to make the results more comparable the rough results were multiplied by either the area of the country or the number of inhabitants. For each parameter first the result for 2015 qnd 2020 are presented. A third table shows the absolute and percentage difference between different scenarios and the trend between 2015 and 2020 on the level of total EU Table 1: Sulphur deposition in 2015 (Meq/year) area (km2) base SEA1 SEA2 SEA3 ASEA1 BSEA1 A B AT Austria BE Belgium BG Bulgaria CH Switzerland CY Cyprus CZ Czech republic DE Germany DK Denmark EE Estonia ES Spain FI Finland FR France GR Greece HU Hungary IE Ireland IT Italy LT Lithuania LU Luxembourg LV Latvia MT Malta NL Netherlands NO Norway PO Poland PT Portugal RO Romania SE Sweden SK Slovakia SL Slovenia UK United kingdom EU AEA 125

132 Table 2: Sulphur deposition in 2020 (Meq/year) Area (km2) base SEA1 SEA3 SEA4 SEA5 ASEA1 BSEA1 ASEA4 BSEA4 AT Austria BE Belgium BG Bulgaria CH Switzerland CY Cyprus CZ Czech republic DE Germany DK Denmark EE Estonia ES Spain FI Finland FR France GR Greece HU Hungary IE Ireland IT Italy LT Lithuania LU Luxembourg LV Latvia MT Malta NL Netherlands NO Norway PO Poland PT Portugal RO Romania SE Sweden SK Slovakia SL Slovenia UK United kingdom EU AEA 126

133 Table 3: Differences between scenarios and years for sulphur depositions (EU27+2 total) abs abs % (Meq/yr (Meq/yr) ) % SEA1-base % % SEA2-base % n.c n.c SEA3-base % % SEA4-base n.c n.c % SEA5-base n.c n.c % BSEA1- ASEA % % ASEA1-SEA % % BSEA1-SEA % % A-SEA4- ASEA1 n.c n.c % BSEA4- BSEA1 n.c n.c % A-base % n.c n.c B-base % n.c n.c B-A % n.c n.c SEA2-SEA % n.c n.c SEA3-SEA % n.c n.c SEA3-SEA1 n.c n.c % SEA4-SEA1 n.c n.c % SEA5-SEA3 n.c n.c % SEA5-SEA4 n.c n.c % ASEA1-base % n.c n.c BSEA1-base % n.c n.c n.c - not calculated AEA 127

134 Table 4: Nitrogen deposition in 2015 (Meq/year) area (km2) base SEA1 SEA2 SEA3 ASEA1 BSEA1 A B AT Austria BE Belgium BG Bulgaria CH Switzerland CY Cyprus CZ Czech republic DE Germany DK Denmark EE Estonia ES Spain FI Finland FR France GR Greece HU Hungary IE Ireland IT Italy LT Lithuania LU Luxembourg LV Latvia MT Malta NL Netherlands NO Norway PO Poland PT Portugal RO Romania SE Sweden SK Slovakia SL Slovenia UK United kingdom EU AEA 128

135 Cost benefit analysis to support the impact assessment AEA/ED45756/Issue 2 Table 5: Nitrogen deposition in 2020 (Meq/year) area (km2) base SEA1 SEA3 SEA4 SEA5 ASEA1 BSEA1 ASEA4 BSEA4 AT Austria BE Belgium BG Bulgaria CH Switzerland CY Cyprus CZ Czech republic DE Germany DK Denmark EE Estonia ES Spain FI Finland FR France GR Greece HU Hungary IE Ireland IT Italy LT Lithuania LU Luxembourg LV Latvia MT Malta NL Netherlands NO Norway PO Poland PT Portugal RO Romania SE Sweden SK Slovakia SL Slovenia UK United kingdom EU

136 Table 6: Differences between scenarios and years for nitrogen depositions (EU27+2 total) abs abs % (Meq/yr) (Meq/yr) % SEA1-base % % SEA2-base % n.c n.c SEA3-base % % SEA4-base n.c n.c % SEA5-base n.c n.c % BSEA1- ASEA % % ASEA1-SEA % % BSEA1-SEA % % A-SEA4- ASEA1 n.c n.c % BSEA4- n.c n.c BSEA % A-base % n.c n.c B-base % n.c n.c B-A % n.c n.c SEA2-SEA % n.c n.c SEA3-SEA % n.c n.c SEA3-SEA1 n.c n.c % SEA4-SEA1 n.c n.c % SEA5-SEA3 n.c n.c % SEA5-SEA4 n.c n.c % ASEA1-base % n.c n.c BSEA1-base % n.c n.c n.c - not calculated 130

