Estimating national exhaust emissions from railway vehicles in Turkey

ent 374 (2007) 127–134www.elsevier.com/locate/scitotenv

Science of the Total Environm

Estimating national exhaust emissionsfrom railway vehicles in Turkey

Faruk Dincer, Tolga Elbir ⁎

Dokuz Eylul University, Faculty of Engineering, Department of Environmental Engineering, Buca, 35160 Izmir, Turkey

Received 2 May 2006; received in revised form 7 December 2006; accepted 14 December 2006Available online 29 January 2007

Abstract

The estimated exhaust emissions from railway vehicles in Turkey were presented. The emissions of nitrogen oxides (NOX),hydrocarbon compounds (HC), carbon monoxide (CO), particulate matter (PM), sulfur dioxide (SO2) and carbon dioxide (CO2)from the diesel locomotives and railcars were calculated using the railway traffic data recorded by Turkish State Railways (TSR)for the period of 2000–2005. EPA emission factors were used for different vehicle types and operation modes such as shunting andline-hauling. Total emissions from railway vehicles in Turkey were estimated as 384 t y−1 for HC, 1016 t y−1 for CO, 6799 t y−1

for NOX, 256 t y−1 for PM, 357 t y−1 for SO2 and 383 537 t y−1 for CO2 for the year 2005. The distribution of emissions withrespect to type of railway vehicles shows that the mainline locomotives contribute ∼91% to the total emissions. The increases of22%, 39% and 49% in the current numbers of mainline locomotives, shunting locomotives and diesel railcars, respectivelycorresponding to the full capacity of railway network in Turkey will increase the annual emissions to 431 t y−1 for HC, 1121 t y−1

for CO, 7399 t y−1 for NOX, 342 t y−1 for PM, 552 t y−1 for SO2 and 420 256 t y−1 for CO2. Total railway emissions constitute0.15%, 0.08% and 4.21% of total Turkish traffic emissions for HC, CO and NOX, respectively.© 2006 Elsevier B.V. All rights reserved.

Keywords: Non-road emission; Railway emission; Diesel locomotives; Emission inventory

1. Introduction

Anthropogenic emissions are classified into threemain categories; mobile, stationary (point), and areasources. Mobile source emissions are further splittedinto on-road (e.g., cars, trucks, and motorcycles) andnon-road emission categories. Because of the significantcontribution of non-road emission sources to the totalmobile source emission inventories, over the pastseveral years it has become critical to prepare their

⁎ Corresponding author. Tel.: +90 232 4127133; fax: +90 2324530922.

E-mail address: [email protected] (T. Elbir).

0048-9697/$ - see front matter © 2006 Elsevier B.V. All rights reserved.doi:10.1016/j.scitotenv.2006.12.043

inventories to satisfy the specific requirements of theClean Air Act Amendments of 1990 (EPA, 2005).

Railway transportation is one of the non-road sourcecategories. Diesel locomotives running with engineshaving no emission control affect the local communitiesand ecosystems throughout a region. The major airpollutants related to railway transportation that canaffect human health include diesel exhaust, nitrogenoxides (NOX), particulate matter (PM), volatile organiccompounds (VOCs), ozone, and sulfur oxides (SOX). Inaddition, diesel exhaust contains an estimated total of450 different compounds, about 40 of which are listed astoxic air contaminants with negative effects on healthand the environment (Bailey and Solomon, 2004).

Fig.1.

Railway

networkin

Turkey.

128 F. Dincer, T. Elbir / Science of the Total Environment 374 (2007) 127–134

129F. Dincer, T. Elbir / Science of the Total Environment 374 (2007) 127–134

EPA has adopted three separate sets of emissionstandards for locomotives (EPA, 1997). These sets ofstandards (Tier 0, Tier 1 and Tier 2) apply to loco-motives and locomotive engines originally manufac-tured between 1973–2001, 2002–2004 and 2005 orlater, respectively. For all cases, locomotives and loco-motive engines are required to meet the applicablestandards at the time of original manufacture and at eachsubsequent remanufacture. Exempted from the emissionstandards are historical steam-powered locomotives,and locomotives originally manufactured before 1973.

