Jensen, S.S., Ketzel, M., Brandt, J., Martinsen, L., Becker, T. 2014: Ren-luftzone i København og sparede eksterne omkostninger ved sundhedsskadelig luftforurening. Aarhus Universitet, DCE – Nationalt Center for Miljø og Energi, 59 s. - Videnskabelig rapport fra DCE - Nationalt Center for Miljø og Energi nr. 58. http://www.dmu.dk/Pub/SR58.pdf
Based on the results of previously conducted air quality assessment of proposed clean air zones for Copenhagen (Jensen et al. 2012b) the Environmental Protection Agency (EPA) and the Municipality of Copenhagen want to carry out an assessment of the potential avoided external costs related to health effects due to proposed clean air zones. The geographic extend of the proposed clean air zone in Copenhagen is identical to the current low emission zone covering Copenhagen and Frederiksberg.
An external cost is defined as the cost incurred by others from an extra kilometers traveled and not paid for. The external costs of air pollution (excluding CO2) is related to health effects of air pollution which are calculated using the so-called "impact pathway" approach where the external costs of air pollution are calculated from the source, its dispersion, population exposure, dose-effect relationships for quantification of the health impacts of both premature death and morbidity, and a pricing of health effects in order to calculate the total external costs. Based on this information unit costs can be calculated e.g. cost per kg emission.
Estimation of the external cost savings due to avoided health effects (benefits) will be included in a comprehensive updated economic assessment of clean air zones where COWI will update costs due to decline in prices for used car etc. (costs).
The project aims to estimate the external cost savings related to health effects by introducing various forms of clean air zones in Copenhagen. The assessment is based on unit costs of air pollution and the calculation of the total saved emissions of the different scenarios.
Calculation of the avoided external costs is conducted for selected scenarios for clean air zones.
The starting point is the so-called Berlin scenario that includes emission requirements for both passenger cars and vans. The requirements are that diesel cars and vans up to and including Euro 3 emission standard and gasoline powered cars and vans up to and including Euro 0 are not allowed to drive in the clean air zone. The calculations were made for 2015.
Another scenario is an additional requirement to the Berlin scenario where all Euro 4 diesel vehicles must have a particulate filter. The additional requirement applies to all diesel Euro 4 vehicles, that is, passenger cars, vans and trucks as well as buses must have a particulate filter. As it is possible to meet this requirement either by retrofitting of closed particulate filters or by switching to Euro 5 or Euro 6 vehicles, the calculations assumes that 50% are retrofitted with particle filters and 50% are distributed proportionally among the remaining categories of vehicles (Euro 5 and Euro 6 ) in the same proportion as before the implementation. The Danish Transport Authority assumes that a closed particle filter on average reduces particulate exhaust emissions by 80% and the same assumption is applied in this study. The calculations were made for 2017 to allow a little longer phasing-in of requirements to the Euro 4 vehicles.
Furthermore, calculations were also made for a separate scenario where all urban buses must meet the Euro 6 emission standard. Buses include urban buses and coaches, and this scenario only includes the urban buses that run on a regular basis. The calculations were made for 2015.
The external cost savings calculated for scenarios in 2015 are compared with the reference scenario in 2015, and the scenario in 2017 is compared with the baseline 2017.
As a sensitivity analysis of the calculations for the reduction of emissions and avoided external costs of the Berlin scenario, this scenario was also calculated in a version based on the assumption that open particulate filters were retrofitted instead of a change of car. Car owners have the option to retrofit open particulate filter, and then be allowed to drive in the clean air zone. EPA has established the following scenario based on what would pay for car owners. It is assumed only to pay for passenger cars as it will be too expensive for vans. It is assumed that 90% of Euro 3 diesel passenger cars are retrofitted with open filters and 25% are retrofitted with open filter for Euro 2 diesel passenger cars. It is assumed that open filters reduce particulate exhaust emissions by 30%.
The overall approach is that the total external costs avoided by a scenario is the unit costs times the avoided emissions. Unit costs for external costs of health effects are expressed in DDK per kg emission, and the avoided emissions are calculated based on traffic performance on the road network in Zealand, emission factors, and changes in the distribution of Euro standards of the car fleet due to the different scenarios. Furthermore, it is taken into account that the effect in the rest of Zealand of a clean air zone in Copenhagen decreases further away from the zone.
Unit costs are based on the EVA-system developed by AU/DCE. The EVA-system (Economic Valuation of Air pollution) is able to calculate the external costs of the individual emissions based on Danish valuations of health effects depending on the geographical locations of the emissions (Brandt et al. 2011a). This system is based on state-of-the-art air quality models for the calculation of air quality and population exposure (Danish Eulerian Hemispheric Model - DEHM), dose-response relationships for calculating the health effects and valuation of health effects. The valuation is based on willingness to pay to avoid such premature deaths and market prices for hospital admissions and increases in medication use.
The EVA-system operates currently with one average price of unit costs for road transport in Denmark for PM2.5. As scenarios for clean air zones takes place in a densely populated area, an indicative methods has been set up to try to create a greater spatial resolution of the external costs of PM2.5. This is done by splitting the unit costs in a regional contribution and an urban contribution where the urban contribution is dependent on population density. It is assumed that there is a significant uncertainty in this indicative method.
The total estimated external costs are projected from 2006 to 2015 based on the assumptions of the Ministry of Transport of an increase in unit costs per kg emission of 1.6% per year. This results in a price increase of 15% from 2006 to 2015. Projections to 2017 gives a price increase of 19% compared to 2006 prices.
The population density (inh./km2) is included in the calculation method for external costs of PM2.5. The data set is based on the Danish Central Person Registry (CPR) from 2008 and joined to the address register to be able to summarize persons per address. Population data is then aggregated to a standardized grid of 1x1 km2.
The above method calculates the avoided costs for a given scenario for a given year, i.e. the first year of introduction of the measure. The scenarios correspond to acceleration of the introduction of newer cleaner Euro standards for vehicles, and the effect in terms of saved emissions are greatest in the first year, after which it will be reduced due to the replacement of the car fleet. The EPA has previously calculated the saved emissions of the Berlin scenario (EPA 2009a). The effect is greatest the first year, and after about 10 years, the savings are down to about 10-15% of the effect in the first year. After 20 years there is no effect of this approach, as the car fleet is replaced. The cumulative avoided emissions compared to the first year of saved emissions are 5.5 times greater for NOx and 5 times greater for PM2.5, and these factors are used to calculate the total external cost savings of a scenario by multiplying these factors on the calculated avoided external costs for the first year.
The saved emissions of the scenarios are calculated on the basis of road traffic volumes and vehicle distributions on the road network on Zealand from DCE's road and traffic data base (Jensen et al 2010), emission factors based on the emission model COPERT 4 and the changes in distribution of Euro emission standards of the car fleet that the different scenarios give rise to. In this way, it is possible to calculate emissions from road traffic with a spatial resolution of 1x1 km2 grid. A standardized grid from the Danish Geodata Agency is used. OML-Highway is used to make the calculations of road traffic performance since it has a built-in tool to easily summarize traffic performance and emissions on a grid (Jensen et al. 2010b).
In the clean air zone, it is assumed that there is no changes in traffic volume, thus there are only changes in the distribution of Euro emission standards of the car fleet due to the requirements of the clean air zone. Therefore, the requirements of the clean air zone have 100% impact within the clean air zone.
Outside the clean air zone, there will also be changes in the car fleet as a result of the clean air zone requirements in Copenhagen but the impact will decrease further away from Copenhagen. This is because car owners who often go to Copenhagen are likely to switch to a car that meets the clean air zone requirements while car owners who are rarely in Copenhagen are not likely to replace their car. It is also assumed that traffic volume is unchanged and that only the car fleet as changed.
COWI has previously reported (EPA 2009b) on the impacts of clean air zones on passenger cars and vans to assess how many cars are affected by clean air zone in different geographical areas of Zealand by combining information on the car fleet and survey data from DTU Transport about transport behavior (TU data). These assessments have COWI updated (COWI 2013), and these new assumptions are taken into account in the present study.
For example, if 60% of older vehicles are affected by the new requirements and if it has been estimated that the emission reduction within the clean air zone in Copenhagen is 10% then emissions saved in the given area is 10% * 60% = 6.0% of the avoided emissions in that area. In this way, the effect outside the clean air zone for different areas of Zealand is calculated.
The Berlin Scenario will reduce NOx emissions from road traffic by 8% within the clean air zone, and 4% for the entire Zealand. The corresponding reduction of PM2.5 emissions are respectively 11% and 7% (Table 3.2). As expected, the percentage reduction of NOx and PM2.5 in the clean air zone, and the relationship between the effect on cars and vans are consistent with results from the previous DCE report on air quality assessment of clean air zones (Jensen et al. 2012b). The saved emissions and costs for vans are estimated to be distributed on 1/3 on vans below 2,500 kg and 2/3 on vans over 2,500 kg.
The Berlin Scenario will the first year save about DKK 44 mio. in external costs and over all the years in which it has an impact it will save about DKK 238 mio. due to reduced NOx emissions (Table 3.3). About 3/4 of the effect is due to reduced NOx emissions from passenger cars and 1/4 from vans.
Due to reduced PM2.5 emission the Berlin Scenario will the first year save about DKK 68 mio. in external costs and over all the years in which it has an impact it will save about DKK 343 mio. About ½ of the effect is caused by passenger cars and ½ by vans. Reduction of NOx and PM2.5 emissions of the Berlin Scenario will the first year save around DKK 112 mio. in external costs and over all the years in which it has an impact it will save about DKK 581 mio. About 2/3 of the effect is caused by passenger cars and 1/3 by vans.
As a sensitivity analysis of the calculations for the reduction of emissions and avoided external costs of the Berlin Scenario, a version was assessed where open particle filters were retrofitted instead of changing cars. Car owners have the option to retrofit open particulate filter, and then be allowed to drive in the clean air zone. EPA has established the following scenario where it is assumed that 90% of Euro 3 diesel passenger cars are retrofitted with open filters and 25% are retrofitted with open filters on Euro 2 diesel passenger cars. The Berlin Scenario incl. retrofitting of open filters on diesel cars will reduce the external costs avoided for approximately DKK 120 mio. for NOx and about DKK 265 mio., totally DKK 385 mio. This means that the avoided external costs are about 34% lower when retrofitting open filters in relation to substitution of Euro classes.
In practice, fewer car owners may retrofitt filters as it is time consuming and inconvenient, and instead change to a car that meets the clean air zone requirements. Therefore, the estimated external cost savings of the Berlin Scenario is expected to lie somewhere between the retrofitting and replacement, that is, DDK 385-581 mio.
Berlin Scenario incl. filter requirements for Euro 4 diesel vehicles
This scenario is as the above Berlin Scenario but with an additional requirement that all diesel Euro 4 vehicles, that is, passenger cars, vans and trucks as well as buses must have a closed particulate filter. As it is possible to meet this requirement either by retrofitting particulate filters or by switching to the Euro 5 or Euro 6 vehicle it is assumed that 50% is retrofitting particle filters and 50% shift to Euro 5 and Euro 6. This scenario reduces NOx emissions by about 13% in the clean air zone, and 6% for the entire Zealand, and PM2.5 emissions are reduced by 18% in the clean air zone, and 10% for the entire Zealand. As expected, the reduction is greater than that of Berlin Scenario as all Euro 4 diesel vehicles are included.
As expected, this scenario has considerably larger avoided external costs for NOx compared to the Berlin Scenario, as NOx emissions from Euro 4 diesel vehicles constitute a significant proportion of the vehicle fleet in 2015 (around 22%) according to Jensen et al. (2012b). 50% of Euro 4 diesel vehicles is assumed to be replaced by Euro 5 and Euro 6 vehicles which generally have lower NOx emissions than Euro 4 (however, only for Euro 6 for the heavy vehicles). A similar pattern is seen for PM2.5 where the effect comes from the retrofitting of particle filters and shift to Euro 5 and Euro 6 where Euro 5 and Euro 6 reduce emissions compared to Euro 4 for light vehicles, while only Euro 6 reduces particulate emissions compared to Euro 4 for heavy vehicles.
The Berlin Scenario incl. filter requirements for Euro 4 diesel vehicles reduces external costs the first year of about DKK 136 mio. and over all the years in which it has an effect it will save about DKK 702 mio.
Compared to the other scenarios the Euro 6 urban bus scenario only has effect within the clean air zone, and it is assumed that none of the Euro 6 buses run outside the Copenhagen outer municipal boundary. As some bus lines in practice cross the outer municipal boundary, this scenario will also have an effect outside the clean air zone but it has not been possible to include this effect and it is considered to be modest.