137 Table 7: Acidification (AAE) in 2015 (Meq/year) area (km2) base SEA1 SEA2 SEA3 ASEA1 BSEA1 A B AT Austria BE Belgium BG Bulgaria CH Switzerland CY Cyprus CZ Czech republic DE Germany DK Denmark EE Estonia ES Spain FI Finland FR France num GR Greece HU Hungary IE Ireland IT Italy LT Lithuania LU Luxembourg LV Latvia MT Malta NL Netherlands NO Norway PO Poland PT Portugal RO Romania SE Sweden SK Slovakia SL Slovenia UK United kingdom EU

138 Table 8: Acidification (AAE) in 2020 (Meq/year) area (km2) base SEA1 SEA3 SEA4 SEA5 ASEA1 BSEA1 ASEA4 BSEA4 AT Austria BE Belgium BG Bulgaria CH Switzerland CY Cyprus CZ Czech republic DE Germany DK Denmark EE Estonia ES Spain FI Finland FR France GR Greece HU Hungary IE Ireland IT Italy LT Lithuania LU Luxembourg LV Latvia MT Malta NL Netherlands NO Norway PO Poland PT Portugal RO Romania SE Sweden SK Slovakia SL Slovenia UK United kingdom EU

139 Table 9: Differences between scenarios and years for acidification (AAE) (EU27+2 total) abs abs % (Meq/yr) (Meq/yr) % SEA1-base % % SEA2-base % n.c n.c SEA3-base % % SEA4-base n.c n.c % SEA5-base n.c n.c % BSEA1- ASEA % % ASEA1-SEA % % BSEA1-SEA % % A-SEA4- ASEA1 n.c n.c % BSEA4- n.c n.c BSEA % A-base % n.c n.c B-base % n.c n.c B-A % n.c n.c SEA2-SEA % n.c n.c SEA3-SEA % n.c n.c SEA3-SEA1 n.c n.c % SEA4-SEA1 n.c n.c % SEA5-SEA3 n.c n.c % SEA5-SEA4 n.c n.c % ASEA1-base % n.c n.c BSEA1-base % n.c n.c n.c - not calculated 133

140 Table 10: Eutrophication (AAE) in 2015 (Meq/year) area (km2) base SEA1 SEA2 SEA3 ASEA1 BSEA1 A B AT Austria BE Belgium BG Bulgaria CH Switzerland CY Cyprus CZ Czech republic DE Germany DK Denmark EE Estonia ES Spain FI Finland FR France GR Greece HU Hungary IE Ireland IT Italy LT Lithuania LU Luxembourg LV Latvia MT Malta NL Netherlands NO Norway PO Poland PT Portugal RO Romania SE Sweden SK Slovakia SL Slovenia UK United kingdom EU

141 Table 11: Eutrophication (AAE) in 2020 (Meq/year) area (km2) base SEA1 SEA3 SEA4 SEA5 ASEA1 BSEA1 ASEA4 BSEA4 AT Austria BE Belgium BG Bulgaria CH Switzerland CY Cyprus CZ Czech republic DE Germany DK Denmark EE Estonia ES Spain FI Finland FR France GR Greece HU Hungary IE Ireland IT Italy LT Lithuania LU Luxembourg LV Latvia MT Malta NL Netherlands NO Norway PO Poland PT Portugal RO Romania SE Sweden SK Slovakia SL Slovenia UK United kingdom EU

142 Table12: Differences between scenarios and years for eutrophication (AAE) (EU27+2 total) abs abs % (Meq/yr) (Meq/yr) % SEA1-base % % SEA2-base % NC NC SEA3-base % % SEA4-base n.c n.c % SEA5-base n.c n.c % BSEA1-ASEA % % ASEA1-SEA % % BSEA1-SEA % % A-SEA4-ASEA1 NC NC % BSEA4-BSEA1 NC NC % A-base % n.c n.c B-base % n.c n.c B-A % n.c n.c SEA2-SEA % n.c n.c SEA3-SEA % n.c n.c SEA3-SEA1 n.c n.c % SEA4-SEA1 n.c n.c % SEA5-SEA3 n.c n.c % SEA5-SEA4 n.c n.c % ASEA1-base % n.c n.c BSEA1-base % n.c n.c n.c - not calculated 136

