SHIPPING AND THE ENVIRONMENT

From Regional to Global Perspectives

Gothenburg, October 24th, 2017

Impact of maritime transport emissions

on coastal air quality in Europe
Mar Viana
Institute of Environmental Assessment and Water Research
mar.viana@idaea.csic.es
 Framework

EEA US EPA EU - Turkey

 Outline

1. Impact of international shipping on European air quality

• Tracers and physico-chemical characteristics
• Impact on ambient PM
• Impact on gaseous pollutants

2. Mitigation strategies:
• Overview
• Case study: environmental and health benefits of designating the
Marmara Sea (Turkey) as an ECA

3. Conclusions
 (1)
Impact of international shipping
on European air quality
 Rationale

Emissions from the marine transport sector

contribute significantly to air pollution globally

Increasing emission source:

• Globalization of manufacturing processes
• Increase of global-scale trade
• Relatively, large efforts to reduce other
sources (industrial, power generation, etc.)
• More future growth expected

Human health Climate Ecosystems

 How much of a problem?

 Different approaches used in different countries

 Not yet achieved the goals for protecting human health
 Chemical tracers

Well-known tracers of combustion based on crude oil:

• V and Ni (>60 publications)
• Others: La, Th, Pb, Zn and SO42- (>18 publications)
Where? PMx V/Ni Reference Where? PMx Tracer Value Reference
Mazzei et al. Spain PM10 V/EC <2 Viana et al. (2009)
Italy PM10 3.2±0.8 (2008)
Mazzei et al.
PM2.5 V/EC <2 Viana et al. (2009)
PM2.5 3.2±0.8 (2008) Spain PM10 La/Ce 0.6-0.8 Pandolfi et al. (2011)
Mazzei et al. PM2.5 La/Ce 0.6-0.8 Pandolfi et al. (2011)
PM10 3.2±0.8 (2008)
Italy PM10 soluble V 80% Becagli et al. (2012)
Agrawal et al.
Ship engine 2.3-4.5 (2008) PM10 soluble V >6 ng/m3 Becagli et al. (2012)
Spain PM2.5 4-5 Viana et al. (2009) PM10 soluble Ni 80% Becagli et al. (2012)
PM10 2-
non-ss SO4 /V 200-400 Becagli et al. (2012)
PM10 4-5 Viana et al. (2009)
Pandolfi et al.
Spain PM10 3 (2011)

PM2.5 3
Pandolfi et al.
(2011)
Tracers may be used in source
apportionment models, BUT:
Europe PM10 3-4 Viana et al. (2014)
changing fuels result in changing tracers
Europe PM2.5 3-4 Viana et al. (2014)
Alastuey et al.
Europe PM10 2.3-2.5 (2016)
 Other tracers
Shipping emissions correlate with:
 NO, NOx, SO2 and VOCs
 Particle number concentration (N): nucleation episodes (SO2) (Reche et al., 2011)
 Particle size distribution (80-500 nm; e,.g., Masiol et al., 2016)

25000 25
Reche et al. (2011)
N
20000 SO2 20

SO2 (µg/m3)
15000 15
N (#/cm3)

10000 10

5000 5

0 0
 Particle size distribution

 Knowledge gap!
o Difficult to discriminate from background
o Depends on measurement location (distance)

• Direct plume: bimodal N size distribution (40 nm, 70 nm) (Isakson et al., 2001)
• In ambient air:
- Stronger contribution to fine than coarse aerosols (Viana et al., 2009)
- Primary particles predominantly submicron (<1 µm) (Petzold et al., 2008;
Healy et al., 2009); modes at <250nm & 350nm (Merico et al., 2016)
- Impact on N, thus ultrafine particles (UFPs, <0.1 µm) (Saxe and Larsen,
- 2004; Reche et al., 2011)
- Particle number or toxicity better metrics than mass?

Zhao et al. (2013) Merico et al. (2016)

 Primary vs. Secondary particles

In Southern-Europe:

Qinhuangdao (China) (Lang et al., 2017):

Primary = 51% of shipping PM2.5 vs. Secondary = 48%; large seasonal variability

Viana et al. (2014)

Premature deaths/year in Europe:

due to primary particles due to secondary particles

(301.000/year) (245.000/year)
More efficient for health to decrease primary PM emissions?
Andersson et al. (2009); Hammingh et al. (2012); Tian et al. (2013); Lang et al. (2017)
 Impact on ambient PMx
Airborne particles
Size fraction / PM
Reference Source Contribution Location
component
Oil
Kim & Hopke (2008) 4-6% PM2.5 US
combustion
Oil
Mazzei et al. (2008) 20% PM1 IT
combustion
Minguillón et al. <5% OC US
(2008) Shipping
<5% PM2.5 US
Source apportionment tools: Oil
Viana et al. (2008) 10-30% PM10 and PM2.5 EU
• dispersion models combustion
5% PM10 ES
• receptor models Amato et al. (2009)
Oil
combustion
6% PM2.5 ES
8% PM1 ES
• chemical tracer methods Viana et al. (2009) Shipping
2-4% PM10 ES
14% PM2.5 ES
Hellebust et al. (2010) Shipping <1% PM2.5-10 and PM0.1-2.5 IE
Limitations: 3-7% PM10 ES
• mixed with other combustion Pandolfi et al. (2011) Shipping
5-10% PM2.5 ES
30% nss SO42- IT
sources (common tracers) 3.9% PM10 IT
Becagli et al. (2012) Shipping
• challenge of unique discrimination 8%
11%
PM2.5
PM1
IT
IT
• lack of comparability 1-5% PM2.5 North Sea
NL, UK,
Hammingh et al. Be, DK, Fr,
(2012) Shipping
1-5% PM2.5 DE, LU,
Norway,
SE, Switz.
Keuken et al. (2014) Shipping 0.5 µg/m3 PM2.5 NL

Pérez et al. (2016) Harbour 9-12% PM10 ES

Pérez et al. (2016) Harbour 11-15% PM2.5 ES

 References:
- Genoa (Italy): Mazzei et al. (2008)
- Melilla (Spain): Viana et al. (2009)
- Cork (Ireland): Hellebust et al. (2010)
- Algeciras (Spain): Pandolfi et al. (2011)
- Lampedusa (Italy): Becagli et al. (2012)
- Barcelona (Spain): Amato et al. (2009)
- Netherlands, UK, Belgium, Denmark,
France, Germany, Sweden, Norway,
Luxembourg, Switzerland: Hammingh et
al. (2012)
- UK: Hadley et al. (2016) 1%
PM10
1-7% PM10 1%

1-20% PM2.5 PM2.5

3%
8-11% PM1 PM1
4%
4% 5%
<1%
Comparability? <10% 3%
2%

2%

2% 1%
13-17% PM2.5 in China
Shanghai; Pearl River Delta 20%
Zhao et al. (2013); Tao et al. (2017)

10-70% PM2.5 in Western USA, Seattle 5% 6% 8%

Hadley (2017)
3-7%

5-10%

14%

2-4% 4% 8% 11%
 Impact on gaseous pollutants

 Fewer number of studies compared to PMx

 Broader spatial coverage across EU (dispersion modelling tools)

Gaseous pollutants
Shipping
Reference Species Location
contribution
Isakson et al. 106%* NO2 Gothenburg (SE)
(2001) 281%* SO2 Gothenburg (SE)
Keuken et al.
5-7 ppb NO2 Rotterdam (NL)
(2005)
North Sea coastal
7-24% NO2
countries
24% NO2 The Netherlands
19% NO2 Denmark
17% NO2 UK
Hammingh et al. 15% NO2 Belgium
(2012)
13% NO2 Norway
9% NO2 Sweden
8% NO2 France
7% NO2 Germany
7% NO2 Ireland
 References:
- Sweden: Isakson et al. (2001)
- Netherlands, Denmark, Germany,
France, Ireland: Hammingh et al. (2012)

NO2

SO2

9%
13%
280% increase with respect
to background levels
7-24% NO2 106% increase with respect
to background levels
19%

Comparability? 7%
17% 24%
7%
15%

8%
Contributions to gases (NO, NO2, SO2) > PM, N

Hotelling: contribution to SO2 < NO & NO2 due to low-S fuels at berth

Contribution to NO >> NO2 and provoked local-scale depletion of O3

Merico et al. (2016)
 Impact of harbour operations

 Knowledge gap! Loading and unloading of vessels, fuelling, etc.

 Studies agree on the relevance of this impact:
• S-Europe: road dust = 26% PM10 in harbours; harbours = 9-12%
urban PM10 (Pérez et al., 2016)
• Los Angeles harbour: vehicular sources + road dust = 54% of
PMx, vs. shipping < 5% of PM2.5 (Minguillon et al., 2008)
• Hotelling, manoeuvring (Merico et al., 2016)
 (2)
Mitigation strategies
 Mitigation strategies

IMO (UN), MARPOL,

SECAs, NECAs EU Directive 2005/33/EC on sulphur
emissions from ships

Technological measures:
• low sulphur fuels
• sulphur scrubbers
• NOx mitigation measures
• liquid natural gas (LNG)
• slow steaming
National regulations • soot particle filters…
 Mitigation strategies

Year 2010 << Years 2000-2006

(50%)

SO2 emissions reduction:

- >2006: use of low-S fuel due to the
Hadley (2017); Western USA SECA regulations in the North Sea
- >2007: MARPOL convention
- >2010: EU directive 2005/33/EC
Velders et al. (2011)
 Mitigation strategies

Directive 2005/33/EC:
• SO2 concentrations in 3 out of 4
harbours decreased (>2010)
• No decrease was observed in Tunis
• Average decrease SO2 = 66% (daily)
• No significant changes for NOx & BC
Schembari et al. (2012)

Sulphur reduction policy in the Baltic Sea SECA (2015):

• for the Baltic Sea only, the latest sulphur regulation is not cost-effective
• Expected annual cost = 465 M€
• Monetized benefit = 105 M€
Annturi et al. (2016)

“Alternative fuel’, “Ship design” or “Operation”:

Highest reductions = “Operation”, with GHG emissions 10% lower than BAU
Winnes et al. (2015)
 Case study: ECA in the Marmara Sea

23 million inhabitants
(and growing)
Istanbul Strait

Marmara Sea
50.000 vessels/year

Canakkale
Strait
 Rationale

The Turkish government aims to apply to International Maritime

Organization (IMO) for the Marmara Sea and the Turkish Straits to be
declared an Emission Control Area (ECA) for SOx

Health benefits modelling:

Only low-sulphur content fuels may be used (<0,1%)
US-EPA BenMAP CE

To support the application to IMO:

quantify the environmental and health benefits
which would derive from designating the Marmara Sea and
Turkish Straits as a sulphur ECA by the year 2020.
 Challenges using BenMAP
Pollutant data: modelled SO2, PM10, PM2.5 with CALPUFF
SO2 before ECA SO2 after ECA

PM10 before ECA PM10 after ECA

Highly spatially-resolved data

 Results: Environmental benefits

Istanbul: Air quality improvement

PM2.5 before ECA PM2.5 after ECA

5% 1.7%

Total PMx
Ship-sourced PMx

SO2 before ECA SO2 after ECA

46% 4.6%
Total SO2
Ship-sourced SO2
 Results: Health benefits

East domain (90% confidence intervals)

Health outcome Scenario PM10 PM2.5 SO2
Hospital admissions Baseline 13,000 18,000 1,200
for respiratory (total burden) (4,900 to 20,000) (6,800 to 20,000) (-830 to 3,200)
diseases Policy scenario 150 330 180
(ICD-10 J00-J99) (number avoided) (57 to 230) (125 to 370) (-108 to 460)
% Change -1% -2% -14%

Baseline 4,300 6,000 1,700

Hospital admissions
(total burden) (770 to 7,800) (1,900 to 9,700) (770 to 2,500)
for circulatory
Policy scenario 45 97 190
system diseases
(number avoided) (8.1 to 82) (30 to 160) (90 to 290)
(ICD-10 I00-I90)
% Change -1% -2% -12%

Baseline 120 670 17

All-cause mortality
(total burden) (50 to 190) (140 to 1,000) (15 to 19)
(ICD-10 A00-R99)
Policy scenario 1 13 2
(number avoided) (0.4 to 1.6) (2.7 to 19) (1.7 to 2.2)
% Change -1% -2% -10%
Viana et al. (2015)
 (3)
Conclusions
 Conclusions & knowledge gaps (1)

• What we know:
o Number of studies on the impact of shipping emissions on air
quality is not large, but increasing
o Impact on PMx, NOx, SO2, and new particle formation (N)
o Ultrafine particles and toxicity, better tracers than mass (?)
o Tracers are available: most commonly, V/Ni =3-5±1 in PM10 and
PM2.5
o Contribution to PMx: 1-20% PMx, with large spatial variability

• What we don’t know (so well):

o Particle size distribution
o Ratio primary to secondary particles? More efficient to reduce
primary emissions (BC, V, Ni…)?
o Discriminating sources with common tracers
o Impact of harbour operations & how to mitigate them
 Conclusions & knowledge gaps (2)

• Mitigation strategies are efficient: 50-66% SO2 reduction, and 2ary PM

• Cost effectiveness?

• Case study: potential improvements in Istanbul

• Environmental benefits: 5% to 2% reduction of ship-sourced PMx;
46% to 5% reduction of ship-sourced SO2 (annual means)
• Health benefits: 12-14% decreased hospital admissions due to SO2;
10% reduced mortality due to SO2; 1-2% decreased hospital
admissions due to PM2.5.
• Overall, beneficial policy from an environmental and health
perspective

• Limitations:
o Uncertainties in emissions modelling & AQ measurements
o Need for regionally-specific health impact functions
Thank you for your attention
mar.viana@idaea.csic.es