There have been several studies estimating the diesellocomotive emissions. Emission estimation studies fordiesel locomotives are based on either fuel consumptiondata (Kean et al., 2000) or direct exhaust measurements(Fritz and Cataldi, 1991; Fritz, 1994; Fritz et al., 1995;Paukert, 2002). Fritz et al. (1995) performed exhaustemission, smoke opacity and fuel consumption tests onthirteen different locomotives from 1992 through 1993.Another study (Fritz, 1994) included steady-stateexhaust emission test results for hydrocarbons, carbonmonoxide, oxides of nitrogen, and particulate matterfrom two passenger locomotives. Kean et al. (2000)described a fuel-based method to determine the emis-sions from diesel locomotives. Jorgensen and Sorenson(1998) studied typical emission factors from railwaydiesel engines along with several traffic parameters suchas passenger occupancy rates, per capita freight andpassenger traffic for European conditions.

In recent years, several studies have focused on de-termination of the impacts of railway emissions. Envi-ronmental impacts of railway transportation (Plakhotniket al., 2005) and air quality modeling studies for railwaytransportation system emissions were also reported (Aly,2002). Aly (2002) derived new models to estimate the

Fig. 2. Lengths of electrified and non-electri

pollution from railway traffic and analyze the factorseffecting energy consumption in railway systems. Pla-khotnik et al. (2005) designed a software for themonitoring of ecological effects of railway transportation.

Several emission estimation studies in Turkey havebeen conducted recently (Muezzinoglu et al., 1998; Elbiret al., 2000; Elbir and Muezzinoglu, 2004). However,these studies only included emissions from industries anddomestic heating. Emissions from mobile sources (i.e.,aircrafts, trains, and ships) have not been included in theTurkish emission inventories due to the lack of data. Onthe other hand, almost all European countries havecompleted their national mobile emission inventories.

This study estimates the exhaust emissions (NOX,HC, CO, PM, SO2 and CO2) from railway traffic inTurkey. The railway traffic data recorded by the TurkishState Railways (TSR) including type and number ofrailway vehicles, number of passengers, and the emis-sion factors from EPA's database were used for thecalculation of railway emissions for the first time inTurkey.

2. Railway network in Turkey

The total length of the railway network in Turkey is10 984 km (TSR, 2006) (Fig. 1). The network is used ona so called mixed traffic basis, i.e. the same railwaysystem serves both passenger and freight traffic. About21% of the railway network (2336 km) is electrified andthe remaining is non-electrified. Electrical trains handleabout 30% of the total railway traffic.

The Turkish railway network has 764 tunnels with atotal length of 180 km and 24 710 bridges and culverts.There are still 4078 level crossings on the railwaynetwork (TSR, 2006).

fied railway lines in the main regions.

Table 1Types, numbers and powers of locomotives and railcars in 2005(TSR, 2006)

Locomotivetype

Enginemanufacturer

Locomotivenumber

Power perlocomotive(hp)

Mainline DE 11000 Krauss-Maffei 63 1065DE 18000 SMTE 1 1500DE 18100 SMTE 16 1500DE 21500 General Electric 5 2000DE 22000 General Motors 86 2000DE 24000 SMTE 272 1900DE 33000 General Motors 18 3300Total 461

Shunting DH 33100 MaK 7 360–450DH 6500 Shunters 5 650DH 3600 MaK 12 360–450DH 7000 Tulomsas 18 700DH 9500 Tulomsas 26 950Total 68

Dieselrailcar

DH 5500 Fiat 8 4×145DH 5600 Tuvasas and Fiat 11 2×225DH 5700 Fiat 30 2×280Total 49

Electricloco

E 40000 Alsthom 5 4000E 43000 Toshiba and

Tulomsas45 4325

E 52500 Koncar 21 5172Total 71

Electricrailcar

E 8000 Alsthom 17 1380E 14000 Tuvasas and

Groupement69 1400

Total 86

130 F. Dincer, T. Elbir / Science of the Total Environment 374 (2007) 127–134

The Turkish State Railway (TSR), acting as an inde-pendent body under the Ministry of Transport in Turkey,is responsible for the railway infrastructure manage-ment. The TSR has seven Regional Directorates for the

Fig. 3. The number of passengers traveling b

management of railway infrastructure at seven mainregions (Istanbul, Ankara, Izmir, Sivas, Malatya, Adana,and Afyon). The lengths of electrified and non-electrified railway network in these regions are shownin Fig. 2.

The Turkish railway vehicle fleet consists of morethan 700 units. The dominant driving power is diesel,with engines ranging in size from about 360 hp forshunting operations, to 3300 hp engines designed formainline operations. The fleet has mainline locomo-tives, shunting locomotives, diesel railcars, electriclocomotives and electric railcars. Table 1 summarizesthe information on these vehicles.