143 Table 13: Particulate matter inhabitant weighted in 2015 (µg PM10/m3*Minh.) inhabitants base SEA1 SEA2 SEA3 ASEA1 BSEA1 A B AT Austria BE Belgium BG Bulgaria CH Switzerland CY Cyprus CZ Czech republic DE Germany DK Denmark EE Estonia ES Spain FI Finland FR France GR Greece HU Hungary IE Ireland IT Italy LT Lithuania LU Luxembourg LV Latvia MT Malta NL Netherlands NO Norway PO Poland PT Portugal RO Romania SE Sweden SK Slovakia SL Slovenia UK United kingdom EU

144 Table 14: Particulate matter inhabitant weighted in 2020 (µg PM10/m3*Minh.) inhabitants base SEA1 SEA3 SEA4 SEA5 ASEA1 BSEA1 ASEA4 BSEA4 AT Austria BE Belgium BG Bulgaria CH Switzerland CY Cyprus CZ Czech republic DE Germany DK Denmark EE Estonia ES Spain FI Finland FR France GR Greece HU Hungary IE Ireland IT Italy LT Lithuania LU Luxembourg LV Latvia MT Malta NL Netherlands NO Norway PO Poland PT Portugal RO Romania SE Sweden SK Slovakia SL Slovenia UK United kingdom EU

145 Table 15: Differences between scenarios and years for particulate matter ( inhabitant weighted, EU27+2 total) abs (µg PM10/m3*Minh.) % abs (µg PM10/m3*Minh.) SEA1-base % % SEA2-base % NC NC SEA3-base % % SEA4-base n.c n.c % SEA5-base n.c n.c % BSEA1-ASEA % % ASEA1-SEA % % BSEA1-SEA % % A-SEA4-ASEA1 n.c n.c % BSEA4-BSEA1 n.c n.c % A-base % n.c n.c B-base % n.c n.c B-A % n.c n.c SEA2-SEA % n.c n.c SEA3-SEA % n.c n.c SEA3-SEA1 n.c n.c % SEA4-SEA1 n.c n.c % SEA5-SEA3 n.c n.c % SEA5-SEA4 n.c n.c % ASEA1-base % n.c n.c BSEA1-base % n.c n.c n.c - not calculated % 139

146 Table 16: SOMO35 in 2015 (ppm.d.*minh.) inhabitants base SEA1 SEA2 SEA3 ASEA1 BSEA1 A B AT Austria BE Belgium BG Bulgaria CH Switzerland CY Cyprus CZ Czech republic DE Germany DK Denmark EE Estonia ES Spain FI Finland FR France GR Greece HU Hungary IE Ireland IT Italy LT Lithuania LU Luxembourg LV Latvia MT Malta NL Netherlands NO Norway PO Poland PT Portugal RO Romania SE Sweden SK Slovakia SL Slovenia UK United kingdom EU

147 Table 17: SOMO35 in 2020 (ppm.d.*minh.) inhabitants base SEA1 SEA3 SEA4 SEA5 ASEA1 BSEA1 ASEA4 BSEA4 AT Austria BE Belgium BG Bulgaria CH Switzerland CY Cyprus CZ Czech republic DE Germany DK Denmark EE Estonia ES Spain FI Finland FR France GR Greece HU Hungary IE Ireland IT Italy LT Lithuania LU Luxembourg LV Latvia MT Malta NL Netherlands NO Norway PO Poland PT Portugal RO Romania SE Sweden SK Slovakia SL Slovenia UK United kingdom EU

148 Table 18: Differences between scenarios and years for SOMO35 (EU27+2 total) abs abs % (ppm.d.*minh.) (ppm.d.*minh.) % SEA1-base % % SEA2-base % n.c n.c SEA3-base % % SEA4-base n.c n.c % SEA5-base n.c n.c % BSEA1-ASEA % 0 0.0% ASEA1-SEA % 0 0.0% BSEA1-SEA % 0 0.0% A-SEA4-ASEA1 n.c n.c 6 0.7% BSEA4-BSEA1 n.c n.c 6 0.7% A-base 0 0.0% n.c n.c B-base 0 0.0% n.c n.c B-A 0 0.0% n.c n.c SEA2-SEA % n.c n.c SEA3-SEA % n.c n.c SEA3-SEA1 NC NC 5 0.6% SEA4-SEA1 NC NC 6 0.7% SEA5-SEA3 NC NC 2 0.3% SEA5-SEA4 NC NC 1 0.1% ASEA1-base % n.c n.c BSEA1-base % n.c n.c n.c - not calculated 142

