Emission Pollutant

8/7/2019 Emission Pollutant

1/134

Ship emissions and air pollution inDenmarkPresent situation and future scenarios

Helge Rrdam Olesen, Morten Winther,Thomas Ellermann, Jesper Christensen

og Marlene Plejdrup

National Environmental Research InstituteAarhus University

Envir on ment al Pr oject No. 1307 2009Mil jpr ojekt


2/134

T he D anish Environmental Protection Agency will, when opportun ity

offers, publish reports and contributions relating to environmental

research and development projects financed via the Danish EPA.

Please note that publication does not signify that the contents of the

reports necessarily reflect the views of the Danish EPA.

T he reports are, however, published because the D anish EPA finds that

the studies represent a valuable contribution to the debate on

environmental policy in D enmark.


3/134

3

Contents

CONTENTS 3PREFACE 7SAMMENFATN ING OG KONKLUSIONER 9

BAGGRUND OG FORM L 9UNDERSGELSEN 10HOVEDKONKLUSIONER 10PROJEKTRESULTATER 12

Den AIS -baserede emissionsopgrelse 12Forudstninger for scenarierne 13

Emissioner i henhold til scenarierne 13Koncentrationer af svovldioxid 14Koncentrationer af NO

216

Koncentrationer af partik ler 17

SUMMARY AND CON CLUSIONS 19BACKGROUN D AND OBJECT IVES 19T HE STUDY 19MAIN CONC LUSIONS 20PROJECT RESULT S 22

T he AIS-based emission inventory 22 Assumptions for scenarios

Emissions according to the scenarios 23Concentrations of sulphur dioxide 24Concentrations of nitrogen dioxide (NO

2) 26

Concentrations of particles 27

1 INTRODUCTION 291.1 BACKGROUND 291.2 AIS DATA 301.3 N EW REGULATIONS AND PROJECT IONS 311.4 RELAT ION T O PREVIOUS STUD IES 321.5 ST RUCT URE OF THE REPORT 33

2 EMISSIONS 352.1 INTRODUCTION 352.2 AIS DAT A PROVIDED BY D AMSA 352.3 VESSEL D ATA PROVIDED BY LLOYDS 372.4 ENGINE T YPES, FUEL T YPES AND AVERAGE ENG INE LIFE TIM ES 372.5 T RAFFIC FORECAST 372.6 ENGINE LOAD FUNCT IONS 38

2.6.1 Auxiliary power in use 392.7 FUEL CONSUMPT ION AND EMISSION FACT ORS 40

2.7.1 Specific fuel consumption 402.7.2 NO

xemission factors 41

2.7.3 SO2 452.7.4 PM 452.7.5 CO and VO C 46


4/134

4

2.7.6 CO2

472.8 CALCULAT ION PROCEDURE 472.9 RESULTS FOR FUEL CON SUMPTION AND EMISSION 472.10 RESULTS:SPATIAL DISTRIBUTION 56

3 SHIP EMISSION INVENT ORIES FOR AIR POLLUT IONMODELLING 61

3.1 INTRODUCTION 613.2 "EMEP-REF" INVENTORY 623.3 "AIS-2007"AND "EMEP-2007" INVENTORIES 643.4 "AIS-SP" INVENTORY 673.5 "AIS-2011" INVENTORY 683.6 "AIS-2020" INVENTORY 693.7 SUMM ARY OF DIFFERENC ES 70

4 AIR POLLUT ION MOD EL CALCULAT IONS 734.1 INTRODUCTION 734.2 INTERPRETING PM 2.5 RESULTS 74

4.3 DESCRIPT ION OF THE MOD EL

744.4 MODEL RUNS AND EMISSION INVENTORIES 764.5 RESULTS:CONCENT RAT ION LEVELS 82

4.5.1 Spatial resolution 834.5.2 SO

2concentration levels by region 86

4.5.3 NO2

concentration levels by region 874.5.4 PM

2.5concentration levels by region 87

4.6 CONT RIBUT ION FROM SHIPS 884.7 COMPARISON WIT H M EASUREMENT S 904.8 SCENARIOS FOR 2011 AND 2020 91

4.8.1 SO2

concentration levels - future scenarios 914.8.2 NO

2concentration levels - future scenarios 92

4.8.3 PM2.5 concentration levels - future scenarios 934.8.4 Contribution from ships according to the scenarios 96

4.9 CONCENT RAT ION LEVELS IN COPENHAGEN 98

5 POLLUT ION FROM SHIPS IN PORT S 1015.1 INTRODUCTION 1015.2 AVAILABLE ST UD IES 1015.3 SULPHUR REGULATION 1025.4 METHODOLOGY:EMISSION INVENTORIES FOR PORTS 103

5.4.1 Activities: At dock, pumping and manoeuvring 1035.4.2 Emission factors

5.5 RESULTS 106

5.6 RESULTS FOR COPENHAGEN 1065.7 CRUISE SHIPS AND AIR QUALITY 109

5.7.1 Cruise ships: NOX

111 5.7.2 Cruise ships: particles and SO

2114

5.8 H IGH RISE BUILDING S 1145.9 RESULTS:T HE PORT OF AARHUS 1155.10 CONCLUSIONS CON CERNING SHIPS IN PORT 118

6 CONCLUSIONS 1217 REFERENCES 123


5/134

5

ANNEXESANNEX A: SHIP CATEGORIES 127ANNEX B: NOX EMISSION FACTORS FOR 2007,2011 AND 2020

FOR SHIP CAT EGORY/ENGINE T YPE COMBINAT IONS 129ANNEX C: EMISSIONS FROM LAND-BASED SOURCES USED FOR

T HE MODEL CALCULATIONS. 131


6/134

6


7/134

7

Preface

T he present report docum ents the results of the project "Contribution fromships to air pollution in Denmark" (in Danish: "Skibsfartens bidrag tilluftforurening i Danmark").

T his project has been carried out on behalf of the Danish EnvironmentalAgency by the N ational Environmental Research Institute (NERI - in D anish

Danmarks Miljundersgelser), which is part of Aarhus University (AU).

T he D anish Environmental Protection Agency has formed a partnership withthe Danish Ship-owners' Association (Danmarks Rederiforening), aimed atcleaner shipping (Partnerskab for renere skibsfart). The current project hasstatus as one of the elements in this partnership.

T he steering group for the project consists of the following individuals: DorteKubel, Jesper Stubkjr and Christian Lange Fogh from the EnvironmentalProtection Agency, Jesper Christensen, Thomas Ellermann, Helge RrdamOlesen and Morten Winther from the N ational Environmental ResearchInstitute, Aarhus University.


8/134

8


9/134

9

Sammenfatning og konklusioner

Baggr und o g f or ml

Skibstrafikken i de danske farvande anses for at spille en betragtelig rolle forluftkvaliteten i de danske byer og i Danmark generelt. DanmarksMiljundersgelser har tidligere opstillet estimater for skibstrafikkensbetydning for luftforurening (bl.a. Olesen et al., 2008) . D isse tidligereestimater er imidlertid behftet med stor usikkerhed. De er baseret pemissionsopgrelser (opgrelser af ud ledninger) m ed ringe geografiskoplsning (50 km x 50 km). Endvidere har de hidtidige emissionsopgrelservret baseret p an tagelser, der ikke ndvend igvis afspejler skibstrafikkensreelle sejlmnstre og omfang.

Siden 2006 er skibstrafikkens bevgelser i danske farvande imidlertid blevetregistreret i det skaldte AIS-system. AIS str for Automatic IdentificationSystem. Systemet indebrer, at alle skibe over 300 bruttotons er forp ligtet tilat medfre en skaldt transponder, der sender information om skibets identitetog position til modtagerstationer p land. Registreringen har gjort det muligtat kortlgge skibsemissioner langt mere detaljeret end tidligere, og dennemulighed er nu blevet udnyttet til at udarbejde en ny emissionsopgrelse forskibstrafikken omkring D anmark.

Den ny AIS-baserede opgrelse af skibsemissionerne har sammen med enlang rkke andre data dannet grundlag for nye modelberegninger af

luftkvalitet i Danmark. Formlet med dette arbejde har vret at vurderebidraget fra skibe til koncentrationen i luften af en rkke forurenende stoffer.Koncentrationsberegningerne er foretaget med en ny version afluftforureningsmodellen D EHM (D anish Eulerian Hemispheric M odel), somhar strre geografisk detaljeringsgrad end den version, som blev anvendt afOlesen et al. (2008).

Den internationale sfartsorganisation (IMO) har vedtaget regler, som sigtermod at reducere skibstrafikkens forurening med svovldioxid (SO 2) ogkvlstofoxider (N OX) i rene frem til 2020. Den foreliggende undersgelsehar ogs haft til hensigt at belyse effekten af disse regler for luftkvaliteten i

Danmark ved at gennemfre scenarioberegninger af luftkvalitet for 2020baseret p de forventede reduktioner af emissionerne.

Der forventes ogs reduktioner i de landbaserede emissioner af SO 2, N O x ogpartikler i perioden frem til 2020. De estimerede reduktioner af delandbaserede emissioner er inddraget i scenarioberegningerne for 2020.

Som en mindre del af undersgelsen er der gennemfrt opdateringer afforskellige tidligere vurderinger af den lokale luftforurening fra skibe i havn.


10/134

10

Under sgel sen

Undersgelsen er gennemfrt af Danmarks Miljundersgelser ved AarhusUniversitet. Et hovedelement i undersgelsen er at der er etableret en ny ogforbedret opgrelse af skibemissionerne i farvandene omkring Danmark.Denne op grelse af skibsemissioner er herefter anvendt til modelberegning afluftkvalitet i Danmark for r 2007. Der er ogs gennemfrt beregninger

baseret p den gamle emissionsopgrelse. Herved kan kvalitetsforbedring afopgrelsen af skibsemissioner vurderes.

Der er endvidere lavet scenarioberegninger af luftkvalitet i 2011 og 2020baseret p forventede reduktioner i skibsudledningerne og et estimat for delandbaserede udledninger i 2020, sledes at effekten af reguleringerne afudledningerne kan vurderes.

Endelig er der udarbejdet vurderinger af den lokale indflydelse afskibeemissionerne p luftkvalitet i danske havne baseret p opdatering aftidligere undersgelser.

Hovedkonkl usion er

Vedrrende emissioner: Et af projektets hovedresultater er, at der nu foreligger en ny og

forbedret emissionsopgrelse over nationale og internationaleskibsemissioner i de danske farvande. Op grelsen er udarbejdet igitterfelter p 1 km x 1 km, hvilket giver en god geografisk beskrivelse afemissionerne i de danske farvande. T idligere opgrelser fra det flleseuropiske luftovervgningsprogram EMEP benyttede gitterfelter p 50x 50 km. Sammenligning mellem mleresultater og modelberegninger

p basis af den gamle (EMEP) og nye emissionsopgrelse (D MU)viser, at den nye opgrelse er vsentligt bedre end den tidligereopgrelse.

Fra 2007 til 2020 forudses der i farvandene omkring Danmark enreduktion i emissionen af svovldioxid fra skibstrafikken p 91% p trodsaf get trafikmngde. Dette skyldes IMO-kravene.

I samme periode forudses en svag stigning i den absolutte emission afkvlstofoxider (N OX) fra skibstrafik, nemlig med 2%. Uden de IMO-krav for NOX, der forventes at blive indfaset i lbet af perioden , villestigningen have vret 15%, foranlediget af get skibstrafik. IM O-kravene betyder store NO X-reduktioner for nye skibe fra 2016, sskibsflden vil gradvis f mindskede NOX-emissioner i rene derefter.

Vedrrende koncentrationer: Siden 1990 er der sket en vsentlig reduktion i SO2-emission fra

landbaserede kilder, og i rene fremover forventes der en kraftigreduktion fra skibstrafik, samtidig med at der forventes en fortsatreduktion til lands. P basis af scenarioberegningerne forventes SO 2-koncentrationen i 2020 som m iddelvrdi i Danmark at vre nede pblot 0,3 g/m3, hvilket er 6% af 1990-niveauet og udgr 1,5% af EU-grnsevrdien. I 2007 kommer omkring 33% af SO 2-koncentrationen iDanmark fra skibsemissioner, m ens dette tal forventes at falde til kun11% i 2020.

I tiden frem til 2020 forventes koncentrationen af NO 2 i by-baggrundeni Kbenhavn at aftage fra ca. 16 til ca. 9 g/m3. Faldet skyldes, at der


11/134

11

forventes reduktioner i NO X-emissionen fra landbaserede kilder.Derimod bidrager skibstrafikken i absolutte tal nogenlunde lige meget nuog i 2020, fordi get skibstrafik og skrpede emissionskrav opvejerhinanden. For Danmark som helhed er NO 2-koncentrationsniveauetvsentligt lavere end i Kbenhavn, og for D anmark forventes et fald fraca. 5,5 til ca. 3,5 g/m3 fra nu og frem til 2020. I jeblikket kommer ca.21% af NO

2

fra skibstrafik, men p rocentdelen vil stige frem til 2020grundet reduktionen for de landbaserede kilder.

Hvad angr koncentrationen af fine partikler (PM 2.5) s er en betydeligdel af partikelmngden af ukendt oprindelse. Dette er internationaltanerkendt som vrende state-of-the-art. Hvad angr den del, som man eri stand til at redegre for (her betegnet mPM

2.5for modelberegnet PM 2.5;

se afsnittet Koncentrationer af partiklerher i sammenfatningen fornrmere forklaring), s stammer ca. 1 g/m3 fra skibe svel iKbenhavn som generelt p landsplan. Beregningerne tyder p et fald iskibes bidrag med ca. 0,2 g/m3 i tiden frem til 2020. H vad skibes

procentvise bidrag til partikelforurening angr, s afhnger talvrdien af,hvilket sted man betragter, og hvorvidt man stter det identificerede

skibsbidrag i forhold til mPM 2.5 eller den totale koncentration af PM 2.5. Iforhold til den totale koncentration af PM 2.5 vil det p rocentvise bidrag if.eks. by-baggrund i Kbenhavn vre af strrelsesordenen 7%.

Sammenligning mellem resultater fra beregning med den gamle(EM EP) og nye emissionsopgrelse (D MU) for skibsudledninger for2007 viser, at den nye opgrelse giver anledning til et fald i deberegnede koncentrationer p 46% for SO 2, 14% for NO 2 og 10% formPM 2.5 set i forhold til beregninger med den gam le opgrelse.

En strre andel af skibstrafikken i danske farvande sejler gennemresund end antaget ved den gamle emissionsopgrelser (EM EP) .Udledningen fra skibe i resund er dermed hjere end tidligere

vurderet. Til gengld er udledningen fra skibe i Storeblt samt Kattegatmarkant lavere end h idtil vurderet. For svovldioxid betyder det, at hvorudslippet i Storeblt tidligere blev vurderet at vre ca. 11 gange s stortsom udslippet i resund, vurderes det nu blot at vre 2,3 gange sstort. Taget under et er udslippene fra farvandene omkring Sjllandlavere end tidligere antaget, og det resulterende n iveau af svovldioxid iluften er lavere end tidligere vurderet gennemsnittet forHovedstadsregionen er sledes beregnet til 1 g/m3 mod tidligere 1,6g/m3.

Vedrrende skibe i havn: I Kbenhavns H avn er krydstogtskibe ansvarlige for 55% af NOX-

emissionen fra samtlige skibe. Krydstogtskibene giver dog ikkeanledning til problemer med at overholde NO 2-grnsevrdier omkringkajpladserne, og giver ej heller p roblemer i relation til grnsevrdier forPM 2.5 og SO 2.

Dog glder, at hvis der er tale om forholdene i stor hjde ved hjhusett ved kaj, s m man udfre mere detaljerede undersgelser for atkunne afgre, om der potentielt er problemer.

Afsnit 5 .10 angiver nogle yderligere konklusioner vedrrende lokalluftforurening fra skibe i havn.


12/134

12

Pr ojekt r esul t at er

Den AIS-baserede emissionsopgrelseDer er udarbejdet en ny, AIS-baseret, emissionsopgrelse for skibstrafik medstor detaljeringsgrad, bde hvad angr den geografiske oplsning, og hvadangr information om forureningen fra de enkelte skibe. Som eksempel viser

Figur 1( tv.) den detaljerede emissionsopgrelse for SO 2 med en oplsning p1 x 1 km. Ved modelberegningerne af koncentrationer er imidlertid benytteten knap s fin oplsning, nemlig p 6 x 6 km, svarende til modellensdetaljeringsgrad.

T il sammenligning viser figuren ogs SO 2-emissioner iht. den ldreemissionsopgrelse fra EMEP (2008), der hidtil har dannet grundlag fortilsvarende modelberegninger, og som har en grov oplsning p 50 x 50 km.Den tidligere opgrelse har bl.a. vret benyttet ved beregningerne iforbindelse med den danske luftkvalitetsovervgning under NOVANA(Ellermann et al., 2009; Kemp et al., 2008), hvor der dog ikke var specieltfokus p skibstrafik. Den t idligere opgrelse benvnesEM EP-ref(se kapitel 3og 4). Det fremgr af Figur 1, at den ny AIS-baserede opgrelse i detaljerviser, hvordan skibsruterne fremstr som veldefinerede "veje" til ss, mens degamle, grovmaskede data indebrer et mere mudret billede, hvor der ogsforekommer emission fra skibe indenlands, fordi de store gitterfelter dkkerbde vand og land.

En rkke andre forskelle er mindre synlige, men betyder noget foremissionerne. I de gamle data b lev der regnet med, at skibene sejler medkonstant belastning af motoren, mens den ny metode tager hensyn til at der erforskelle i emissionerne, afhngigt af skibets hastighed. Den gamle metode

byggede p antagelserom sejlruten mellem havnene, mens AIS-metoden tagerudgangspunkt i de faktiske sejlruter, baseret p data fra 24 dgn i 2007.

Figur 1 Skibsemissio ner af SO2 pr km2. Til venst r e i henh ol d t il den ny, AIS-baser ede

emissionsopgrel se med detal jer in gsgr ad p 1 x 1 km. Til hjr e i h enh ol d t il dent idl igere benyt t ede, hvor emissionen f or el igger i f el t er p 50 x 50 km.


13/134

13

Den ny metode anses for et vsentligt fremskridt i forhold til den tidligere;dog er der fortsat mu lighed for forbedringer f.eks. udvidelse af omrdet medAIS-signaler og inddragelse af en strre mngde AIS-data end de 24 dage,der nu udgr basis for beregningerne.

Forudstninger for scenarierneDe vsentligste forudstninger for scenarie-beregningerne er som flger.

De vedtagne IMO reguleringer for svovl og NO X gennemfres somplanlagt. Specielt er det forudsat, at farvandene omkring Danmarkudpeges til en skaldt NO

XEmission Control Area (N ECA) som

defineret af IM O. sters-landene forbereder en ansgning herom.Status som N ECA indebrer skrpede krav om N OX-udledninger,specielt for nye skibe efter 2016.

2020-scenariet bygger p en forudstning om undret skibstrafik fremtil 2011 og derefter en stigning p 3,5% pr r for fragtskibe i perioden

2011-2020. Der antages undret passagertrafik. Hvad angr landbaserede emissioner er benyttet et specifikt stforudstninger. Der forventes i nr fremtid fremsat forslag til nytinternationalt direktiv om nationale emissionslofter (NEC-direktivet),som skal vre overholdt i 2020. Direktivet er imidlertid ikke fremlagtendnu, s derfor anvendes et scenario for udledninger fra delandbaserede kilder i Europa i 2020, der er sammensat ud fra EUstemastrategi for ren luft i Europa (se bl.a. Bach et al., 2006) og scenarierfor EU-27 udarbejdet af International Institute for Applied SystemsAnalysis (Amann et al., 2008), som led i det forberedende arbejde iforbindelse med et nyt NEC-direktiv. Detaljer findes i afsnit 4.4.

Emissioner i henhold ti l scenarierneFigur 2 viser brndstofforbrug samt emissioner fra skibstrafik for de trescenarier 2007, 2011 og 2020 i farvandene omkring D anmark - nrmerebestemt det omrde, der er farvet i Figur 1 tv.

Brndstofforbrug

0

1020

30

40

50

60

70

80

90

100

110

120

130

2007 2011 2020

(PJ)

NOX emissioner

0

20000

40000

60000

80000

100000

120000

140000

160000

180000

200000

2007 2011 2020

(tons)

SO2 emissioner

0

10000

20000

30000

40000

50000

60000

70000

2007 2011 2020

(tons)

Emissioner af primr PM2.5

0

1000

2000

3000

4000

5000

6000

2007 2011 2020

(tons)

Figu r 2 Br ndst of f or br ug (Pet ajou l e) samt emissioner f r a skibst r af ik i far vandene

omkr in g Danmark (jvn f . Figur 1) f or de tr e scenari er 2007, 2011 og 2020.