Transport demand in Turkey has grown significantlyover the past five decades. Overall, the demand (asmeasured by passenger-kilometers and ton-kilometers)has grown at an annual rate of nearly 8% since 1950.Demand for road transport has grown at an annual rateof about 7.6% while rail transport demand has grown atabout 2%, demand for water transport by 5% and air atover 16% per year (Audige, 2006). Road and railwaytransports represented about 37% and 55% of the totaltransport market in 1950, respectively. By the year 2000,road, transport represented 93%, of the total market,while rail, water, and air represented 4%, 2%, and 1% ofthe total intercity transport market in Turkey, respec-tively (Audige, 2006). While the current rail transporttask is not insignificant, it is certainly less important forthe Turkish economy than it was in the past. Althoughthe Turkish State Railway's rail traffic market sharehave declined significantly, overall railway traffic hasgrown somewhat.

The number of passengers traveling by railway vehi-cles in 2005 is around 77 million while it was more than85 million in 2000 (TSR, 2006). Fig. 3 shows the number

y railway for the period of 2000–2005.

Table 2Locomotive-kilometers and operating hours of Turkish railway network for the year 2005

Operation mode Diesel system Electrical system Total

Locomotive Railcar Total Locomotive Railcar Total

Locomotive kilometers (103 km) Train haulage 30 583 1666 32 249 8596 4551 13 147 45 396Assisting 3675 1754 5429 321 3827 4148 9577Light 1278 0 1278 164 1 165 1443Actual Loco km 35 536 3420 38 956 9081 8379 17 460 56 416Shunting 2234 1 2235 0 0 0 2235Hot waiting 1162 87 1249 399 776 1175 2424Total 38 932 3508 42 440 9480 9155 18 635 61 075

Operating hours (h) Train haulage 922 115 59 261 – 180 037 94 523 274 560 –Assisting 100 933 28 328 129 321 11 210 69 297 80 507 209 828Light 29 407 2558 31 965 5858 2825 8683 40 648Actual Loco km 1 052 455 90 147 161 286 197 105 166 645 363 750 250 476Shunting 279 225 98 279 323 0 0 0 279 323Hot waiting 290 429 21 520 311 949 99 785 194 104 293 889 605 838Total 1 622 109 111 765 752 558 296 890 360 749 657 639 1 135 637

131F. Dincer, T. Elbir / Science of the Total Environment 374 (2007) 127–134

of passengers transported by railway in Turkey between2000 and 2005. Other activity patterns of Turkish railwaysystem such as locomotive-kilometers and operatinghours are given in Table 2. Diesel and electricity con-sumption for the year 2005 are 143 973 m3 and218 697 000 kW h, respectively (TSR, 2006).

Ninety three percent of the goods are transported on-road and only 6% on railways in Turkey (TSR, 2006).The amount of freight transported by the railway isapproximately 19 million tons including agriculturalproducts and livestock, foodstuff and fodder, solid min-erals, petroleum products, ores, metallurgical products,construction materials, fertilizers, chemicals, machinesand vehicles (TSR, 2006). The railway network servesfor the present foreign trade export and import transportstraffic between European and Asian countries (Roma-nia, Hungary, and Iran) (TSR, 2006).

The TSR has seven ports (Fig. 1) with freight andpassenger ship capacities of 12 768 and 5109 in 2005,respectively (TSR, 2006). These ports serve for thepresent foreign trade export and import transports aswell as the transit traffic. The highest freight transportvolume occurs in Mersin Port with 16.3 million tons bythe year 2005 (TSR, 2006).

Table 3Emission factors for locomotives, g l−1

Manufacture year Locomotive type HC CO NOX PM

1973–2001 (Tier 0) Line-haul 2.64 7.03 47.02 1.77Switch 5.55 10.06 69.21 2.43

2002–2004 (Tier 1) Line-haul 2.59 7.03 36.72 1.77Switch 5.55 10.06 53.36 2.43

After 2004 (Tier 2) Line-haul 1.43 7.03 27.21 0.95Switch 2.91 10.06 40.15 1.14

Although TSR has a small portion in passengertransport, new projects are under construction to in-crease this portion. Three big projects are under con-struction. The first project is Ankara–Istanbul FastTrain. A trip on this route will take about 3 h at themaximum speed of 250 km h−1. Also, this project willshorten the railway line from 576 km to 416 km. Thesecond project is Ankara–Konya Fast Train. This newrailway will shorten the travel time between Ankara andKonya to 1 h. In both projects the trains will be SpanishCAF trains. Third project is the Marmaray Project. Thisproject is a shared rail underground system that willconnect the European and Asian side of Istanbul underthe Istanbul Bosphorus. At the end of this project,passenger transport in Istanbul (Region 1) will increaseup to 10 fold. The railway lines of these three projectsare shown in Fig. 1.