149 Table 19: AOT40c in 2015 (ppm.d.*minh.) inhabitants base SEA1 SEA2 SEA3 ASEA1 BSEA1 A B AT Austria BE Belgium BG Bulgaria CH Switzerland CY Cyprus CZ Czech republic DE Germany DK Denmark EE Estonia ES Spain FI Finland FR France GR Greece HU Hungary IE Ireland IT Italy LT Lithuania LU Luxembourg LV Latvia MT Malta NL Netherlands NO Norway PO Poland PT Portugal RO Romania SE Sweden SK Slovakia SL Slovenia UK United kingdom EU

150 Table 20: AOT40c in 2020 (ppm.d.*minh.) Inhabitants base SEA1 SEA3 SEA4 SEA5 ASEA1 BSEA1 ASEA4 BSEA4 AT Austria BE Belgium BG Bulgaria CH Switzerland CY Cyprus CZ Czech republic DE Germany DK Denmark EE Estonia ES Spain FI Finland FR France GR Greece HU Hungary IE Ireland IT Italy LT Lithuania LU Luxembourg LV Latvia MT Malta NL Netherlands NO Norway PO Poland PT Portugal RO Romania SE Sweden SK Slovakia SL Slovenia UK United kingdom EU

151 Table 21: Differences between scenarios and years for AOT40c (EU27+2 total) abs abs % (ppm.d.*minh.) (ppm.d.*minh.) % SEA1-base % % SEA2-base % n.c n.c SEA3-base % % SEA4-base n.c n.c % SEA5-base n.c n.c % BSEA1-ASEA % 0 0.0% ASEA1-SEA % 1 0.0% BSEA1-SEA % 1 0.1% A-SEA4-ASEA1 n.c n.c % BSEA4-BSEA1 n.c n.c % A-base 0 0.0% n.c n.c B-base 1 0.0% n.c n.c B-A 1 0.0% n.c n.c SEA2-SEA % n.c n.c SEA3-SEA % n.c n.c SEA3-SEA1 NC NC % SEA4-SEA1 NC NC % SEA5-SEA3 NC NC % SEA5-SEA4 NC NC 8 0.4% ASEA1-base % n.c n.c BSEA1-base % n.c n.c n.c. - not calculated 145

152 Table 22: AOT40f in 2015 (ppm.d.*minh.) inhab. base SEA1 SEA2 SEA3 ASEA1 BSEA1 A B AT Austria BE Belgium BG Bulgaria CH Switzerland CY Cyprus CZ Czech republic DE Germany DK Denmark EE Estonia ES Spain FI Finland FR France GR Greece HU Hungary IE Ireland IT Italy LT Lithuania LU Luxembourg LV Latvia MT Malta NL Netherlands NO Norway PO Poland PT Portugal RO Romania SE Sweden SK Slovakia SL Slovenia UK United kingdom EU

153 Table 23: AOT40f in 2020 (ppm.d.*minh.) inhabitants base SEA1 SEA3 SEA4 SEA5 ASEA1 BSEA1 ASEA4 BSEA4 AT Austria BE Belgium BG Bulgaria CH Switzerland CY Cyprus CZ Czech republic DE Germany DK Denmark EE Estonia ES Spain FI Finland FR France GR Greece HU Hungary IE Ireland IT Italy LT Lithuania LU Luxembourg LV Latvia MT Malta NL Netherlands NO Norway PO Poland PT Portugal RO Romania SE Sweden SK Slovakia SL Slovenia UK United kingdom EU

154 Table 24: Differences between scenarios and years for AOT40f (EU27+2 total) abs abs % (ppm.d.*minh.) (ppm.d.*minh.) % SEA1-base % % SEA2-base % n.c n.c SEA3-base % % SEA4-base n.c n.c % SEA5-base n.c n.c % BSEA1-ASEA % 0 0.0% ASEA1-SEA % 1 0.0% BSEA1-SEA % 1 0.0% A-SEA4-ASEA1 NC NC % BSEA4-BSEA1 NC NC % A-base 0 0.0% n.c n.c B-base 1 0.0% n.c n.c B-A 1 0.0% n.c n.c SEA2-SEA % n.c n.c SEA3-SEA % n.c n.c SEA3-SEA1 n.c n.c % SEA4-SEA1 n.c n.c % SEA5-SEA3 n.c n.c % SEA5-SEA4 n.c n.c % ASEA1-base % n.c n.c BSEA1-base % n.c n.c n.c - not calculated 148