Due to the current low emission zone rules that require particulate filter up to and including Euro 3 heavy vehicles, the majority of all urban buses are Euro 5, and then comes Euro 4, and only a very small proportion is Euro 6 in 2015 (Jensen et al. 2012b). A shift to Euro 6 will reduce NOx emissions and particulate exhaust emissions compared to Euro 4 and 5, if the Euro 6 lives up to the expected reduction in emissions in practice.
This scenario reduces NOx emissions by about 3% and PM2.5 by 0.4% in the clean air zone. If the reduction is compared to the emission from all buses (both urban buses and coaches) the NOx emission reduction is about 54% and for particulate exhaust emissions about 28% but the overall reduction in the clean air zone is small because buses represent a very small part of the traffic volume, and many streets have no buses.
This scenario saves the first year about DKK 4 mio. in external costs and over all the years in which it has an effect, it will save about DKK 20 mio. due to reduced NOx emissions. Reduced PM2.5 emission provides the first year only a saving of about DKK 1.5 mio. in external costs and over all the years in which it has an effect, it will save about DKK 7 mio. Due to reduced NOx and PM2.5 emissions this scenario will the first year save about DKK 5 mio. in external costs and over all the years in which it has an effect, it will save about DKK 28 mio.
However, this measure is very effective for streets with bus traffic. According to the DCE report on air quality assessment of clean air zones (Jensen et al. 2012b) this scenario reduces NOx emissions by about 9% at H.C. Andersens Boulevard in 2015 which is the same effect as the Berlin Scenario. This street has about 1.2% urban buses and H.C. Andersens Boulevard is one of the most busy urban streets in Copenhagen.
Comparison of Unit Costs of the Ministry of Transport
The unit costs of the present study has been compared with the unit costs of the Ministry of Transportation that represents the official transportation unit costs for the calculation of the external costs avoided since they are used for welfare economic assessment of policy measures in the transportation sector.
The unit costs for PM2.5 in this study is depending on population density. For Copenhagen with an average density of about 7,700 inh./km2 the urban contribution is (7.700/128) / 2 * 46/ kg = DKK 1,384/ kg plus the regional contribution of DKK 285/kg given a total unit cost of ~ DKK 1,700/kg (2006 prices). This is in line with the Ministry of Transport that has unit cost for a city of DKK 1718/kg (2010 prices).
The unit cost for NOx is DKK 75 per kg NOx emission (in 2006 prices). Since the effect of NOx is primarily at the regional scale it is considered the same for urban and rural areas, as both the Ministry of Transport and the EPA have previously assumed. The unit cost of EVA is here about 50% higher than that of the Ministry of Transport (DKK 52 per kg).
The largest uncertainties are estimated to be in the unit costs, and especially the indicative method for greater spatial resolution in the cities for PM2.5. This uncertainty could be further reduced by developing the EVA system to include a local scale air quality model as part of the system, so that calculations could be performed on 1x1 km2 instead of presently 16,7 x16, 7 km2.
There is less uncertainty on emissions that are based on emission models that are based on measurements of emissions. Further, emissions can be indirectly validated by comparing measured and calculated concentrations, where air quality models calculate concentrations based on emissions data.
Emissions from future emission standards such as Euro 6 is an uncertainty as there is still relatively few vehicles on the market that meet the Euro 6 and few emission measurements available. This is particularly important for the scenario where only Euro 6 urban buses are allowed. Past experience has shown that the Euro 5 for heavy vehicles did not deliver the expected NOx emission reduction as compared to Euro 4 and emission factors for Euro 5 had to be assumed as Euro 4.
There is also some uncertainty on the impact of the scenarios outside the clean air zone, which uses a method developed by COWI. This method uses information about the number of vehicles likely to be affected by the requirements. The main purpose of the method is to estimate the number of affected vehicles to estimate the loss in value etc. This method is compiled from car fleet data and expected user behavior, and it generally provides plausible results since the effect decreases with distance from the clean air zone. In the present study the results of the method are used to estimate the effect outside the clean air zone but the method is not directly related to road traffic which is the basis for calculating the avoided emissions.