3. Methods for calculation of emission

Emissions estimates for railway vehicles are basedprimarily on a methodology described in EPA's theNon-road Engine and Vehicle Emissions Study (EPA,

Table 4Annual emissions from railway transport in Turkey for the period of2000–2005, t y−1

Year HC CO NOX PM SO2 CO2

2000 470.7 1244.0 8327.7 313.2 437.4 469 640.62001 372.2 983.2 6581.9 247.5 345.6 371 055.92002 373.9 988.3 6615.7 248.8 347.5 373 093.82003 388.6 1027.4 6877.5 258.6 361.3 387 908.02004 368.1 974.0 6519.6 245.2 342.7 367 933.82005 384.1 1015.6 6798.6 255.7 357.2 383 536.5

Table 5Regional railway emissions in 2005, t y−1

Region HC CO NOX PM SO2 CO2

1 80.9 214.0 1432.7 53.9 75.3 80 824.42 90.9 240.4 1609.3 60.5 84.6 90 789.03 52.6 139.0 930.7 35.0 48.9 52 503.34 61.3 162.1 1085.0 40.8 57.0 61 210.95 18.4 48.5 325.0 12.2 17.1 18 333.66 57.6 152.3 1019.5 38.3 53.6 57 515.17 22.4 59.2 396.4 14.9 20.8 22 360.3Total 384.1 1015.6 6798.6 255.7 357.2 383 536.5

132 F. Dincer, T. Elbir / Science of the Total Environment 374 (2007) 127–134

1991). This study includes a major assessment of off-road engine activity and emissions at the national andlocal levels, including emissions from over 80 differenttypes of equipment.

Emissions, E, were estimated by EPA for most off-road engine applications as follows:

E ¼ N � HP� LF� HRS� EF ð1Þ

where N is the engine population (number of engines inuse); HP is the average rated power of these engines; LFis the load factor, which indicates the typical operatingload of the engines relative to their average rated power;HRS is the average annual hours of use per engine; andEF is an emission factor expressed in units of work doneby the engine (e.g., g kW h−1 or g bhp h−1).

EPA has established emission standards for oxides ofnitrogen (NOx), hydrocarbons (HC), carbon monoxide(CO), particulate matter (PM) and smoke for newlymanufactured and remanufactured diesel-powered loco-motives and locomotive engines, which have previouslybeen unregulated. Three separate sets of emissionstandards have been adopted, with applicability of thestandards dependent on the date a locomotive is firstmanufactured. EPA has estimated the average emissionfactors for all locomotives, which reflects the penetra-tion of the Tier 0–Tier 2 locomotives into the fleetover time (EPA, 1997). The first set of emission factors(Tier 0) apply to locomotives and locomotive enginesoriginally manufactured from 1973 through 2001. The

Table 6Comparison of annual railway emissions per passenger-kilometers traveled a

Region/country

Year (g passenger-km−1)

NOX CO HC PM SO2 CO2

Europe 2000 0.28 0.04 – – – –UK 2002 0.44 0.10 – – 0.16 52.80USA 2002 1.12 0.08 0.05 0.05 – 53.86Canada 2004 – – – – – –Turkey 2005 1.35 0.20 0.08 0.05 0.07 76.16

second set of emission factors (Tier 1) apply to loco-motives and locomotive engines originally manufac-tured from 2002 through 2004. The final set of emissionfactors (Tier 2) apply to locomotives and locomotiveengines originally manufactured in 2005 and later.These average emission factors estimated by EPA aregiven in Table 3.

The types of locomotive engines used in Turkey arecommonly diesel-electric and manufactured in early1970s and 1980s. In order to estimate the currentnational emissions in this study, the emission factors ofNOX, HC, CO and PM for Tier 0 locomotives were usedfrom Table 3. The emissions of SO2 and CO2 wereestimated by fuel use data. The emission factorsestimated for Tier 2 locomotives which will have acleaner technology were used to estimate the nationalemissions from new future locomotives.

Diesel fuel quality has an influence on the type andamount of pollutants in the exhaust emissions. Espe-cially, sulfur levels influence the amount of compositePM and SO2 emitted. The sulfur content of the dieselfuel used in railway operations is around 0.5% in Turkey(Icingur and Altiparmak, 2003).

4. Results and discussion

The main source of Turkish railway emissions is thediesel usage on railway vehicles. The annual dieselconsumption is around 143 973 m3 for railway transportin 2005 (TSR, 2006). The percentages of diesel usagefor mainline and shunting operations are 99.1 and 0.9,respectively.

The exhaust emissions of HC, CO, NOX, PM, SO2

and CO2 from the Turkish locomotives were estimatedby using EPA's emissions factors and diesel usage dataon these vehicles. The amounts of the annual railwayemissions estimated for the 2000–2005 period are listedin Table 4. Estimations showed that the contribution ofmainline operations to the total emissions from railwaysis around 98–99%.

The distribution of emissions according to type ofrailway vehicles shows that the mainline locomotives

nd tons freight-kilometers carried at different countries

(g t-km−1) Reference

NOX CO HC PM SO2 CO2

0.16 0.03 0.01 – – – EEA (2006)0.37 0.11 – – 0.06 63.12 TWUK (2006)0.43 0.03 0.02 0.01 – 20.61 DOT (2005)0.32 0.05 0.02 0.01 0.01 16.58 EC (2003)0.74 0.11 0.04 0.03 0.04 41.91 (This study)

133F. Dincer, T. Elbir / Science of the Total Environment 374 (2007) 127–134

comprise approximately 91% of the total emissions. Thediesel railcars and the shunting locomotives comprise8% and 1% of the total emissions respectively.

The distribution of railway emissions according tothe regions is shown in Table 5. The most pollutingregion is Region-2 (Ankara) with the total emissions of91 t y−1 for HC, 240 t y−1 for CO, 1609 t y−1 for NOX,61 t y−1 for PM, 85 t y−1 for SO2 and 90 789 t y−1 forCO2. Region-1 (Istanbul) has the second highest emis-sions in 2005.

Comparing the Turkish railway emissions estimatedin this study for the year 2005 to the total amount ofemissions estimated for the traffic sector in Turkey forthe year 2005 (Muezzinoglu et al., 1998), it can be statedthat the railway emissions constitute 0.15% of HC,0.08% of CO and 4.21% of NOX emissions.

Table 6 compares the annual railway emissions inTurkey, UK, Canada, and USA. The results of Turkishrailway emissions are comparable to the emissions fromother countries. Emissions are almost similar comparedto these countries except for NOX emissions which arehigher than the other national emissions.

The 22%, 39% and 49% increase in the currentnumbers of mainline locomotives, shunting locomotivesand diesel railcars, respectively (the full capacity of rail-way network in Turkey) will increase annual emissions to431 t y−1 for HC, 1121 t y−1 for CO, 7399 t y−1 for NOX

and 342 t y−1 for PM, 552 t y−1 for SO2 and 420 256 t y−1

for CO2.Current annual capacity of railway passengers is

77 million in Turkey (TSR, 2006) and the railway trafficgrew at a slower pace of about 2% per year (WorldBank, 2005). The annual capacity will be 86.7 millionpassengers in 2010. Assuming that the increase in thenumber of passengers is the same with the increase inthe diesel consumption, a projection to estimate thediesel consumption of railway vehicles in Turkey from2005 to 2010 shows that the consumption will be158 960 m3 in 2010. This will cause a total amount ofemissions of 406 t y−1 for HC, 1121 t y−1 for CO, 7208 ty−1 for NOX, 270 t y−1 for PM, 394 t y−1 for SO2 and423 455 t y−1 for CO2.

5. Conclusion

The estimation of exhaust gas emissions (NOX, HC,CO, PM, SO2 and CO2) of railway vehicles in Turkey ispresented for the first time. The emissions from thediesel locomotives and railcars were estimated. Theestimated emissions were 384 t y−1 for HC, 1016 t y−1

for CO, 6799 t y−1 for NOX, 256 t y−1 for PM, 357 t y−1

for SO2 and 383 537 t y−1 for CO2 in 2005. After the on-

going projects are finished, passenger and freightcapacities of TSR network will be increased about10%. This will cause a total amount of emissions of406 t y−1 for HC, 1121 t y−1 for CO, 7208 t y−1 forNOX, 270 t y

−1 for PM, 394 t y−1 for SO2 and 423 455 ty−1 for CO2 in the year 2010.

Results show that the estimated emissions fromrailway network in Turkey were relatively low ascompared to on-road traffic emissions. Among the airpollutants, NOx emissions had the highest contribution(4%) to the total traffic NOx emissions.

High portion of the locomotives (58%) in the TSRfleet were manufactured before 1973, so these locomo-tives have been exempted from the EPA emission stan-dards. Since the early 2000s, TSR have been renewingtheir fleets with fuel-efficient and higher horsepowerlocomotives. Comprehensive studies will be necessaryfor the estimation of emissions from these new vehicles.

References

Aly MHF. Models to estimate emissions from railway transportsystems. Alex Eng J 2002;41:143–52.

AudigeM. Transport Sector Overview; 2006. http://wbln0018.worldbank.org/ECA/Transport.nsf/Countries/Turkey?Opendocument, November.

Bailey D, Solomon G. Pollution prevention at ports: clearing the air.Environ Impact Asses Rev 2004;24:749–74.

Department of Transportation (DOT). Transportation Statistics AnnualReport; Bureau of Transportation Statistics. Washington, DC:Research and Innovative Technology Administration; 2005.

Elbir T, Muezzinoglu A. Estimation of emission strengths of primaryair pollutants in the city of Izmir, Turkey. Atmos Environ 2004;38:1851–7.

Elbir T, Muezzinoglu A, Bayram A. Evaluation of some air pollutionindicators in Turkey. Environ Int 2000;26:5-10.

Environment Canada (EC). Locomotive Emissions MonitoringProgram 2003. Transportation Systems Division Air PollutionPrevention Directorate 2003; Report Number EPS 3/HA.

EPA. Non-road engine and vehicle emission study. Report No. EPA-21A-2001. Ann Arbor, MI: Office of Mobile Sources; 1991.

EPA. Emission factors for locomotives, technical highlights. ReportNo.EPA420-F-97-051. Ann Arbor, MI: Office of Mobile Sources;1997.

EPA. User's guide for the final NONROAD2005model. Report No.EPA420-R-05-013. Office of Transportation and Air Quality; 2005.

European Environment Agency (EEA). Specific Emissions of AirPollutants; 2006. http://themes.eea.europa.eu/Sectors_and_activities/transport/indicators/technology/TERM28,2003.09/TERM2003_28_EEA31_Specific_emissions_of_air_pollutantsfinal.pdf, November.

Fritz SG. Exhaust emissions from two intercity passenger locomotives.ASME J Eng Gas Turbine Power 1994;116:774–83.

Fritz SG, Cataldi GR. Gaseous and particulate emissions from diesellocomotive engines. Trans ASME 1991;113:370–6.

Fritz SG, Markworth VO, Cataldi GR. Exhaust emissions from in-uselocomotives. ASME Paper NO 95-ICE-4; 1995.

Icingur Y, Altiparmak D. Experimental analysis of the effects of fuelinjection pressure and fuel cetane number on direct injection dieselengine emissions. Turk J Eng Environ Sci 2003;27:291–7.

134 F. Dincer, T. Elbir / Science of the Total Environment 374 (2007) 127–134

Jorgensen MW, Sorenson SC. Estimating emissions from railwaytraffic. Int J Veh Des 1998;20:210–8.

Kean AJ, Sawyer RF, Harley RA. A fuel-based assessment of off-roaddiesel engine emissions. JAPCA 2000;50:1929–39.

Muezzinoglu A, Elbir T, Bayram A. Inventory of emissions frommajor air pollutant categories in Turkey. Environ Eng Policy1998;1:109–16.

Paukert H. Diesel emissions of European railway vehicles. Rail Int2002;33:20–9.

Plakhotnik VN, Onyshchenko YV, Yaryshkina LA. The environmentalimpacts of railway transportation in the Ukraine. Transp Res Part D2005;10:263–8.

Transport Watch UK (TWUK). Comparing Emissions by Road andRail; 2006. http://www.transwatch.co.uk/transport-pdfs/transport-fact-sheet-5a.pdf, November.

TSR. Turkish State Railways Annual Statistics 2001–2005; 2006.World Bank. Railways restructuring project. Report No: AB1186;

2005.