155 Table 25: O3avg in 2015 (ppm.d.*minh.) inhabitants base SEA1 SEA2 SEA3 ASEA1 BSEA1 A B AT Austria BE Belgium BG Bulgaria CH Switzerland CY Cyprus CZ Czech republic DE Germany DK Denmark EE Estonia ES Spain FI Finland FR France GR Greece HU Hungary IE Ireland IT Italy LT Lithuania LU Luxembourg LV Latvia MT Malta NL Netherlands NO Norway PO Poland PT Portugal RO Romania SE Sweden SK Slovakia SL Slovenia UK United kingdom EU

156 Table 26: O3avg in 2020 (ppm.d.*minh.) Inhabitants base SEA1 SEA3 SEA4 SEA5 ASEA1 BSEA1 ASEA4 BSEA4 AT Austria BE Belgium BG Bulgaria CH Switzerland CY Cyprus CZ Czech republic DE Germany DK Denmark EE Estonia ES Spain FI Finland FR France GR Greece HU Hungary IE Ireland IT Italy LT Lithuania LU Luxembourg LV Latvia MT Malta NL Netherlands NO Norway PO Poland PT Portugal RO Romania SE Sweden SK Slovakia SL Slovenia UK United kingdom EU

157 Table 27: Differences between scenarios and years for O3avg (EU27+2 total) abs abs % (ppm.d.*minh.) (ppm.d.*minh.) % SEA1-base % % SEA2-base % n.c n.c SEA3-base % % SEA4-base n.c n.c % SEA5-base n.c n.c % BSEA1-ASEA % % ASEA1-SEA % % BSEA1-SEA % % A-SEA4-ASEA1 n.c n.c % BSEA4-BSEA1 n.c n.c % A-base % n.c n.c B-base % n.c n.c B-A % n.c n.c SEA2-SEA % n.c n.c SEA3-SEA % n.c n.c SEA3-SEA1 n.c n.c % SEA4-SEA1 n.c n.c % SEA5-SEA3 n.c n.c % SEA5-SEA4 n.c n.c % ASEA1-base % n.c n.c BSEA1-base % n.c n.c n.c - not calculated 151

158 Table 28: Difference between 2015 and 2020 per effected area NS, B, MED, NS, B NS, B NS, B NS, B BS SECA Baseline SEA1 SEA3 ASEA1 BSEA1 S-deposition N-deposition Acidification Eutrophication PM10 AOT40f abs % -2% -4% -5% -3% -3% abs % -3% -5% -5% -5% -5% abs % -4% -11% -11% -11% -10% abs % -11% -13% -13% -13% -13% abs % -6% -8% -9% -8% -8% abs % -7% -11% -11% -11% -11% 152

159 17 Appendix 5 - Health Benefit Diagrams SEA1 Scenario SEA2 Scenario

160 SEA3 Scenario AS1 Scenario

161 BS1 Scenario SEA1 Scenario

162 SEA3 Scenario SEA4 Scenario

163 SEA5 Scenario ASEA1 Scenario

164 BSEA1 Scenario ASEA4 Scenario

165 BSEA4 Scenario

166 18 Appendix 6 - Methodology for cost calculations Scenario description The following parameters are specified in each scenario: 1. Fuel use [PJ] 2. Implementation of Tier I, II, III NO X control technologies [% of fuel] 3. Sulphur control: * The sulphur content of fuel used, [mass % in fuel] or * Implementation rate of new or retrofit scrubbers [% of vessels] 4. Costs for NO X and SO 2 control Scenario background data specification Scenarios: Baseline, SEA1, SEA2, SEA3, SEA4, SEA5 (see chapter 6 for more info) Years: 2015 and 2020 Sea regions: Baltic Sea Black Sea Mediterranean North East Atlantic North Sea & English Channel Vessel types: Ferry Bulk carrier Chemical tanker Container ship General Cargo LG tanker Oil tanker RoRo cargo Engine types: Abbreviation SSD-MG/DO MSD-MG/DO HSD-MG/DO GT-MG/DO ST-MG/DO SSD-RO MSD-RO HSD-RO GT-RO ST-RO Description Slow Speed engines using Marine Diesel or Gas Oil Medium Speed engines using Marine Diesel or Gas Oil High Speed engines using Marine Diesel or Gas Oil Gas turbines using Marine Diesel or Gas Oil Steam turbines using Marine Diesel or Gas Oil Slow Speed engines using Residual Oil Medium Speed engines using Residual Oil High Speed engines using Residual Oil Gas turbines using Residual Oil Steam turbines using Residual Oil 160