14/134

14

Der beregnes en vkst p 15% i brndstofforbruget fra 2007 til 2020, hvilketer et resultat af get trafikmngde (36% forgelse for fragtskibe og undretpassagertrafik), kombineret med at motorerne i skibsflden gradvis blivermere effektive.

For kvlstofoxider (NO X) forventes en svag stigning i emissionen p 2% fra2007 til 2020. Uden strengere krav til emissionerne ville stigningen havesvaret til forgelsen i brndstofforbrug (15%). Resultatet af IMO-restriktionerne er alts en gennemsnitlig reduktion i NO X-emissionerne p11%per kg brndstof. Sdanne reduktioner fortstter efter 2020.For svovldioxid (SO 2) forudses fra 2007 til 2020 en reduktion p 91%, mensreduktionen for pr imr PM 2.5 er 54%. Primre part ikler udgr kun en lille delaf den mngde partikler, man finder i luften (se afsnittet Koncentrationer af

partiklerher i sammenfatningen for forklaring af primre og sekundrepartikler).

De mest markante reduktioner i skibsemissionerne sker for SO 2. Dissereduktioner illustreres p kortene i Figur 3. K ortene giver et overblik over

skibstrafikken omkring D anmark. Meget synlige er hovedskibsruterne fra denBotniske Bugt til Nordsen, de strre indenlandske frgeruter, samtfrgeruterne der forbinder Danmark, Sverige, Tyskland og Polen.

Koncentrationer af svovldioxidHvad angr forurening med SO 2 har skibstrafik hidtil vret en meget markantkilde. Fra ca. 2007 har skibe i stersen og Nordsen vret underlagt et kravom at svovlindholdet i olien hjst m vre 1,5%, mens det tidligere typisk var2,7% (sektion 3.1) . Skibstrafik fremstr dog stadig som en vsentlig kilde,sammenlignet med kilder p landjorden. Figur 4 viser beregnedekoncentrationer af SO 2, dels for en 2007-situation ( med 1,5% svovl), dels forsituationen i 2020. I 2020 vil svovlkravet vre betydeligt skrpet, nemlig til0,1%, og det afspejler sig tydeligt i koncentrationen i luften.

Siden 1990 er der sket en vsentlig reduktion af SO 2-emission fralandbaserede kilder, og reduktionen forventes at fortstte. Resultatet vil i

Figur 3 Kor t o ver esti mer ede emission er af SO2 f r a skibst r afi k i hver t af rene 2007, 2011 og 2020. Enheden erSO2 pr . km

2.


15/134

15

2020 vre, at SO 2-koncentrationen som middelvrdi i Danmark vil vre nedep blot 0,3 g/m3, hvilket er 6% af 1990-niveauet.

Figur 5 viser det relative bidrag fra skibstrafik - i forhold til samtlige kilder - tilkoncentrationsniveauet af SO 2 samt flere andre forureningskomponenter. D erer tale om et gennemsnit over hele Danmark. Figuren viser resultater fra tost beregninger: Dels beregninger p grundlag af den nye, mere prciseemissionsopgrelse (betegnet AIS-2007), dels p grundlag af den ldreopgrelse fra EMEP (betegnetEM EP-ref).

Den ny opgrelse adskiller sig fra den gamle p flere mder: den forudstter

at svovlprocenten i brndslet er 1,5%, mens den gamle refererede tilsituationen fr 2007 og benyttede 2,7%; den ny har hjere geografiskoplsning; den er baseret p mere korrekte skibsruter end den gamle; og denbruger mere realistiske data for skibenes m otorbelastning. Alle disse faktorertilsammen betyder, at den ny opgrelse frer til lavere emissioner.Reduktionen er mest markant for svovldioxid og primr PM 2.5, men gr sigogs gldende for de andre forureningskomponenter. Hvor det p grundlag afden ldre emissionsopgrelse fra EM EP blev vurderet at skibe stod for 68% afSO 2-niveauet i Danmark, er andelen nede p 37% i henhold til den nyopgrelse.

Figur 4 Ber egnede kon cent r at ioner af SO2 i g/ m3. Til venstr e er vist sit uat ion en i 2007, til hjr e sit uat ion en i

2020. For 2007-sit uati on en ses skibst r af ikkens indf l ydel se t ydel igt ; f or 2020-scenar iet r esterer der nstenint et synl igt bidr ag f r a skibst r af ik.


16/134

16

Koncentrationer af NO2

Det er ved beregningerne forudsat, at farvandene omkring D anmark udpegestil et skaldtNOX

Emission Control Area, som defineret af IMO , hvor derglder skrpede NOX-krav. Det mest markante krav glder nye skibe fra2016, hvor der krves 80% reduktion af NO X-udledningen. D er er andre,mindre vidtgende krav til ldre skibe.

Resultatet af beregningerne er, at IMO-kravene ikke formr at opveje denforudsete stigning i skibstrafikken. I r 2020 forudses stort set samme n iveauaf NOX-emissioner fra skibsfart som i 2007-situationen. Efter 2020 vil IMO -kravene formentlig ytre sig i lavere skibsemissioner, men det er ikke tilfldet i2020. D erimod sker der markante ndringer p land i henhold til detovennvnte scenario for emissionerne i 2020.

Resultatet ses i Figur 6. NO X leder til dannelse af NO 2, som der gldersundhedsrelaterede grnsevrdier for. Figur 6 viser beregnedekoncentrationer af NO 2, dels for en 2007-situation, dels for situationen i 2020.De store skibsruter forbliver srdeles synlige, mens koncentrationen plandjorden aftager. Andelen af skibes bidrag til N O 2-koncentrationsniveauet eri 2007 21%, mens den vil stige til 34% i 2020 - fordi de vrige kilders emissionreduceres.

Det er et meget interessant resultat af beregningerne, at bybaggrundsniveauetaf NO 2 i Kbenhavn forventes at falde med ca. 7 g/m

3 i tiden fra 2007 til2020. NO

2

-niveauerne p mange kbenhavnske gadestrkninger overstigerEU's grnsevrdi p 40 g/m3, og i lyset heraf er et s stort fald ibaggrundsniveauet interessant.

0

10

20

30

40

50

60

70

80

SO2 NO2 Primary PM2.5 mPM2.5

%

froms

hipemissions

EMEP-ref

AIS-2007

Figur 5 Rel at ivt bidr ag fr a skibe t il koncent r at ion sni veauet i l uf t en af f or skel l igef or ur enende sto f f er . Far ven p sjl er ne viser , hvil ken emissionsopgrel se der l iggert il gr und f or de ber egnede kon cent r at io ner. De l ysebl sjl er bet jener sig af den n ye,mer e prcise opgr el se ("AIS-2007"). De mr kebl af den l dr e emissio nsopgr el se f r aEMEP ("EMEP-ref ") .


17/134

17

Koncentrationer af partiklerFine partikler med en diameter p mindre end 2,5 mikrometer betegnes somPM 2.5. M an kan yderligere sondre mellem forskellige typer partikler, derindgr i PM 2.5. Primre partiklerforefindes som partikler umiddelbart efter atde har forladt kilden. Sekundre partiklerer derimod p artikler, der ikke er"fdt" som partikler, men er dannet ved om dannelse af gasarter - typisk mangetimer efter at forureningen er sendt ud i atmosfren. De sekundre partiklerkan yderligere underopdeles i sekundre uorganiske og sekundre organiske.

Skibes udledning af SO 2 og N OX giver anledning til dannelse af sekundreuorganiske partikler, og hele denne mekanisme indgr i den benyttedeluftforureningsmodel, D EHM. Hvad angr de sekundre organiske partiklerhar man ikke tilstrkkelig viden t il at kunne beskrive dem modelmssigt.Dette glder ogs p internationalt niveau (se bl.a. Yttri et al., 2009). Derforbeskriver modellen kun en del af den m ngde partikler, man finder iatmosfren, nemlig primre og sekundre uorganiske partikler. I rapportenomtales denne del som mPM

2.5.

Figur 7 viser beregnede koncentrationer afmPM2.5

. PM 2.5 kan transporteresover store afstande, mens SO 2 og NO 2 ikke har lige s lang levetid. Derforadskiller det geografiske mnster i Figur 7 sig markant fra mnstret i deforegende figurer. Skibenes forurening er ikke synlig, mens det er markant, atforurening fra det centrale Europa gradvis klinger af hen over Danmark.

Skibstrafikkens bidrag til mPM 2.5 er 18%, mens bidraget set i forhold tilatmosfrens totale PM 2.5-niveau er en del mindre. Det vurderes at i forhold tiltotal PM 2.5 i Kbenhavn er skibstrafiks bidrag omkring 4-7%, afhngigt af omman betragter luften i en trafikeret gade eller i by-baggrund (dvs. vk fra selvegaden).

Frem til 2020 vil der ske en reduktion af det generelle niveau af mPM 2.5 overDanmark p lidt over 2 g/m3, hvilket hovedsagelig skyldes reduktion ilandbaserede emissioner.

Figur 6 Ber egnede kon cent r at io ner af NO2 i g/ m3. Til venstr e er vist sit uat ion en i 2007, til hjr e sit uat ion en

i 2020. Der sker nsten i ngen ndr in g i udsl ippet af NOx f r a skibst r af ik mel l em de t o t idspunkt er , mens der ermar kant e r edukti on er f or l andbaser ede kil der .


18/134

18

Figur 7 Ber egnede kon cent r at ioner i g/ m3 af mPM2.5, der er den del af part ikel massen, som kan ber egn es afl uf t f or ur eni ngsmodel l en. Der er t al e om pr imr e par t ikl er samt om sekundr e uor ganiske par t ikl er , mens manved f or l idt om sekundr e organiskepar t ikl er t il at kunn e beskr ive dem model mssigt . Til venst r e er vistsit uat ion en i 2007, ti l hjr e sit uat ion en i 2020.


19/134


20/134

20

(EMEP, 2008) emission inventories have been applied for model calculationsof air quality in Denmark, thus allowing an assessment of the effect of therevised inventory.

Furthermore, scenario calculations for 2011 and 2020 have been carried out,in order to evaluate the effect of the IM O regulations. T he scenariocalculations have been based on expected reductions in ship emissions and anestimate for land-based emissions in 2020.

Finally, the contr ibution to local air pollution from ships at ports has beenassessed in various ways, based on updates of previous stud ies.

Main concl usion s

Concerning emissions: One of the main results of the project is that a new, improved emission

inventory for national and international ship emissions in Danish marine

waters has been established. T he inventory has a spatial resolution of 1x 1 km. Previous inventories from the European Monitoringprogramme EM EP applied a grid size of 50 x 50 km. A comparisonbetween the measured concentrations and model results based on theold (EMEP) inventory and the new (N ERI) inventory shows that thenew inventory is a considerable improvement.

Between 2007 and 2020 an emission reduction as large as 91% isenvisaged in the marine waters around Denmark, in respect to sulphurdioxide from ship traffic - despite an increase of traffic. T his is due tothe IMO regulations.

Within the same period a marginal increase is expected in totalemissions of nitrogen oxides (NOX) from ship traffic, namely by 2%

from 2007 to 2020. Without stricter emission standards the increasewould have corresponded to the increase in fuel consumption, i.e. 15%.T he IMO requirements imply large NO X reductions for new ships from2016. As a consequence, the ship fleet will gradually experience areduction in its average NOX emission factor from 2016 onwards; thisdevelopment will continue after 2020.

Concerning concentrations: Since 1990, SO 2 emissions from land-based sources have been

substantially reduced, and in the years to come a continued largereduction from ship traffic is expected, along with an expected

reduction for land-based sources. Based on scenario calculations theSO 2 concentration level as an average for D enmark will decreaseconsiderably in the period up to 2020, so in 2020 it will reach a level of0.3 g/m3, which is only 6% of what it was in 1990, and corresponds to1.5% of the EU limit values. In 2007 around 33% of the SO 2concentration level in Denmark was due to ship emissions, this numberwill be reduced to about 11% in 2020.

T he concentration of nitrogen dioxide (N O 2) in urban background airin Copenhagen is expected to be reduced from 16 to 9 g/m3 in theperiod up to 2020. T he reduction is due to expected reductions in N OXemissions from land-based sources. For ship traffic, however, thecontr ibution is essentially unchanged, because increase in ship traffic

and stricter emission requirements balance each other. For D enmark asa whole the N O 2 concentration level is considerably lower than in


21/134

21

Copenhagen, and it will decrease from 5.5 to around 3.5 g/m3 from2007 up to 2020. Presently, as an average for Denmark, 21% of NO 2can be at tributed to ship traffic, but the relative share from ships willincrease in the years up to 2020 due to reductions for the land-basedsources.

A considerable amount of fine particles (PM 2.5) is of unknown origin.Internationally, this is recognised as being state-of-the-art. T he fractionthat can be explained and modelled is here designated mPM

2.5

(modelled PM 2.5, see the section Concentrations of particles here in thesummary for further explanation). Around 1 g/m 3 of the mPM 2.5 levelcan be attributed to ships, both in Copenhagen and generally inDenmark. Calculations point to a slight decrease by approximately 0.2g/m3 during the period to 2020. If one wishes to express the shipcontribution in terms of percent, the fraction attributable to ships willdepend on the location considered, as well as whether the mPM 2.5contribution is compared to the value for mPM 2.5 or for total PM 2.5.Compared to total PM 2.5, the percent wise contribution in urbanbackground air in Copenhagen will be on the order of 7%.

A comparison between concentration results based on the previous(EM EP) and the new (N ERI) inventories show that the new inventoryresults in concentrations that are lower by, respectively, 46% for SO 2,14% for NO 2 and 10% for mPM 2.5 (average concentrations overDenmark).

According to the new inventory the amount of ship traffic through theresund is larger than according to the previous (EMEP). As aconsequence, emissions in the resund are larger than previouslyassessed. On the other hand, in the Storeblt and Kattegat shipemissions are substantially smaller than previously estimated. In the caseof sulphur dioxide, the emission in Storeblt was previously assessed to

be 11 times larger than that of resund; now, the emission in Storebltis assessed to be only 2.3 times larger. Altogether, the emissions aroundZealand are smaller than previously assessed, and the resulting level ofsulphur dioxide in ambient air is smaller than previously assessed. Forthe C openhagen region the average SO 2 concentration is estimated at 1g/m3, to be contrasted with the previous estimate of 1.6 g/m3.

Concerning ships at port: In the port of Copenhagen cruise ships are responsible for 55% of the

total NOX emission from ships at port. Anyhow, this does not lead toproblems related to N O 2 limit values around the docks, and neither doesit create problems with respect to limit values for PM 2.5 and SO 2.

When considering conditions at high rise buildings close to berths withvery heavy ship traffic, there may potentially be problems with N O 2exceedances very close to the berth (within the nearest hundred or fewhundred meters). In such cases it can be wise to conduct detailedstudies.

Section 5.10 of the report indicates some further conclusionsconcerning local air pollution from ships at port.


22/134

22

Pr oject r esul t s

The AIS-based emission inventoryThe project has resulted in a new, AIS-based emission inventory for shiptraffic in the Danish marine waters. It is based on detailed information, bothin respect to geographical resolution, and to the underlying information on the

pollution from the individual ships. As an examp le Fig. 1 shows the detailedemission inventory for SO 2 with a resolution of 1 x 1 km. For the modellingpurposes, however, a spatial resolution of 6 x 6 km is used, corresponding tothe resolution of the air pollution model.

For comparison, Fig. 1 also shows SO 2 emissions according to the previousinventory from EM EP (2008) , which has been the base for model calculationsuntil now, and which has a crude resolution of 50 x 50 km. T he previousinventory has been used for calculations in the framework of the Danishmonitoring programme N OVANA (Ellermann et al., 2009; Kemp et al.,

2008) - which, however, did not have particular focus on ship t raffic. T heprevious inventory is referred to as EM EP-ref(se Chapters 3 and 4).

It appears from Fig. 1 that the new AIS-based inventory in detail shows howthe ship rou tes appear as well-defined 'roads' at sea, as opposed to the oldmethod, which has grid cells extending over both sea and land, and thusplaces some emissions from ships over land.

Other d ifferences are less visible, but have consequences for the emissions:The previous inventory assumed that the engine load for ships was constant,whereas the new method takes account of differences in emissions, depending

on the speed of the ship. T he old method app lied assumptions concerning theship routes, whereas the AIS method is based on actual routes, based on asample of 24 days in 2007.

Fig. 1 Sh ip emissions of SO2 pr km2. The l ef t panel displ ays val ues f r om the new, AIS-based emission in vent or y wit h a resol ut ion of 1 x 1 km. The r igh t panel il l ust r at es th epr eviousl y used invent or y, wher e t he emission is assigned t o g r id cel l s of size 50 x 50km (EMEP, 2008).


23/134

23

T he new method is regarded as a considerable improvement compared to theprevious; however, there is still room for improvement e.g., extension of thearea with AIS signals, and use of a larger amount of AIS data than the 24 dayswhich form the basis for the present calculations.

Assumptions for scenariosT he main assumptions underlying the scenario calculations are as follows.

The IMO regulations for sulphur and NO X are implemented asplanned. In particular, it is assumed that the marine waters aroundDenmark will be designated aNO

XEmission Control Area (N ECA) as

defined by IM O. T he countries around the Baltic Sea are preparing anapplication with this purpose. T he status as N ECA implies additionalNOX emission restrictions for new ships from 2016.

Concerning the amoun t of ship traffic, an annual increase of 3.5 percenthas been assumed for transport of goods from 2011 and onwards, whilepassenger traffic is assumed unchanged.

Concerning land-based emissions, a specific set of assumptions havebeen used. It is expected that in near future a new EU directive onnational emission ceilings for 2020 ( NEC directive) will be negotiated.However, there is not yet an official prop osal for the directive.Therefore a scenario for emissions from land-based sources in Europein 2020 has been set up, which is based on emission scenarios preparedin connection with EU's thematic strategy for clean air in Europe (seee.g. Bach et al., 2006), and scenarios for EU-27, p repared by theInternational Institute for Applied Systems Analysis (Amann et al.,2008) as part of the preparatory work for a new NEC directive. Furtherdetails are found in Section 4.4.

Emissions according to the scenariosFig. 2 displays the fuel consumption and emissions from ship traffic for thethree scenarios: 2007, 2011 and 2020. The values refer to the area which iscolored in Fig. 1(left).

Fuel consumption

0

1020

30

40

50

60

70

80

90

100

110

120

130

2007 2011 2020

(PJ)

NOX emissions

0

20000

40000

60000

80000

100000

120000

140000

160000

180000

200000

2007 2011 2020

(tons)

SO2 emissions

0

10000

20000

30000

40000

50000

60000

70000

2007 2011 2020

(tons)

Emissions of primary PM2.5

0

1000

2000

3000

4000

5000

6000

2007 2011 2020

(tons)

Fig. 2 Fuel consumpt io n (Pet ajou l e) and emissions f r om ship t r af f ic in t he wat er s

aro und Denmark (cf . Fig . 1) f or t he th r ee scenar io years 2007, 2011 and 2020.


24/134

24

The calculations predict a 15% increase in fuel consumption from 2007 to2020 as a result of increased traffic (36% increase for freight ships andunchanged passenger traffic), combined with increased efficiency for shipengines.

For emissions of nitrogen oxides (NOX) there is an expected growth of 2%between 2007 and 2020. Without stricter emission requirements the increasewould have corresponded to the increase in fuel consumption (15%).Accordingly, the result of the IMO reductions is an average reduction in NO Xemisssions of 11%per kg of fuel. Such relative reductions are expected tocontinue after 2020.

For sulphur dioxide (SO2) a reduction of 91% is envisaged between 2007 and2020, while for primary PM 2.5 the reduction is 54%. Primary particles accountfor only a minor fraction of the total amount of particles found in ambient air(see the section Concentrations of particles in this summ ary).T he most pronounced reductions in ship emissions take place for SO 2. T hese

reductions are illustrated in Fig. 3. T he maps yield an overview of sea trafficaround Denmark. Clearly visible are the main shipping lanes from the innerBaltic Sea (Bothnian Bay) to the N orth Sea, the m ajor D anish domestic ferryroutes, and the ferry routes connecting Denmark, Sweden, G ermany andPoland.

Concentrations of sulphur dioxideShip emissions are a major source to pollution with SO 2. Since 2007 it hasbeen a requirement for ships in the Baltic Sea and the N orth Sea that thesulphur content in fuel should not exceed 1.5%, whereas previously it typicallywas 2.7% (Section 3.1). Nevertheless, ships are still an important source incomparison with land-based sources. Fig. 4 shows calculated concentrationsof SO 2, referring to, respectively, 2007 (with 1.5% sulphur in fuel) and 2020.T he requirement to sulphur content in 2020 is considerably more restrictive

than now (max. 0.1% sulphur), and this is clearly reflected in computedconcentrations for 2020.

Fig . 3 Maps of year l y emissions of SO2f or t he th r ee years o f 2007, 2011 and 2020. The uni t is SO

2per km2..


25/134


26/134

26

Concentrations of nitrogen dioxide (NO2)

IMO has adopted regulations of NO X emissions. In the present study it isassumed that the marine waters around D enmark will become designated aso-calledN O

XEmission Control Area as defined by IMO. This implies several

restrictions on NOX emissions, in particular for new ships from 2016, wherean 80 percent reduction in N O X emission is required. There are other, lessdemanding requirements to older ships.

However, at the same time an increase in ship traffic is foreseen. Thereduction in emissions is not able to completely outbalance the foreseenincrease in ship traffic, so the N OX emission from ships will be slightly higherin 2020 than in 2007. It is likely that the IM O requirements will result in lowership emissions after 2020, but this is not the case for 2020. On the other hand,

there are marked changes for land-based sources according to the scenario for2020.

The result in terms of concentrations is seen in Fig. 6. NO X leads to formationof N O 2, for which health-related limit values exist. F ig. 6 shows modelcalculated concentrations of NO 2 for 2007 and 2020, respectively.respectively. T he large shipping routes remain clearly visible, but inlandconcentrations decrease. T he share from ships to the N O 2 concentration levelis 21% in 2007, but increases to 34 % in 2020 - because the absolutecontribution from other sources is reduced.

It is an interesting result of the calculations that the concentration of NO 2 inurban background in C openhagen is expected to be reduced by 7 g/m3 (from16 to 9 g/m3) in the period up to 2020. In several highly trafficked streets inCopenhagen the EU limit value of 40 g/m3 is exceeded; this fact makes alarge drop in background levels very interesting.

0

10

20

30

40

50

60

70

80

SO2 NO2 Primary PM2.5 mPM2.5

%

froms

hipemissions

EMEP-ref

AIS-2007

Fig. 5 Rel at ive cont r ibut ion f r om shi ps t o t he aver age concent r at ion s of var iouspol l ut ant s in Denmar k. The col our in dicat es which emission in vent or y th ecal cul at io ns are based on. The l igh t bl ue col umns use t he new, mor e pr ecisein vent or y ("AIS-2007"). The dark bl ue use t he ol d emission in vent or y f r om EMEP("EMEP-ref ").


27/134

27

Concentrations of particlesFine particles with a diameter less than 2.5 micrometer are referred to asPM 2.5. One can distinguish betweenprimary particles and secondary particles.Primary part icles exist as particles immediately after they have left the source.Secondary particles were not 'born ' as particles, but are created from gases,which undergo chemical transformation during transport a process thatcontinues for several hours or days after the pollution has left the source.Secondary particles can be further characterised as secondary inorganicpart icles or as secondary organic particles.

Emissions of SO 2 and NOX from ships contribute to the formation ofsecondary inorganic particles. The D EHM model which is used forcalculations of concentrations takes account of the processes involved.However, there is insufficient knowledge to describe the formation ofsecondary organic particles, and D EHM does not account for these.Internationally, this is recognised as being state-of-the-artfor current models(see Yttri et al., 2009). Accordingly, the DEHM model accounts only for acertain fraction of the particles found in the atmosphere, namely the primaryand the secondary inorganic. This part is here designated mPM

2.5(modelled

PM 2.5).

Fig. 7 shows calculated concentration levels of mPM 2.5. PM 2.5 can betransported over large distances, whereas SO 2 and NO 2 have a shorter

atmospheric lifetime. This is the reason why the geographical pattern in Fig. 7is very dissimilar to the previous figures. T he pollution from ships is notvisible, while it is a dominant feature that p ollution from central Europegradually declines over Denmark. The share of mPM 2.5 attributable to shiptraffic is 18%, while in terms oftotal PM 2.5 the share from ships is considerablysmaller. Relative to total PM 2.5 in Copenhagen, the share from ship traffic canbe estimated to around 4-7%, depending on whether one considers the air in ahighly trafficked street or in urban background (e.g. a park).

In the time up to 2020 a reduction of the general level of mPM 2.5 will takeplace, amounting to slightly more than 2 g/m3. This is mainly due toemission reductions for land-based sources.

Fig. 6 Cal cul at ed concent r at ion s of NO2 In g/ m3. The l ef t panel r epr esent s t he sit uati on in 2007, t he ri gh t

panel t he sit uat ion i n 2020. Ther e is har dl y any change in t he emissions of NOX f r om shi p t r af f ic bet ween t het wo year s, whi l e t her e ar e l ar ge r educt ions f or l and-based sour ces.


28/134

28

Fig. 7 Cal cul ated concent r ati on s of mPM2.5 in g/ m3. mPM2.5 is t he f r acti on of par t icl e mass th at can be deal t

wit h in t he air pol l ut ion model . Thi s is pr imar y par t icl es pl us secondary ino r ganic par t icl es. The l ef t panelr epr esent s th e sit uat ion in 2007, th e r igh t panel t he sit uat ion in 2020.


29/134

29

1Introduction

The work described in the present report covers several aspects of thecontribution from ships to air pollution in D enmark. T he main focus of thereport is to describe a newly developed and more accurate emission inventoryfor air pollution related to ship traffic in the Danish marine waters, and to usethis inventory to determine the contribution from ships to concentration levelswithin Denmark. Further, the study includes scenario calculations for years2011 and 2020. These scenario calculations are used to assess the effect thatinternational regulations of ship emissions and other regulations will have forthe future air pollution in D enmark. A minor comp onent of the workconcerns the contribution from ships at port to local air pollution in the portsof Copenhagen and Aarhus.

1.1 Backgroun d

Emissions from ship engines contribute to air pollution with sulphur dioxide(SO 2), nitrogen oxides (N OX), particulate matter (PM ) and Volatile OrganicCompounds (VOC ). Furthermore, ship engines emit the greenhouse gas CO 2.Focus of the present report is, however, on traditional pollutants, not on CO 2.

Previous investigations have made clear that the contribution from ships toemission of SO 2 and NOX is considerable, compared to land-based sources,and that there is a substantial effect on inland concentrations. It is importantto note that even though N OX and SO 2 are gases, they will contr ibute to the

formation of particles. Some of the particles present in ambient air have beenemitted as part icles (primary particles), while others (secondary particles) havebeen formed through chemical and physical reactions in the atmosphere.Thus, although SO 2 and NOX are emitted as gases, they will affect inlandconcentration levels for both these gases themselves, as well as for reactionproducts, including particles.

Pollution with particles is of considerable interest, because particles areassociated with negative health effects. A commonly used measure for particlepollution is concentration in terms of PM 2.5, i.e. part icles smaller than 2 .5micrometer. This is referred to as the fine fraction of particle pollution.

As mentioned above, one can distinguish between primary particles andsecondary particles. Secondary particles can be further characterised assecondary inorganic particles or as secondary organic particles. All types ofparticles are present in ambient air. Note, however, that emission inventoriestake account on ly ofprimary particles, because only these are actually emittedfrom the sources. T he DEHM m odel - which is applied to compu teconcentration levels in ambient air in Chapter 4 - also takes account of theformation of secondary inorganic particles. Some of these originate from shipemissions of SO 2 and NO X. However, DEHM does not account for secondaryorganic particles. This question is elaborated in Chapter 4, in particularsections 4.2 and 4.9.


30/134

30

The present study represents no attempt at evaluation of health effects; itconsiders only concentration levels of various compounds.

1.2 AIS dat a

Since 2006 the so-called AIS system - Automatic Identification System - has

been implemented in the marine waters around Denmark. It is required thatall ships greater than 300 G T carry an AIS transponder. With very short timeintervals the transponder transfers signals to land-based stations, providinginformation on ship identity, position, destination etc. T his type ofinformation p rovides the basis for estimates of the ship emissions and theirspatial distribution in the present study. The methodology is described indetail in Chapter 2. T he spatial distribution of emissions becomes much moreprecise with use of AIS data than with previously available data.

In the present study, AIS data were requested from DaMSA (DanishMaritime Safety Administration). Since AIS signals represent a huge amountof data, only data from a limited area around Denmark were requested fromDaMSA as illustrated in Figure 1.1. Furthermore, in order to represent theyear 2007, not all data for the entire year was considered. The present projectuses 12 two-days periods, one period for each month, representing bothweekend days and normal working days. From DaMSA data the followinginformation was used: Vessel IMO and MMSI codes, call sign, time of AISsignal, and latitude-longitude coordinates. For each ship, the sailing speedbetween two AIS signals was calculated from the time between the signals,and the corresponding latitude-longitude registrations.

Figu r e 1.1 Fuel consumpt io n accor din g to AIS dat a wit h appl icati on o f t hemet hodo l ogy descr ibed i n chapt er 2. The un it is TJ/km2. The col our ed area on t he mapil l ust r at es where th e AIS dat a ar e appl ied in t he pr esent st udy. In t he f ol l owing, itwill be r ef er r ed t o as t he AIS in vent or y ar ea.


31/134

31

79 % of the ships were identified as entries in L loyds Registers technicaldatabase. A reasonable assumption was that the ships not included in thelatter database were merely small vessels. Further assessment of the number ofships and fuel consumption estimates, also made it reasonable to assume thatemission results the 12 periods could be extrapolated to cover a full yearemission estimate.

1.3 New r egul at ion s and pr oject ions

The present study considers the current situation, as well as two projections.In 2008, the International Maritime Organisation ( IMO ) agreed on new,stricter rules for the emissions of SO 2 and NOX, to be implemented stepwise inthe period towards 2020.

Figure 1.2 illustrates the requirements for sulphur content in heavy fuel oil,which is the most common marine fuel. SECA stands for SO

xEmission

Control Area. T he curve labelled N on-SECA represents the general

requirements, applicable everywhere. In SECA's requirements are stricter.T he Baltic Sea and the N orth Sea1 have been appointed SEC A areas, with thestricter requirements implemented in 2006 and August 2007, respectively (seechapter 2 for details). Prior to 2006, the sulphur content in fuel used by shipsin these areas has assumed to be 2.7%, in accordance with Cofala et al. (2007)- see Section 3.1 for more details.

The basic scenario that is taken to represent the current situation is the year2007. In computations for this scenario it has been assumed that the BalticSea and N orth Sea were SECA areas during all of 2007.

1

In the con text of SECA's the Baltic Sea and the N orth Sea are larger than theirnormal delimitation; the two SEC A's are adjacent and include the inner Danishwaters.

0

1

2

3

4

5

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021

Sulphurcontent(%)

SECA

Non-SECA

Figur e 1.2 Limit t o sul phur cont ent in h eavy f uel oil , accor ding t o IMO r egul at ion s.The SECA r egul ati on s wer e impl ement ed in t he Bal t ic Sea in 2006 and in t he Nor t h Seain August 2007. For cal cul ati on s in t he pr esent r epor t , SECA r egul ati on s have beenassumed in ef f ect dur in g t he ent ir e year of 2007. The br oken l in e in dicates 2.7% wh ichwas the assumed l evel pr io r t o 2007.


32/134

32

NO x emissions are regulated by IM O (M arpol 73/78 Annex VI, and further,amendments) in a three tiered emission regulation approach, as follows:

Tier I: Diesel engines (> 130 kW) installed on a ship constructed onor after 1 January 2000 and prior to 1 January 2011.

Tier II: Diesel engines (> 130 kW) installed on a ship constructed onor after 1 January 2011.

Tier III: Diesel engines (> 130 kW) installed on a ship constructed onor after 1 January 2016.

T ier III applies only in areas, which have been designated N OX

EmissionControl Area (N ECA). T he countries around the Baltic Sea are preparing anapplication to obtain this status for the Baltic Sea. It is assumed for the 2020scenario that the marine waters around Denmark are a NEC A.

As it appears from Figure 1.2 there will be substantial reductions in sulphurcontent in SECA areas in 2010 and 2015. Furthermore, the NO x emissions

from new engines become significantly lower in the period considered. Theeffect of these measures has been studied, based on p rojected emissionscenarios for the years 2011 and 2020, in addition to the baseline scenario for2007.

The scenarios take account of a general increase in ship traffic during theperiod up to 2020. In accordance with expectations from the D anish Ship-owners' Association (see chapter 2) it has been assumed that, due to thefinancial crisis, ship traffic has dropped since 2007, but will again in 2011reach the level of 2007. From 2011 and onwards, an annual increase of 3.5percent has been assumed for transport of goods, while passenger traffic isassumed unchanged.

1.4 Rel at ion t o pr evious st udies

Many previous studies from NERI have included emissions from ships as oneof the sources to air pollution, whereas they have not addressed thecontribution from ships separately. Among such studies are annual reports onthe D anish monitoring programmes (e.g. Kemp et al., 2008; Ellermann et al.,2009).

The present study is the first to make use of AIS data for air pollution studiesin a Danish context. In other countries the use of AIS data for construction ofemission inventories has recently started.

A related study is the report by Winther (2008) on fuel consumption andemissions from navigation in D enmark, which is, however, focussed onnational Danish ship traffic.

In 2001, emission inventories for 1995/1996 and 1999/2000 were compiledby Wismann (2001), estimating the fuel consumption and emissions for allsea transport in D anish waters.

Other D anish p rojects were the assessment of the fuel consumption and

emissions for ships in D anish ports (hotelling, manoeuvring, landing/loading)by Oxbl and Wismann (2003) , and the examination of air quality effects


33/134

33

from cruise ship activities in the Port of Copenhagen by Olesen andBerkowicz (2005). Other studies concerning pollution in ports are alsorelevant and will be discussed in Chapter 5.

1.5 St r uct ur e of t he r epor t

T he report has the following structure.

Chapter 2 provides a comprehensive description of the methodology appliedfor constructing the detailed AIS-based emission inventories, which form thebasis for the further calculations. The inventories are for the years 2007, 2011and 2020. It is described how AIS data obtained from D aMSA are processedin order to determine time, position and speed for each ship in the AISinventory area. Further, the chapter explains how speed based engine loadfunctions are used to estimate the instantaneous ship engine load. T heemission factors are described, as well as relevant ship technical data obtainedfrom e.g. Lloyds Register. T he emissions are calculated as a product ofemission factors, engine size, engine load and time between AIS signals.Emission results are shown for 2007, 2011 and 2020 in details per enginetype, fuel type and ship category. Results are also listed per flag state, and G ISmaps show the geographical distribution of emissions in the AIS inventoryarea.

Chapter 3 goes one step further from the AIS-based inventories. In order toapply these inventories in an air pollution model, it is necessary to combinethem with emission data covering a much larger geographical data. Chapter 3presents the inventories for ship emissions that are used for model runs withthe air pollution model DEH M (D anish Eulerian H emispheric Model). It isdescribed how the AIS-based inventories from Chapter 2 are combined with

emission data provided by EMEP ( Co-operative Programme for M onitoring andEvaluation of the Long-R ange T ransmission of Air Pollutants in Europe). T heEMEP estimates are characterised by a cruder geographical resolution, andthere are interesting differences between the two sets of inventories which arediscussed.

Chapter 4 describes how the ship emissions are combined with emissionsfrom land-based sources and aviation. However, the bulk of the chapter isdevoted to a presentation and discussion of results for concentrations,obtained with the air pollution model DEHM (D anish Eulerian H emisphericModel). T he chapter includes a quantification of the contribution from shipsto concentration levels in Denmark. Further, it presents scenario calculations,

demonstrating the combined effect on concentration levels of the IMOmeasures and the expected changes in land-based emissions. Finally, itconsiders the influence of ship traffic on general concentration levels in theCopenhagen area.

Chapter 5 presents results of studies of local pollution from ships at port inCopenhagen and Aarhus.

Conclusions are found in the section Summary and conclusions in front of thereport.


34/134

34


35/134

35

2Emissions

2.1 Introduction

T his chapter provides a comprehensive description of the methodologyapplied for constructing the detailed AIS-based emission inventories of shipemissions, which form the basis for the further calculations. T he inventoriesare for the years 2007, 2011 and 2020. Furthermore, the chapter summarisesresults.

After a brief overview of the entire approach, the subsequent sections providedetails.

For the 'AIS inventory area' defined in Figure 1.1, AIS data have beenrequested from the D anish Maritime Safety Administration (D aMSA; inDanish: Farvandsvsenet). In order to calculate fuel consumption andemissions for ships in the area, these data are u sed together with vesseltechnical information from Lloyds Register, engine load functions providedby DT U (T echnical University of Denmark), and general fuel consumptionand emission factors provided by NERI. In order to facilitate the inventorycalculations, assumptions are also made regarding fuel type used, engine typesand average engine life times.

For each single vessel in the AIS dataset, ship category, engine type, fuel type,main engine and auxiliary engine size are determined using the technical data

from Lloyds Register and supplementary information from DT U and M ANDiesel. The vessel sailing speed is found between each AIS signal, and theinstantaneous engine load is calculated from basis functions (rep resenting fivecommon ship classes), using main engine size and vessel sailing speed asinput.

The fuel consumption and emissions from each vessel during the timebetween two consecutive AIS signals are calculated by combining engine size,engine load, time duration between the AIS signals, and fuelconsumption/emission factors corresponding to the vessels engine and fueltype. The baseline results are calculated for the year 2007, and results for the

prognosis years 2011 and 2020 are obtained by using forecasted fuelconsumption/emission factors from NERI and expectations for sea trafficgrowth from Danish Ship-owners' Association.

2.2 AIS dat a pr ovided by DaMSA

AIS signals represent a huge amount of data. In order to restrict the volume ofdata, a limited area around Denmark (F igure 1.1) was appointed as being ofprimary interest, and only data from that area were requested from D aMSA.Furthermore, in order to represent the year 2007, not all data for the entireyear was considered. The present project uses 12 two-days periods, one

period for each month, representing both weekend days and normal workingdays. The following AIS data are used: Vessel IMO and MMSI codes, call


36/134

36

sign, time of AIS signal, and latitude-longitude coordinates. For each ship, thesailing speed between two AIS signals is calculated from the time between thesignals, and the corresponding latitude-longitude registrations.

Table 2.1 shows the selected date and day combinations for the differentmonths in 2007, and the corresponding percentage of AIS signals (from theDaMSA dataset) pertaining to identified ships, and the num ber of ships thatcould be identified as entr ies in L loyds Registers technical database. T hetotal number of AIS signals in D aMSA data is 15.725 mio. (not shown), andthe total percentage of identification is 79 %. Based on information fromLloyd's (pers. comm., A. Halai, Lloyds Register, 2009) the unidentified shipsare assumed to be m erely small sized vessels. T hus, it is well known that manymodern pleasure craft have AIS systems installed. Consequently, the biasintroduced in the subsequent fuel consumption and emission calculations isregarded as marginal.

Figure 2.1 displays the number of ships for each months 48-hour period. T henumber of ships varies to some extent from period to period; however, therange of variation is limited. The same conclusion can be drawn from the

corresponding fuel consumption results shown in Figure 2.7 in section 2.9.Based on these observations it is considered a reasonable assumption that thetotal number of signals obtained from DaMSA for the time periods inquestion can be extrapolated to cover a full year by using a multiplicationfactor of 365/24, and that th is will result in a reasonably reliable estimate offuel consumption and emissions from the subsequent inventory calculations.

Tabl e 2.1 Days sel ect ed f or AIS dat a capt ur e.

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Date 13-14 19-20 8-9 11-12 6-7 15-16 24-25 18-19 11-12 26-27 12-13 3-4

DaySat-Sun

Mon-Tue Thu-Fri

Wed-Thu

Sun-Mon Fri-Sat

Tue-Wed

Sat-Sun

Tue-Wed Fri-Sat

Mon-Tue

Mon-Tue

Identifiedsignals (%) 83 80 81 79 82 80 78 79 77 79 76 74Identifiedships (No.) 1231 1348 1269 1388 1524 1490 1256 1499 1429 1525 1286 1184

0

200

400

600

800

1000

1200

1400

1600

1800

J an Fe b M ar Apr Ma y J un Jul Au g S ep Okt Nov Dec

Figur e 2.1 Number o f sh ips f or each mont h s 48-hour per io d in DaMSA data, ident if iedas ent r ies in t he Ll oyds Register s t echn ical database.


37/134

37

2.3 Vessel data pr ovided by Ll oyds

Relevant technical data for each ship is found by linking the vessel's IMO andMMSI codes and call sign from AIS data to the technical registrations fromLloyds Registers database. T he latter data consists of main engine size,engine stroke type ( 2-stroke/4-stroke), vessel flag and general ship category.T he information on general ship category is used to group the vessels into five

ship classes (representative types) for which engine load functions can beestablished (c.f. section 2.6). The Lloyds ship categories and therepresentative ship classes are shown in Annex A.

2.4 Engine t ypes, f uel t ypes and aver age engine l if e t imes

From Lloyds' database a distinction between 2-stroke and 4-stroke engines aswell as gas turbine engines is given. It is necessary to further allocate thesedata into the general engine types: Gas turbine, slow speed, medium speedand high speed engines for which fuel consumption and emission data areavailable. T he following table shows the applied classification , which is basedon information from M AN Diesel (pers. comm., Flemming Bak, 2009) andWinther ( 2008).

2.5 Tr af f ic f or ecast

The Danish Ship-owners' Association was requested to provide a forecast for

the development in ship traffic for the present study. The Association expectsthat in 2011 the amount of traffic will be back at the 2007 level after adecrease related to the current financial crisis (pers. comm. Jacob Clasen,Danish Ship-owners' Association, 2009). In recent years there has been alarge annual growth in traffic of around 5 %. However, in the second part of2008 the traffic began to decrease due to the global financial crisis, and theexpectation is that th is drop in sea transport activity will not becomeoutbalanced until 2011.

From 2011 to 2020 the D anish Ship-owners expect an annual traffic growthof between 3-4 % for goods carrying vessels, and hence 3.5 % is used in thepresent survey. T he traffic levels for passenger ships are expected to be the

same as for 2007.

Tabl e 2.2 Esti mated main engine t ype and f uel t ype f or ship engines in t he pr esentinventory

Engine type Enginesize

Engine type Fuel type Engine li fe time

(LloydsRegister)

(kW) (estimated) (estimated) (years, estimated)

Gas turbine Gas turbine Diesel 30

2-stroke Slow speed (2-stroke)

HFO 30

4-stroke 4000 Medium speed (4-stroke)

HFO 30


38/134

38

2.6 Engine l oad f unctio ns

An extensive database with ship data is maintained by Hans Otto Kristensen(Senior researcher, DTU 2). As part of the p resent p roject, H.O . Kristensenhas used this database to produce basic equations for service speed (V s) as afunction of main engine service power (P s). T echnical data from a largenumber of vessels (in brackets) form the base for the equations for the five

most common ship categories: Container ships (240), T ankers (199), BulkCarriers (74), Ro-Ro cargo (59) and Ro-Ro passenger ships (91) . T he directdata source is the Royal Institution of Naval Architects. Since 1990 they haveannually published technical details for around 50 ships of all classes andtypes in their publication series Significant Ships (http://www.rina.org.uk).T he latter source of information is regarded throughout the maritimecommunity as the most reliable source of information of technical datadirectly linked to individual vessels (pers. comm. Hans Otto Kristensen,DT U, 2009).

For data validation p urposes, a comparison was made between the proposed

speed power relations from D T U, and regression functions that could beobtained from the Swedish ShipPax database for Ro-ro cargo ships (870ships), and from L loyds Registers technical database for container ships(5150 ships). A perfect match between regression curves was found in bothsituations (DT U, 2009).

The speedpower relations for each of the five ship categories are representedby the curves in Figure 2.2. The curves are estimated based on regressionanalysis, and hence they are well suited as input data for fuel consum ptionand emission inventories comprising groups of vessels. T he curve valuescannot be u sed directly for specific ships, since in the individual case thespeed power relations to some extent differ from the general curves.

In a few cases P s for other ship categories than Ro-Ro passenger ships aresmaller than the lower limit for P s on the curves depicted in Figure 2.2, andhence, for ships with such small engines the service speed cannot be directlydeduced from the curves in Figure 2.2. Instead, in these situations thefunction for Ro-Ro passenger ships is used as an approximate approach.

Based on experience, the necessary engine power (P cal) for ship p ropulsion atan observed sailing speed ( Vs) is found by adjusting the engine service power(PS) with the observed:service speed ratio to the 3

rd power:3

S

obs

scal V

VPP (1)

Finally, the engine load percentage, %M CR, is expressed as:

100/100%

3

ME

S

obss

ME

cal PV

VP

P

PMCR (2)

If the calculated %MCR exceeds 100%, %MCR is set to 100%.

2 Department of mechanical engineering, Section of coastal, maritime and structuralengineering


39/134

39

2.6.1 Auxiliary power in useFor the ship categories: Container ships, Tankers, Bulk Carriers, Ro-Ro cargothe installed auxiliary power is expressed as a function of main engine size inthe state of the art AE-ME functions agreed by IMO M EPC (M arineEnvironment Protection Committee) at its 59th meeting (IMO M EPC, 17August 2009). The AE-ME equations have been developed by DTU, and theunderlying data base is fuel consumption data reported by large shippingcompanies. Prior to the IMO MEPC 59th meeting, the AE-ME data was

provided by D T U ( 2009) for use within the present inventory:

Container Ships

VS = 3,7Ln(PS) - 14.8

R2= 0.95

8

12

16

20

24

28

0 10000 20000 30000 40000 50000 60000 70000

PS = 85 % MCR [kW]

VS[knots]

Tankers

VS = 1.12Ln(PS) + 4.6

R2= 0.49

9

11

13

15

17

0 4000 8000 12000 16000 20000 24000 28000 32000

PS= 85 % MCR [kW]

VS[knots]

Bulk carriers

VS = 1.31Ln(PS) + 2.8

R2= 0.73

9

11

13

15

17

0 3000 6000 9000 12000 15000 18000 21000 24000

PS = 85 % MCR [kW]

VS[knots]

Ro-Ro passenger ships

VS = 0.00037 PS + 14.5

R2 = 0.78

8

12

16

20

24

28

32

0 9000 18000 27000 36000 45000 54000

PS = 75 % MCR

VS(knots)

Ro-Ro Cargo ships

Speed = 3.82Ln(PS) - 15.9

R2 = 0.61

14

16

18

20

22

24

0 5000 10000 15000 20000 25000 30000

PS = 75 % MCR [kW]

VS[knots]

Figu r e 2.2 Ser vice speed (Vs) as a f unct io n of main eng in e ser vice power (Ps) f or Con t ain er ships, Tankers, Bul k

Car r ier s, Ro-Ro cargo and Ro-Ro passenger shi ps. MCR stands f or Maximum Con t in uous Rat in g.


40/134

40

AE = 5% M E, ME < 10000 kW (3)

AE = 250 kW + 2.5% ME, ME > 10000 kW (4)

For Ro-Ro p assenger ships the installed auxiliary power is estimated to be20% of the ships main engine size, based on queries from the database kept byDTU.

For all ship categories, the average auxiliary engine load is assumed to be 50%(pers. comm. H ans Otto Kristensen, DT U, 2009).

2.7 Fuel consumpt ion and emission f actor s

Generally, the fuel consumption and emission factors in g/kWh depend on toengine type, fuel type and engine production year.

2.7.1 Specific fuel consumptionThe standard curves for specific fuel consumption, sfc (g/kWh), are shown inFigure 2.3 for slow-, medium- and high-speed engines, as a function of engineproduction year. For gas turbines, a mean fuel consumption figure of 240g/kWh is used. All fuel consump tion data come from the D anish T EM A2000emission model (Ministry of T ransport, 2000), and are based on the enginespecific fuel consumption data from several engine manufacturers (pers.comm. Hans Otto Kristensen, DT U, 2009).

Considering the fuel consumption trend graph in F igure 2.3, the first part of itapplies to engines manufactured up until the mid 1990s, and was produced

in the late 1990s for the D anish T EM A 2000 model. Because the regressioncurve is supported by actual fuel consumption factors for these engines, thispart of the graph is regarded as being the most accurate. For newer engines,

0,0

50,0

100,0

150,0

200,0

250,0

300,0

1949

1953

1957

1961

1965

1969

1973

1977

1981

1985

1989

1993

1997

2001

2005

Engine production year

sfc(g/kWh)

Medium speed (4-stroke) Slow speed (2-stroke) High speed (4-stroke)

Figu r e 2.3 Specif ic fuel con sumpt io n f or mar in e eng in es r el at ed to t he eng in epr oduct ion year (g/kWh)


41/134

41

the fuel consumption trend is established based on expert judgement. T hegraph is, however, still regarded as valid in relation to its use in estimatingemission for engines in the situation which prevails today (pers. comm. HansOtto Kristensen, DT U, 2006) . T he sfc figures for 2005 are used for enginesbuilt from 2006 onwards to provide the basis for the fuel consumptioncalculations for future years.

Using the average engine life times, LT , listed in T able 2.2, the average sfcfactors per inventory year, X, is calculated from:

k

LTXyear

Xyear

yk

XkLT

sfc

sfc

,

, (5)

Where sfc = specific fuel consumption (g/kWh), X = inventory year, k =engine type, y = engine production year, LT = engine life time.

The average sfc factors per inventory year are shown in Figure 2.4 for the

inventory years 2000-2020.

2.7.2 NOx emission factors

2.7.2.1 IM O emission regulations for NOx

For seagoing vessels, NO x emissions are regulated as explained in Marpol73/78 Annex VI, formulated by IMO (International Maritime O rganisation),and further, amendments to MARPO L Annex VI has been agreed by IMO inOctober 2008. A three tiered emission regulation approach is considered,which comprises the following:

Tier I: Diesel engines (> 130 kW) installed on a ship constructed on

or after 1 January 2000 and prior to 1 January 2011.

0

50

100

150

200

250

2000

2002

2004

2006

2008

2010

2012

2014

2016

2018

2020

Inventory year

Sfc(g/kWh)

High speed (4-stroke) Medium speed (4-stroke) Slow speed (2-stroke)

Figu r e 2.4 Aver age sfc f act or s f or mar in e eng in es f or t he invent or y year s 2000-2020(g/kWh)


42/134

42

Tier II: Diesel engines (> 130 kW) installed on a ship constructed onor after 1 January 2011.

Tier III3: Diesel engines (> 130 kW) installed on a ship constructed onor after 1 January 2016.

T he NO x emission limits for ship engines in relation to their rated enginespeed (n) given in RPM (Revolutions Per Minute) are shown in T able 2.3.

Following the IMO emission regulations, the NO x T ier I limits are also to beapplied for existing engines with a power output higher than 5000 kW and adisplacement per cylinder at or above 90 litres, installed on a ship constructedon or after 1 January 1990 but prior to 1 January 2000.

2.7.2.2 NOx

emission factors for engines built before 2006

T he NO x emission factors (g/kWh) for slow- and medium-speed engines areobtained from M AN DIESEL ( 2006). With a global market share of 75%,

MAN Diesel is by far the worlds largest ship engine manufacturer, and hencein terms of representativity, the emission factors are well suited as inpu t forinventory emission calculations comprising many ships. For a relevant year ofcomparison, 2000, Winther (2008) finds a good accordance between theMAN Diesel emission factors and from other important studies (Whall et al.2002; Endresen et al. 2003) per engine type. T he concordance is the best forslow speed engines which is the most dominant source for NOx emissions.

T he NO x emission factors provided by MAN Diesel for slow- and mediumspeed engines are shown in Figure 2.5, together with NO x emission factors forhigh-speed engines. For gas turbines, a mean NO x emission factor of 4 g/kWhis used. The emission information for high-speed engines and gas turbinescomes from the D anish T EM A2000 emission model (M inistry of Transport,2000) . For h igh speed engines the emission factor level is determined byKristensen (2006) through discussion with relevant engine manufacturers,considering engine operation at a normal engine speed range (1000 RPM) forhigh speed ferries. For high speed engines build in 2000, the NO x emissionfactor from Figure 3.4 fits with the IMO T ier I emission standard derivedfrom the relevant equation in T able 3.2.

The increase in fuel efficiency up to year 2000 caused the NO x emissionfactors to increase. However, in the beginning of the 1990s ( for slow-speedengines) and by the end of the 1990s (for medium-speed engines), NO x

3 For ships operating in a designated Em ission Control Area. Outside a designatedEmission C ontrol Area, T ier II limits apply.

Tabl e 2.3 Tier I-III NOx emissio n l imit s f or shi p eng ines (amendment s t o MARPOLAnnex VI)

NOx limit RPM (n)Tier I 17 g/kWh

45 x n-0.2 g/kWh9,8 g/kWh

n < 130130 n < 2000n 2000

Tier II 14.4 g/kWh44 x n-0.23 g/kWh7.7 g/kWh

n < 130130 n < 2000n 2000

Tier III 3.4 g/kWh

9 x n-0.2 g/kWh2 g/kWh

n < 130

130 n < 2000n 2000


43/134

43

emission performance is improved, mainly due to improved engine design.The emission improvements are of a sufficient size to enable the IMO Tier INO x emission requirements in 2000 to be met.

2.7.2.3 NOx

emission factors for engines built from 2006 onwards

The Tier III requirements for new ships built after 2016 will apply indesignatedNO

XEmission Control Areas (N ECA). It is assumed that the AIS

inventory area is appointed a N ECA.

T hus, for newer engines in compliance with T ier II (2011) and T ier III(2016) emission standards, emission factors are estimated by adjusting the

Tier I emission factors (2000-2005) in two steps, relative to the Tier II:Tier I

0,0

5,0

10,0

15,0

20,0

25,0

1949

1953

1957

1961

1965

1969

1973

1977

1981

1985

1989

1993

1997

2001

2005


NO

x(g/kWh)

Medium speed (4-stroke) Slow speed (2-stroke) High speed (4-stroke)

Figur e 2.5 NOx emission f act or s f or ship eng in es buil t bef or e 2006 (g/kWh)

02

4

6

8

10

12

14

16

18

2006

2007

2008

2009

2010

2011

2012

2013

2014

2015

2016

2017

2018

2019

2020


NOx

(g/kWh)

High speed (4-stroke) Medium speed (4-stroke) Slow speed (2-stroke)

Figur e 2.6 NOx emission f acto r s fo r shi p engi nes buil t f r om 2006 onwar ds (g/ kWh)


44/134

44

and T ier III:T ier I ratios. T he estimated emission factors for the engineproduction years 2006-2020 are shown in Figure 2.6.

2.7.2.4 T he effect of IMO N Ox

emission requirements for engines built from 1990

and prior to 2000

For slow speed engines, the new T ier I emission standards for existing enginesbuilt from 1990 to 1999 are somewhat lower than the emission factors actuallymeasured by MAN Diesel. As mentioned in section 2.7.2.2, most of the shipengines today are built by MAN Diesel, and from their side it is expected that75% of all slow speed engines built between 1990 and 1999 in the 7500-22500 kW segment will become retrofitted with more emission efficient slidevalves in order to meet the IMO emission standards. Other engine sizes willalso be considered at a later stage.

T he slide valves are designed in a way which improves fuel atomization,penetration and mixing in the engine combustion chamber. On the same timeengine performance adjustments are made (pers. comm. M ichael F. Pedersen,MAN Diesel, 1999). The retrofit scheme is envisaged to take place following

a linear time schedule from 1st of July 2011 to 1st of July 2016.The retrofit emission effect is incorporated in the present inventory in similarway so that emission factors for 1990-1999 engine production years aregradually being replaced by T ier I emission factors, going from 0% to 100%representat ion between 1st of July 2011 and 1st of July 2016.T he correction factor KR is found as:

yk

RRTierIkRRyk

REF

PSXSEFPSXSEFXPykK

,

,, )()())()(1(),,,(

(6)

Where:k = engine type (slow speed)X = Inventory year, X = [2011;2020]P = M ain engine size (M E), P = [5000 kW;22500 kW]y = engine production year, y = [1990;1999]

Outside the criteria for k, y, P and X, K R = 1.

2.7.2.5 Average NOx

emission factors per inventory year

T he average NO x emission factors take into account engine production year,average engine life times and the 2011-2016 retrofit scheme, as explained in

the previous sections. It thus assumes that the sea area considered isappointed a NECA area, so Tier III regulations will apply after 2016.

Using the average engine life times listed in T able 2.2, the average NO xemission factors per inventory year, X, is calculated from:

k

R

LTXyear

Xyear

yk

XykLT

XPykKEF

EF

),,,(,

,,

(7)

Where EF = emission factor (g/kWh) , X = inventory year, k = engine type, y= engine production year, LT = engine life time.


45/134

45

For the inventory years 2007, 2011 and 2020 the average N Ox emissionfactors are shown in Annex II for ship category/engine type combinations, andwith/without retrofit.

2.7.3 SO2

In relation to the sulphur content in heavy fuel and marine gas oil used by

ship engines, Table 2.4shows the current legislation in force, and theamendment of MARPOL Annex VI agreed by IMO in October 2008. Thesesulphur contents are also used in the present inventory.

From 2006/20074, the SEC A areas enter into force. T he current inventory for2007 assumes that the SECA regulations have been in place for the entire yearof 2007, so a fuel sulphur content of 1.5% is used for heavy fuel oil in 2007.In 2011 and 2016 the sulphur content gradually become lower, as prescribedby the IMO fuel standards.

In order to obtain emission factors in g/kWh, the sulphur percentages fromT able 2.5 are inserted in the following expression:

100

%2)(

2

sfcSSOEF

(8)

Where EF = emission factor in g/kWh, S% = sulphur percentage, and sfc =specific fuel consumption in g/kWh. The sfc factor is taken from equation 5.Equation 1 uses 2.0 kg SO 2/kg S, the chemical relation between burnedsulphur and generated SO 2 provided in EM EP/CORIN AIR (2007).

2.7.4 PM

For diesel fuelled ship engines the emission of particles (primary part icles -see Chapter 4 for details on secondary particles) depends on the fuel sulphurcontent, S%. T he emission factors in g/kg fuel are calculated as:

4 2007 is used as the effective year in the inventory

Tabl e 2.4 Cur r ent l egisl at ion in r el at ion t o mar in e f uel qual it yLegislation Heavy fuel oil Gas oil

S-% Impl. date S-% Impl. dateEU-directive 93/12 None 0.21 1.10.1994EU-directive 1999/32 None 0.2 1.1.2000EU-directive 2005/332 SECA - Baltic sea 1.5 11.08.2006 0.1 1.1.2008

SECA - North sea 1.5 11.08.2007 0.1 1.1.2008Outside SECAs None 0.1 1.1.2008

MARPOL Annex VI SECA Baltic sea 1.5 19.05.2006

SECA North sea 1.5 21.11.2007Outside SECA 4.5 19.05.2006

MARPOL Annex VI amendments SECAs 1 01.03.2010SECAs 0.1 01.01.2015Outside SECAs 3.5 01.01.2012Outside SECAs 0.5 01.01.20203

Notes:

1Sulphur content limit for fuel sold inside EU2From 1.1.2010 fuel with a sulphur content higher than 0.1 % must not be used in EU ports for ships at berth exceeding twohours3 Subject to a feasibility review to be completed no later than 2018. If the conclusion of such a review becomes negative theeffective date would default 1 January 2025


46/134

46

)745.0%(854.0)/( SPM ekgfuelgEF (9)

The PM emission factor equation is experimentally derived frommeasurements made by Lloyds (1995). taken from T EM A2000(T rafikministeriet, 2000).

Subsequently, the PM emission factor in g/kWh is found from:

1000

)/()/(

sfckgfuelgEFkWhgEF PMPM

(10)

Based on information from MAN DIESEL (N. Kjemtrup, 2006), the PM 10and PM 2.5 shares of total PM (= T SP) are 99 and 98.5%, respectively.

T he PM emission factors are experimentally derived from m easurementsmade by L loyds (1995). T he latter source make up the basis for most of theinventories and methodology descriptions carried out so far on aninternational level, and hence the PM emission factors corresponds well withthe input data used by e.g. the global inventories set up by Corbett andKoehler (2003) and Endresen et al. (2003) , the inventory made for Europe byWhall et al. (2002), the European calculation methodology description fromthe MEET project (Georgakaki et al. 2004) as well as the EMEP/CORINAIRguidebook methodology (EMEP/CORINAIR 2009). The PM emissionfactors from L loyds (1995) are also used in the D anish T EM A2000 emissionmodel (Ministry of Transport, 2000).

2.7.5 CO and VOC

In general the emission factors for CO and VOC in g/kWh are regarded asvery uncertain. The general experience from the ship engine manufacturer'sside is a decrease in CO and VOC when fuel efficiency improves (pers.comm. S. H enningsen, M AN Diesel, 2006). H owever, due to missingconsistent emission data as a function of engine year, constant factors for allengine production years are chosen for the present inventory. The emissionfactors originate from the emission measurement programme carried out byLloyds (1995) .

As for PM , the CO and VOC emission factors from Lloyds (1995) make upthe basis for most of the inventories and methodology descriptions carried outso far internationally, and in th is respect corresponds well with the emission

factors used by e.g. the global inventories set up by Corbett and Koehler(2003) and Endresen et al. (2003) , the inventory made for Europe by Whallet al. (2002), the European calculation methodology description from theMEET project (Georgakaki et al. 2004) a

Emission Pollutant

Documents

Transcript of Emission Pollutant