167 Engine size classification: Main engine Small Medium Large range [kw] Average ME size [kw] SSD 2 share [%] Auxiliary engine size Small Medium Large range [kw] Average AE size [kw] The engine size classification follows the engine size classification as specified in the ENTEC 2005 reports. The distribution of the engine sizes relative to the total fuel use is derived from ENTEC 2005 and is expressed as % of total fuel use. Share of fuel used in respective engine size category: Small Medium Large 25% 41% 34% Scenario specific parameters Fuels considered and sulphur content The following fuels are available for use in the scenarios. The use of each fuel is scenario-specific and depends on whether the sea region is considered as a Sulphur Emission Control Area (SECA) in the scenario. When scrubbers are used instead of low-sulphur fuel oils, a sulphur content of maximum 2.94 % is used for the region. Fuel type Sulphur content Marine Diesel (MD) 0.5 % MD 0.1 % Heavy Fuel Oils (HF) / Residual Oils (RO) 2.94 % HF 1.5 % HF 0.5 % HF 0.1 % Implementation of Tier I, II, III NO x control technologies The implementation of Tier control technologies correspond to the vessel vintages specified in Extremis. In the cost estimates it is assumed that the age distribution will remain the same for the years 2005, 2015 and For 2015 and all sea regions, the following implementation rates of Tier technologies are obtained: No NO 2015 Tier III Tier II Tier I 30 % Tier I x - control Ferry 0% 16% 28% 17% 39% Bulk carrier 0% 20% 37% 13% 30% Chemical tanker 0% 29% 44% 8% 19% Container ship 0% 32% 46% 7% 16% General Cargo 0% 7% 26% 20% 47% LG tanker 0% 29% 44% 8% 19% Oil tanker 0% 29% 44% 8% 19% RoRo cargo 0% 16% 28% 17% 39% 161

168 For 2020 in sea regions considered as NO X -ECAs, the following Tier technology implementation rates are obtained: 2020 NO X -ECA Tier III Tier II Tier I 30% Tier I No NO X - control Ferry 16% 16% 20% 15% 34% Bulk carrier 20% 22% 27% 9% 22% Chemical tanker 29% 23% 33% 5% 11% Container ship 32% 29% 25% 4% 10% General Cargo 7% 15% 22% 17% 39% LG tanker 29% 23% 33% 5% 11% Oil tanker 29% 23% 33% 5% 11% RoRo cargo 16% 16% 20% 15% 34% For 2020, in sea regions considered as outside NO X -ECA, the following Tier technology implementation rates are obtained: 2020 outside No NO Tier III Tier II Tier I 30% Tier I X - NO X -ECA control Ferry 0% 32% 20% 15% 34% Bulk carrier 0% 42% 27% 9% 22% Chemical tanker 0% 52% 33% 5% 11% Container ship 0% 61% 25% 4% 10% General Cargo 0% 22% 22% 17% 39% LG tanker 0% 52% 33% 5% 11% Oil tanker 0% 52% 33% 5% 11% RoRo cargo 0% 32% 20% 15% 34% Sulphur control - scrubbers In the calculations where SO 2 scrubbers are used instead of low-sulphur fuels, the following implementation rates are obtained from considering the age distribution in Extremis The scrubber implementation rates takes into consideration whether retrofit is needed or if new installations are sufficient. The 'Retrofit' categories include the vessels constructed before 2011, while the 'New' category includes the vessels constructed after Scrubbers -new and retrofits Retrofit New Retrofit New Ferry 84% 16% 68% 32% Bulk carrier 80% 20% 58% 42% Chemical tanker 71% 29% 48% 52% Container ship 68% 32% 39% 61% General Cargo 93% 7% 78% 22% LG tanker 71% 29% 48% 52% Oil tanker 71% 29% 48% 52% RoRo cargo 84% 16% 68% 32% 162

169 The Gemini Building Fermi Avenue Harwell International Business Centre Didcot Oxfordshire OX11 0QR Tel: Fax: