Transportation Research Part D 57 (2017) 237–250

Contents lists available at ScienceDirect

Transportation Research Part D

journal homepage: www.elsevier.com/locate/trd

Adapting the shipping sector to stricter emissions regulations: Fuel MARK

switching or installing a scrubber?
⁎
Luis María Abadiea, Nestor Goicoecheab, Ibon Galarragaa,c,
a
Basque Centre for Climate Change (BC3), Edificio sede 1, 1st floor, Parque Científico UPV-EHU, Sarriena s/n, 48930 Leioa, Spain
b
Escuela de Ingeniería de Bilbao, Universidad del País Vasco-Euskal Herriko Unibertsitatea (UPV-EHU), Ingeniero Torres Quevedo Plaza, 1, 48013
Bilbao, Bizkaia, Spain
c
Economics for Energy, Gran Vía 3, 3°E, 36204 Vigo, Spain

AR TI CLE I NF O AB S T R A CT

Keywords: This paper examines how the existing fleet in the shipping industry can be adapted to the new
Marine fuel emission regulations through the two main techniques that currently exist: (a) the use of low-
Uncertainty sulphur marine diesels; and (b) the installation of scrubbers. A method is presented here for
Sulphur emissions drawing up an economic assessment of both these techniques under uncertainty. It enables the
Technology decision
best option to be selected at any given time taking into account fuel prices (spot and futures),
Stochastic models
scrubber installation costs, the time that the vessel operates in an Emission Control Area (ECA)
and the remaining useful lifetime of the vessel. The paper also considers the possibility of an
unexpected change from a non-ECA navigation area to an ECA. The assessment is carried out in a
manner consistent with marine diesel and crude oil spot and futures market quotes. Our results
show the net present value of investing in the installation of scrubbers and investing in changing
fuel types for different assumptions on how vessels are operated. We also analyse increases in fuel
consumption and CO2 emissions as a consequence of using scrubbers and how they affects the
financial analysis if such incremental emissions must be paid under a CO2 pricing mechanism.

1. Introduction

Maritime transport has increased significantly over the past 40 years in terms of both volume and the number of ships operating.
This growth has been driven by an increase in the demand for transportation in an ever-more globalised economy. In fact the shipping
industry is responsible for carrying about 90 per cent of world trade (ICS, 2014).
According to the International Chamber of Shipping (ICS, 2014), the forecast is for around 17 billion tonnes to be carried in 2030,
while Global Marine Trends 2030 forecasts that the volume of seaborne trade will double from nine billion tonnes to around 20
billion tonnes by the same year.
The trends denote a continued increase in both seaborne trade and the size of the worldwide fleet. This increase in demand affects
the consumption, availability and price of marine fuel (UNCTAD, 2016).
Note that fuel consumption is a factor in the shipping industry mainly because it is the largest cost item in the operational
expenses (OPEX) of a vessel. Up to 50 per cent of the cost of a voyage is accounted for by fuel costs (Bialystocki and Konovessis, 2016;
Stopford, 2009; Ronen, 2011). Fuel consumption in this sector is also an important driver of air quality and greenhouse gas (GHG)
emissions.
Fig. 1 shows the estimated fuel consumption of the world fleet resulting from different activity-based estimates and statistics

⁎
Corresponding author at: Basque Centre for Climate Change (BC3), Edificio sede 1, 1st floor, Parque Científico UPV-EHU, Sarriena s/n, 48930 Leioa, Spain.
E-mail address: ibon.galarraga@bc3research.org (I. Galarraga).

http://dx.doi.org/10.1016/j.trd.2017.09.017

Available online 05 October 2017

1361-9209/ © 2017 Elsevier Ltd. All rights reserved.
L.M. Abadie et al. Transportation Research Part D 57 (2017) 237–250

Fig. 1. Worldwide fleet fuel. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Source: IMO, 2009, Second IMO GHG Study 2009, under IMO authorisation. “Symbols indicate the original estimates for individual years and the solid lines show the
original estimates of trend. Dashed lines show the backcast and forecast, calculated from the time evolution of freight tonne-miles with the point estimates. The blue
square shows the activity-based estimate from IMO (2009) and the blue range bar indicates the high and low bound estimates.”

Table 1
CO2 emissions from different modes of transport..
Source: own work based on the following data: http://www.
worldshipping.org/industry-issues/environment/air-emissions/
carbon-emissions and http://fluglaern.de/hamburg/klima.htm

Mode of Transport CO2 emissions

Aircraft 450–500 g
Truck 60–150 g
Train 30–100 g
Ship 10–40 g

(Corbett and Köhler, 2003; Eyring et al., 2005; Endresen et al., 2003; IMO, 2009). These estimates show a high degree of uncertainty
depending on the source of the study and the methods used. Two main approaches are used: (1) based on activity data (bottom-up
approach); and (2) based on fuel statistics (top-down approach).
Fuel consumption in the shipping industry is, as in other sectors, closely related to emissions into the atmosphere. Indeed, it has
been suggested that the international fleet contributes significantly to global anthropogenic emissions (Endresen et al., 2003). Its key
emissions are of carbon dioxide (CO2), nitrogen oxides (NOx), sulphur oxides (SOx) and volatile organic compounds (Hydrocarbons).
These pollutants are related to climate change and health problems, as well as to localised ocean acidification and fertilisation
(Hassellöv et al., 2013; Fuglestvedt et al., 2009). The adverse health effects of air pollution include premature mortality due to
cardiopulmonary diseases and lung cancer (Silva et al., 2016).
The shipping industry accounted for about 4 per cent of global CO2 emissions in 2007 (Reynolds, 2009), 2.8 per cent of the
world’s total GHG emissions in 2007 and 2.2 per cent in 2012 (ICS, 2014). The 3rd IMO GHG study (IMO, 2014) recognises that
maritime transport emits around 1000 million tonnes of CO2 annually and is responsible for about 2.5% of GHG emissions. Maritime
shipping is by far the most carbon-efficient form of trade (see Table 1).
Table 1 below shows the amount of CO2 in grams emitted per metric tonne of freight and per kilometre of transportation.
The International Maritime Organisation (IMO) is a specialised agency of the United Nations in charge of the safety and security of
shipping, and of preventing marine pollution by ships. One of its main roles is to implement and review international conventions for
shipping. The most important international convention related to environmental quality in the shipping sector is the International
Convention for the Prevention of Pollution from Ships (MARPOL).
As these regulations become stricter with respect to emissions, the shipping sector is considering all its fuelling options. These
include changing to low sulphur marine fuels, LNGs or biofuels, or even retrofitting their vessels and installing scrubbers1 (Boer and
Hoen, 2015). Adopting scrubbing technology enables sulphur oxides (SOx) and particulate matter (PM) to be removed. In this paper

1
Scrubbers are air pollution control devices that remove particulates and/or gases from exhaust streams.

238
L.M. Abadie et al. Transportation Research Part D 57 (2017) 237–250

we focus solely on the options available to the existing fleet, and therefore consider only the options of switching to low-sulphur
marine fuels and installing a scrubber. In addition to these two measures, other authors have considered reducing the speed of vessels
to reduce emissions (of CO2 in this case) (Corbett et al., 2009). High retrofitting costs mean that the remaining options are usually
only feasible for new vessels, which places them outside the scope of this paper. Other papers on the topic include Psaraftis and
Kontovas (2010), which examines environmental and economic trade-offs in the shipping industry.
The choice of one option or the other depends on various factors such as the price of fuels, the area in which the ship usually
operates, the regulations applicable to it, the number of days at sea and the remaining useful lifetime of the ship, among others. Some
of these factors are known but others involve a certain degree of uncertainty. In this paper we propose the use of stochastic modelling
to deal with uncertainties concerning the price of the different fuels, the possibility of an unexpected change from a non ECA
(Emission Control Area)2 navigation zone to an ECA and the incremental impact of CO2 allowances in the hypothetical case of the
European Trading Scheme (EU ETS) coming to include the shipping sector.3 These calculations enable us to compare the costs and
benefits of a fuel switch with those of installing a scrubber. The possibility of a new ECA designation is analysed by Panagakos et al.
(2014) in the case of the Mediterranean Sea and the possibility of extension to the entire European coastline is studied by Schinas and
Bani (2012).
This paper contributes to the literature by providing a stochastic model that enables future marine fuel distributions to be obtained
based on their spot prices and the crude oil futures market. This method is consistent with crude oil market quotes and the gaps between
marine fuels and crude oil. Expected future marine fuel prices could be used for hedging transactions. The work presented here over-
comes the problem of having to select scenarios, which can be limited and often not related to fuel market quotes. The estimates resulting
from the stochastic model are then used to decide between installing scrubbers and using more expensive marine fuels in ECAs, con-
sidering various factors such as the remaining lifetime of the vessel, the price of marine fuels, the area where the ship usually operates and
the number of days at the sea. With the 2020 sulphur cap coming up, deciding between the options available has become a very relevant
issue: the decision must be taken under uncertainty in regard to many factors and considering additional effects such as incremental fuel
consumption and emissions. These decisions usually imply investments that are irreversible, with a significant initial fixed cost and
uncertain benefits. Stochastic modelling and assessment can thus be a very powerful tool for guidance.
The paper does not set out mainly to analyse carbon emissions from shipping, but the model presented enables the incremental
emissions that would result if a scrubber is installed to be calculated, along with the incremental cost if those additional emissions
were subject to the EU ETS. Note that slow steaming or reducing the speed of vessels is not considered in this paper, but the authors
acknowledge that it could also reduce CO2 emissions, as stated in Cariou (2011).
The paper is organised as follows: after this general introduction, Section 2 looks at the regulation that affects emissions in the
maritime sector. Section 3 explores the existing literature on the topic. Section 4 presents the stochastic model, the analysis and the
results. The final section gives conclusions and suggestions for further research.

2. Regulations affecting the shipping sector

The IMO has been enforcing a new legislative framework aimed at achieving greener shipping and a sustainable global maritime
transportation system. Up to 1990 the IMO’s main area of work was in preventing and controlling pollution by oil, chemicals, sewage
and garbage. Most of the regulations on these points entered into force in 2001. Since then its regulatory work has been expanded to
include atmospheric protection, prevention of air pollution and reduction of GHG emissions. The technical and operational measures
related to energy-efficiency and marine environment adopted by the IMO are collected as regulations in various annexes of MARPOL
73/78.
In October 2016, the IMO decided to implement a new SOx regulation.4 This required a global sulphur cap of 0.50 per cent m/m to
be set by 2020. Note however, that this global cap does not change the limits for SOx in ECAs, which were set at 0.1 per cent m/m by
the IMO on 1 January 2015.
For the case of NOx emissions the regulatory limits depend on maximum engine operating speed.5
According to MARPOL (see Annex VI) there are two main areas of emissions and fuel quality requirements defined: one referring
to ECAs and the other global. Current ECAs include the Baltic Sea (SOx: adopted 1997, entered into force 2005; NOx: 2016/2021), the
North Sea (SOx: 2005/2006; NOx: 2016/2021), the North American ECA, including most of the US and Canadian coast (NOx & SOx:
2010/2012) and the US Caribbean ECA, including Puerto Rico and the US Virgin Islands (NOx & SOx: 2011/2014). Further designated
areas are under discussion, such as Australia, Japan, Singapore, the Mediterranean Sea and Mexico (Andersson and Brynolf, 2015).
Note that the possibility that new areas may be defined explains part of the analysis presented later in this paper. Other regulations
include the new Chinese regulation for coastal waters published in December 20156 and EU directive 2005/33/EC.7

2
Note that Emission Control Areas (ECAs) impose different environmental restrictions on different sailing zones around the globe. These may involve unexpected
regulatory changes.
3
Regulated by Directive 2009/29/EC.
4
See MARPOL Annex VI Fuel sulphur limits. Sulphur oxides (SOx) – Regulation 14.
5
See MARPOL Annex VI NOx Emission limits. Nitrogen Oxides (NOx) – Regulation 13.
6
This is the case for vessels that operate in the Pearl River delta, Yangtze River delta and Bohai Sea, where it is mandatory to use fuel containing less than 0.5
sulphur m/m.
7
This extended the 1.5 m/m sulphur limit to ferries operating to and from any EU port. The EU directive also sets a maximum limit of 0.1 m/m % on the sulphur
content of marine fuels used by vessels when they berth for more than two hours.

239
L.M. Abadie et al. Transportation Research Part D 57 (2017) 237–250

The first study on CO2 emissions was adopted in 1997. It led to the presentation of the “IMO study of Greenhouse Gas Emissions
from Ships” at the Marine Environment Protection Committee (MEPC) in 2000. Based on this study two policy mechanisms were set
to tackle GHG emissions: the Energy Efficiency Design Index (EEDI) and the Ship Energy Efficiency Management Plan (SEEMP). A
second IMO GHG study (IMO, 2009) was presented later to provide better foundation for future decisions. In 2014 the MEPC
approved the Third IMO GHG Study (IMO, 2014), which provided updated estimates.
However, the IMO has not yet decided on a specific policy mechanism for reducing CO2 emissions and supporting energy effi-
ciency.

3. Adapting to atmospheric emissions regulations

In the past decision-making with respect to engines in the shipping sector was focussed mainly on energy efficiency, but now it is
more oriented towards reducing emissions in engine combustion to avoid detrimental effects on the environment (Tormos-Martínez
et al., 2016).
Yewen et al. (2016) state that the latest regulation on sulphur emissions in ECAs forces shipping companies to optimise their
maritime bunker management (MBM). Their paper analyses MBM from an integrated point of view including bunker hedging and fuel
consumption by making optimal routing and speed decisions. Bengtsson et al. (2011) study different options related to SOx and NOx
abatement technologies and to fuel changes. They analyse numerous alternatives using life-cycle assessment (LCA). Jiang et al.
(2014) assess the costs and benefits of two sulphur abatement options: sulphur scrubbers and marine gas-oil. Using a bottom-up
approach, they conclude that if the remaining lifetime of a ship is less than 4 years it is not suitable for scrubber installation. Their
study does not consider the engine curve (i.e. the curve that relates fuel consumption to revolutions per minute), weather conditions
or routage.
Brynolf et al. (2014) focus on switching fuels and look at LNG, liquefied biogas (LBG), methanol and bio-ethanol as alternatives
from an LCA approach. They conclude that the impact on climate change is of the same order of magnitude as with the use of Heavy
Fuel Oil (HFO). Moreno-Gutiérrez et al. (2015) study and compare nine different bottom-up methods regarding fuel consumption and
emission.
Several papers look at scrubber-based solutions, which have the advantage of allowing high-sulphur fuel to be used, while
reducing SOx emissions by as much as 98 per cent (Kristensen, 2012). Scrubber technology includes three types of exhaust gas
cleaning systems: open loop, which uses only sea water; closed loop, which uses fresh water mixed with caustic soda; and hybrid.
Panasiuk and Turkina (2015) analyse the investments required for different scrubbers to comply with the SOx requirements of
MARPOL. They show that in general the open loop system is cheaper and smaller in size but that the capital costs of scrubbers,
operating costs and maintenance costs vary widely for different ship sizes, engine capacities and scrubber technologies. Using cash
flow modelling based on a particular ship, they conclude that using low-sulphur fuel together with the installation of a scrubber is the
best way of complying with regulations.
There are several sources of uncertainty: Cullinane and Bergqvist (2014) look at the difficulties of estimating the payback time of
a fuel switching option. The uncertainties that they consider include the number of operational hours inside ECAs, the cost of
abatement technologies and estimated future fuel costs.
In this context of several sources of uncertainty, more sophisticated methods are called for. There are many such methods,
including stochastic modelling of uncertain sources. Good examples include Trigeorgis (1996) and Dixit and Pindyck (1994) for
resource management, and more recently Abadie and Chamorro (2013) for energy assets, Abadie et al. (2017) for technological
change under uncertainty in the cement industry and Abadie et al. (2014) for carbon capture and storage. As far as the authors are
aware, this is the first time that these methods have been used to analyse investment decisions in the shipping industry.

4. The stochastic model

There are cases in which one commodity is strongly correlated with another for which long-maturity futures prices are available,
e.g. crude oil products. Cortazar et al. (2008) propose a multi-commodity model that enables long-maturity futures prices available
for one commodity to be used to reliably estimate futures prices for another using the so-called Kalman filter. In our case we propose a
different model to estimate future marine fuel prices based on future crude oil quotas and the stochastic differentials between those
marine fuels and crude oil. This is also a very robust way of obtaining estimates. Using this type of approach, Chamorro et al. (2012)
estimate historical electricity margins using spot natural gas, coal and carbon prices.

4.1. Stochastic crude oil, carbon and marine fuel prices

Crude oil and marine fuel prices are strongly correlated. Fig. 2 shows this price behaviour for Brent8 crude oil, IFO180,9 IFO380,10
LSMGO11 and ULSFO.12 (See Annex 1 for clarification with respect to marine fuels).

8
Brent prices converted to USD/tonne at a rate of 7.53 barrels/tonne (http://shipandbunker.com).
9
Intermediate fuel oil with a maximum viscosity of 180 Centistokes (< 3.5% sulphur).
10
Intermediate fuel oil with a maximum viscosity of 380 Centistokes (< 3.5% sulphur).
11
Low-sulphur (< 0.1%) Marine Gas Oil.

240
L.M. Abadie et al. Transportation Research Part D 57 (2017) 237–250

Fig. 2. Crude oil and marine fuel prices (from 10/19/2015 to 11/7/2016)..
Source: prepared by the authors using Brent crude oil prices from the Energy Information Administration (EIA) and marine fuel prices from Rotterdam Bunker prices
(http://shipandbunker.com)

Table 2 below also shows the correlation values for these prices. These values suggest that marine fuel prices are determined by
the crude oil price plus a differential (positive or negative). This being so, it can be shown that marine fuel prices are determined by
Eq. (1):
Sti = Mti + Bt , where i ∈ {IFO380,IFO380,LSMGO,ULSFO} (1)
where Sti is the price of marine fuel i in the time t, Mti is the gap in price between marine fuel i and the Brent crude oil price Bt at time t.
Note that Mti and Bt are modelled as correlated stochastic processes.
Fig. 3 below shows the gap in prices for different marine fuels using Eq. (1). Fig. 3 suggests that marine fuel price gaps follow a
mean-reverting stochastic process. For a more comprehensive understanding of the gaps see the statistics in Table 3.
The correlation between Brent and marine fuel price gaps is shown in Table 4. The table shows that when Brent crude oil price
rises the positive gap between ULSFO and LSMGO usually also increases, while the negative gap between IFO180 and IFO 380
decreases.
As noted in Fig. 3, the gaps are very volatile but revert quickly to an equilibrium value. Furthermore, gaps can be positive or
negative and in the ULSFO case the gap can change signs. The Ornstein-Uhlenbeck processes defined by Eq. (2) as follows are
stochastic mean-reverting processes that enable this behaviour.
dMti = k i (M ∗i−Mti ) dt + σ idW i (2)
where M ∗iis the long-term equilibrium value towards which Mti ki
tends to revert in the long-term, is the reversion speed, is the σi
volatility and dW i = ε (t ) dt is the increment of a Wiener process with ε (t ): N (0,1) .
Eq. (2) is the continuous-time version of a first-order autoregressive process AR(1). The corresponding discrete-time version is

Table 2
Brent and marine fuel correlations.

Brent ULSFO IFO180 IFO380 LSMGO

Brent 1.000
ULSFO 0.952 1.000
IFO180 0.912 0.937 1.000
IFO380 0.923 0.947 0.992 1.000
LSMGO 0.973 0.976 0.932 0.939 1.000

241
L.M. Abadie et al. Transportation Research Part D 57 (2017) 237–250

Fig. 3. Marine fuel price gaps (from 10/19/2015 to 11/7/2016).

Table 3
Summary statistics of marine fuel price gaps.

Marine Fuel Mean Median Minimum Maximum

ULSFO 30.487 32.647 −23.941 79.614

IFO180 −92.625 −92.07 −130.94 −46.01
IFO380 −119.76 −121.98 −151.51 −73.51
LSMGO 54.116 53.558 15.82 91.123

Marine Fuel Std. Dev. C.V. Skewness Ex. kurtosis

ULSFO 18.989 0.62286 −0.14201 −0.099297

IFO180 19.194 0.20723 −0.11138 −0.71919
IFO380 17.909 0.14953 0.24721 −0.86147
LSMGO 15.641 0.28903 −0.13191 −0.57273

Marine Fuel 5 per cent percentile 95 per cent percentile IQ range

ULSFO 0.05 61.738 25.54

IFO180 −126.08 −61.734 28.64
IFO380 −147.18 −88.734 29.532
LSMGO 26.255 78.391 23.277

Table 4
Brent and marine fuel price gap correlation.

Brent ULSFO IFO180 IFO380 LSMGO

Brent 1.000
ULSFO 0.402 1.000
IFO180 −0.289 0.364 1.000
IFO380 −0.324 0.374 0.955 1.000
LSMGO 0.547 0.765 0.218 0.193 1.000

242
L.M. Abadie et al. Transportation Research Part D 57 (2017) 237–250

Table 5
Marine fuel price gap parameters.

ULSFO IFO180 IFO380 LSMGO

ai 5.618 −8.359 −10.771 8.985

bi 0.819 0.908 0.908 0.834
σεi 11.123 7.907 8.467 8.694
M ∗i 30.982 −90.906 −117.678 54.246
ki 50.420 24.309 24.191 45.632
σi 194.498 131.595 140.949 150.674
M0i 71.99 −46.01 −73.51 76.99

depicted in Eq. (3).

i i
Mti+ Δt = M ∗i [1−e−k Δt ] + Mti e−k Δt + εti+ Δt = ai + biMti + εti+ Δt (3)
As shown in Abadie and Chamorro (2013), the parameter values can be obtained from Eqs. (4)–(6).
ai
M ∗i =
1−bi (4)

lnbi
ki = −
Δt (5)

2(σ iε)2lnbi
(σ i)2 =
Δt [(bi )2−1] (6)
where the values ai , bi σεi
and are estimated with the Ordinary Least Squares (OLS) model.
1
Assuming 252 trading days per year, then Δt = 252 . The results of the estimation are shown in Table 5.
It can be seen that the long-term equilibrium values M ∗i are nearest to the Table 3 mean values. The M0i values are the gap on the
last day of the series (11/07/2016), and these figures are the initial values.
For the Brent crude oil price we use a somewhat different mean-reverting stochastic process: an inhomogeneous geometric
Brownian motion process (igbm). Now with Bt in the stochastic part of Eq. (7) it is not possible to obtain negative values.
dBt = k B (B∗−Bt ) dt + σ BBt dW B (7)
The figures for past Brent crude oil volatility can be obtained using Eq. (8), which is the discrete version of Eq. (7).
Bt + Δt −Bt 1
= −k B Δt + k BL∗Δt + σ B Δt εtB
Bt Bt (8)
Note the different concept of this volatility. In this model σ B is the volatility of returns Bt + Δt − Bt while in marine fuel oil price gaps
Bt
it is the price gap volatility. That is, the former refers to the volatility of the returns and the latter to the volatility of prices.
B
Using Eq. (8) the annualized return volatility of Brent crude oil can be calculated as σ = 0.441.
For carbon prices we use a geometric Brownian motion model, as in Eq. (9).
dAt = αAt dt + σ AAt dW A (9)
where α is the trend and σ is its volatility.
A

Eq. (10) below is the discrete version.

At + Δt −At
= α Δt + σ A Δt εtA
At (10)
Carbon prices in €/tonne are converted to $/tonne using the exchange rate of the European Central Bank (ECB). Eq. (10) enables
the annualised return volatility of carbon to be calculated as σ A = 0.466.

Table 6
Brent, carbon and marine fuel price gap correlations.

Brent ULSFO IFO180 IFO380 LSMGO Carbon

Brent 1.000
ULSFO −0.536 1.000
IFO180 −0.833 0.613 1.000
IFO380 −0.850 0.640 0.927 1.000
LSMGO −0.560 0.613 0.650 0.705 1.000
Carbon −0.089 0.082 0.057 0.078 0.075 1.000

243
L.M. Abadie et al. Transportation Research Part D 57 (2017) 237–250

Fig. 4. Brent Crude Oil Futures Prices (from 1/2/2014 to 11/7/2016).

The correlations can now be calculated using the residual of the regressions. These results are shown in Table 6.
In all cases the Brent crude oil return are negatively correlated with marine fuel price gaps. This is explained by the denominator
of Bt + Δt − Bt . The correlations with carbon return prices are very low, so we assume zero correlation in this case.
Bt
To calculate the remaining parameters of the Brent crude oil and carbon stochastic processes we use a different approach based on
Brent futures market price quotes. The approach is explained below.
In the futures market (risk-neutral world) the Brent crude oil price behaves a little differently. This is mainly because the market
price of risk must be subtracted.
Eq. (11) shows this behaviour where λBBt is the market price of risk.

dBt = [k B (B∗−Bt )−λBBt ] dt + σ BBt dW B (11)

In this equation B∗ is the long-term equilibrium value towards which Bt tends to revert in the long term in the real word and
k BB∗
B∗∗ = is the equivalent value in the risk-neutral world.
k B + λB

Fig. 5. Brent Crude Oil Futures Prices (11/7/2016).

244
L.M. Abadie et al. Transportation Research Part D 57 (2017) 237–250

Table 7
Parameters of stochastic processes for Brent and Carbon.

Brent Carbon

k B + λB 0.5289 α A−λ A 0.0068

B** ($/tonne) 442.311 –
σB 0.441 σA 0.451
B0 ($/tonne) 322.510 A0 (€/tonne) 6.230

Fig. 4 shows the mean-reverting behaviour of Brent prices for all the days considered. The last day’s futures prices (11/07/2016)
can be seen in Fig. 5.
Fig. 5 shows the actual futures Brent prices for the last day of the series and the estimated function using ordinary non-linear least
squares.
Similarly the carbon price is shown by Eq. (12) in the risk-neutral world (futures market).
dAt = (α A−λ A) At dt + σ AAt dW A (12)
The Brent price is converted from $/barrel to $/tonne at a rate of 7.53 barrels per tonne according to data on marine fuels from
Ship and Bunker.13.
We take the last day of the price series (11/07/2016) as our initial values.
Table 7 shows all the parameters that result from the calibration process.

4.2. The Net Present Value (NPV) of marine fuel in tonnes/year

As shown in Abadie and Chamorro (2013) the expected value at time t of Brent crude oil under the equivalent martingale measure
takes the form shown in Eq. (13).
k BB∗ B B B B B B B B
E Q (Bt ) = (1−e−(k + λ ) t ) + B0 e−(k + λ ) t = B∗∗ (1−e−(k + λ ) t ) + B0 e−(k + λ ) t .
k B + λB (13)
Calculating the value of an annuity between times τ1 and τ2 requires computing the integral in Eq. (14):
τ2
PV = ∫τ 1
E Q (Bt ) e−rt dt
(14)
which yields Eq. (15),
B0−B∗∗ B B B B B∗∗ −rτ1 −rτ2
NPV (B0,τ1,τ2) = (e−(k + λ + r ) τ1−e−(k + λ + r ) τ2) + (e −e )
kB B
+λ +r r (15)
If τ1 = 0 then Eq. (16) emerges:
B0−B∗∗ B B B∗∗
NPV (B0,0,τ2) = (1−e−(k + λ + r ) τ2) + (1−e−rτ2)
kB B
+λ +r r (16)
And similarly for the marine fuel price gap Eq. (17) is obtained:

M0i −M i ∗ i M i∗
NPV (M0i ,τ1,τ2) = (1−e−(k + r ) τ2) + (1−e−rτ2)
i
k +r r (17)
And for the marine fuel price Eq. (18) is obtained:

B0−B∗∗ B B B∗∗ M i −M i ∗ i M i∗
NPV (i,0,τ2) = (1−e−(k + λ + r ) τ2) + (1−e−rτ2) + 0i (1−e−(k + r ) τ2) + (1−e−rτ2)
kB B
+λ +r r k +r r (18)
14
For r we use the yield of US bonds to 20 years from 11/07/2016, which gives r = 0.0226 for marine fuels. For emission
allowances we use the yield to 20 years of AAA-rated euro area government bonds from 11/07/2016,15 which gives r = 0.00672. It
can be seen that this rate is almost identical to the trend for the carbon futures market: α A−λ A = 0.0068.
For a carbon annuity Eq. (19) is obtained;
A0 A
NPV (A0 ,0,τ2) = (e (α − λ − r ) τ2−1)
α−λ A−r (19)
Table 8 shows the present value of fuel costs for 1 tonne/year throughout the remaining lifetime of a ship. Note that low-sulphur

12
Ultra-Low Sulphur Fuel Oil (< 0.1%).
13
http://shipandbunker.com.
14
https://www.treasury.gov.
15
https://www.ecb.europa.eu/stats/money/yc/html/index.en.html.

245
L.M. Abadie et al. Transportation Research Part D 57 (2017) 237–250

Table 8
Marine fuel annuity ($/tonne year).

Remaining lifetime (years) ULSFO IFO180 IFO380 LSMGO LSMGO-IFO180 LSMGO-IFO380

5 2035.0 1459.8 1333.2 2144.7 684.9 811.5

10 4020.7 2930.8 2691.1 4228.6 1297.8 1537.5
15 5804.9 4255.3 3914.6 6100.5 1845.2 2185.9
20 7399.1 5438.9 5008.1 7773.1 2334.2 2765.0
25 8823.1 6496.1 5984.8 9267.1 2771.0 3282.3
30 10,094.9 7440.4 6857.1 10,601.3 3160.9 3744.2

marine diesel such as ULSFO has a high impact on maritime transport costs.
According to the IMO (2014) the emission factor for HFO and low-sulphur fuel oil is 3114 kg/tonne. Using Eq. (19) and mul-
tiplying by 3.114 it is possible to calculate the emission cost arising from the consumption of one tonne of marine fuel per year.
Results are presented in Table 9.
Table 9 shows that the Net Present Value (NPV) of emission costs of 1 tonne of fuel/year with a remaining useful lifetime of
30 year is €582.7/tonne/year. This would be the case for marine companies that have to buy all their emission allowances in the EU
ETS. This amount is converted to €/tonne of fuel per year using the exchange rate $1.11/€ (set by the European Central Bank on 11/
07/2016). This gives $646.8/tonne of fuel per year.

4.3. The Net Present Value (NPV) of total marine fuel consumption during the remaining useful lifetime

We use the tonnes/day consumption figures from Green Ship of the Futures (2012) as presented in Table 10 below.
We now assume two possible behaviours: one in which the ship is at sea 80 per cent of the time and in harbour 20 per cent (10 per
cent idling and 10 per cent unloading), and the other in which the ship is at sea 70 per cent of the time and in harbour 30 per cent (15
per cent idling and 15 per cent unloading). Using the percentage of time that a ship stays in ECAs (pECA), we obtain the figures shown
in Table 11.
Table 11 shows how many tonnes of fuel are used under each scenario of time at sea using scrubbers and switching fuels for both
types of fuel and two points in time: in the initial period (up to the end of 2019) and in the second period (2020 and beyond). Note
that in the initial period, when the fuel switching option is considered different percentages of low-sulphur could be used depending
on how long is spent in ECAs. During the second period, low-sulphur fuel should always be used.
By contrast, if scrubbers are installed the ship is allowed to consume high-sulphur fuels. In this case the consumption is a little
greater than in the fuel-switching case due to the additional energy that is required to operate the scrubber, as shown in Table 12.
Taking the present time as the end of 2016, the Total Net Present Value (TNPV) of fuel cost is given by Eq. (20).

TNPV (i,τ2) = cini NPV (i,0,3) + c fii NPV (i,3,τ2) (20)

where is the initial annual consumption of marine fuel i in the years 2017–2019 and is the final annual consumption of marine
cini c fii
fuel i in year 2020 and beyond, taken from Table 11. Note that the TNPV depends on the remaining useful lifetime τ2 .
Table 12 shows the present value of fuel costs for both options, assuming that 80 per cent of the time is spent at sea. Note that the
last two columns show the savings in fuel cost when the option of installing a scrubber is decided on. For a financial assessment this
cost saving can be compared with the cost of investing in a scrubber. The equivalent calculations for 70 per cent of time spent at sea
are shown in Table 13.
The information provided by these tables provides an understanding of the effects that variables such as the remaining lifetime of
a ship, the percentage of time that vessels spend in ECAs and the percentage of time spent at sea have on investment decisions.
Logically, the longer the remaining lifetime of the vessel is, the longer vessels spend in ECAs and the longer they spend at sea, the
more attractive the option of investing in scrubbers becomes. However, it is also noted that when scrubbers are used fuel con-
sumption is higher, so emissions per year are also higher. Table 14 shows the extent of these additional emissions as a result of higher
fuel consumption.
Tables 15 and 16 show the net impact of considering the increases in emissions for vessels that spend 80 per cent and 70 per cent
of their time at sea. The net impact is calculated assuming that emissions are paid for at the price of EU ETS allowances. This is to

Table 9
Marine fuel emissions per tonne/year.

Remaining lifetime (years) Cost ($) Cost (€)

5 107.7 97.0
10 215.5 194.1
15 323.2 291.2
20 431.0 388.3
25 538.9 485.5
30 646.8 582.7

246
L.M. Abadie et al. Transportation Research Part D 57 (2017) 237–250

Table 10
Daily consumption by vessels.

Consumption by vessels (tonnes/day) Additional consumption by scrubbers (tonnes/day)

HFO Low Sulphur

At Sea 32.4 30.5 0.8

Main engine 28.7 27.0
Auxiliary engine 3.7 3.5

In Harbour
Idling 4.3 4.1 0.2
Unloading 4.3 4.1 0.4

Table 11
Fuel consumption in tonnes according to percentage of time in ECAs and at sea

Percentage in ECA 80 per cent at sea 70 per cent at sea

Switching Fuel Type (tonnes) With Scrubbers (tonnes) Switching Fuel Type (tonnes) With Scrubbers (tonnes)

Low Sulphur High Sulphur Low Sulphur High Sulphur Low Sulphur High Sulphur Low Sulphur High Sulphur

100 Year < 2020 9205.3 0.0 0.0 10,030.2 8241.7 0.0 0.0 8986.3
100 Year ≥ 2020 9205.3 0.0 0.0 10,030.2 8241.7 0.0 0.0 8986.3
75 Year < 2020 6904.0 2443.7 0.0 9966.3 6181.3 2187.3 0.0 8927.0
75 Year ≥ 2020 9205.3 0.0 0.0 10,030.2 8241.7 0.0 0.0 8986.3
50 Year < 2020 4602.6 4887.3 0.0 9902.4 4120.9 4374.5 0.0 8867.7
50 Year ≥ 2020 9205.3 0.0 0.0 10,030.2 8241.7 0.0 0.0 8986.3
25 Year < 2020 2301.3 7331.0 0.00 9838.6 2060.4 6561.8 0.0 8808.4
25 Year ≥ 2020 9205.3 0.00 0.00 10,030.2 8241.7 0.0 0.0 8986.3
0 Year < 2020 0.00 9774.7 0.00 9774.7 0.00 8749.1 0.0 8749.1
0 Year ≥ 2020 9205.3 0.00 0.00 10,030.2 8241.7 0.0 0.0 8986.3

Table 12
Net Present Value in $ for fuel switching versus scrubber installation (80 per cent at sea)

Percentage in Remaining Fuel Switching Using Scrubbers Differences between

ECA Lifetime (years) Switching & Scrubbers

LSMGO + IFO LSMGO + IFO High-Sulphur IFO High-Sulphur IFO High-Sulphur IFO High-Sulphur IFO
180 360 180 360 180 360

100 10 38,925,455 38,925,455 29,396,028 26,992,345 9,529,428 11,933,111

20 71,554,051 71,554,051 54,553,370 50,232,395 17,000,680 21,321,656
30 97,588,835 97,588,835 74,628,772 68,778,337 22,960,063 28,810,498

50 10 37,234,424 36,854,803 29,287,983 26,894,227 7,946,440 9,960,576

20 69,863,019 69,483,398 54,445,326 50,134,277 15,417,693 19,349,121
30 95,897,803 95,518,183 74,520,727 68,680,219 21,377,076 26,837,964

0 10 35,543,603 34,784,354 29,180,023 26,796,186 6,363,580 7,988,168

20 68,172,198 67,412,949 54,337,366 50,036,236 13,834,832 17,376,713
30 94,206,983 93,447,734 74,412,767 68,582,178 19,794,215 24,865,556

illustrate the hypothetical case of the shipping industry being included in the sectors regulated by the EU ETS or a similar scheme. It is
currently not affected by EU ETS but the European Commission has been calling for approaches to reduce emissions from this growing
sector. In fact in 2018 large ports in the EU will be required to report emissions.16

5. Conclusions and further research

In the current context of an international community that is seeking to reduce pollutant emissions from the shipping sector,
regulation is becoming stricter. Changes in legislation may have a significant impact on investment decisions, particularly with
respect to fuelling alternatives. Among the options that the sector is considering two main approaches can be distinguished: one for

16
Regulation (EU) 2015/757 of the European Parliament and of the Council of 29 April 2015 on the monitoring, reporting and verification of carbon dioxide
emissions from maritime transport, and amending Directive 2009/16/EC.

247
L.M. Abadie et al. Transportation Research Part D 57 (2017) 237–250

Table 13
Net Present Value in $ of fuel switching versus scrubber installation (70 per cent at sea)

Percentage in Remaining Fuel Switching Using Scrubbers Differences between

ECA Lifetime (years) Switching & Scrubbers

LSMGO + IFO LSMGO + IFO High-Sulphur IFO High-Sulphur IFO High-Sulphur IFO High-Sulphur IFO
180 360 180 360 180 360

100 10 34,850,784 34,850,784 26,336,616 24,183,098 8,514,169 10,667,687

20 64,063,857 64,063,857 48,875,691 45,004,423 15,188,166 19,059,433
30 87,373,351 87,373,351 66,861,731 61,620,184 20,511,620 25,753,166

50 10 33,335,870 32,996,082 26,236,349 24,092,043 7,099,521 8,904,038

20 62,548,943 62,209,154 48,775,424 44,913,369 13,773,519 17,295,785
30 85,858,437 85,518,648 66,761,464 61,529,130 19,096,972 23,989,518

0 10 31,820,914 31,141,329 26,136,082 24,000,989 5,684,832 7,140,341

20 61,033,987 60,354,401 48,675,157 44,822,314 12,358,830 15,532,087
30 94,206,983 93,447,734 74,412,767 68,582,178 17,682,283 22,225,820

Table 14
Additional fuel consumption and additional emissions when scrubbers are used (tonnes)

Percentage in ECA 80 per cent at sea 70 per cent at sea

Fuel Consumption Emissions Fuel Consumption Emissions

100 824.9 2568.7 744.6 2318.7

Year < 2020
100 824.9 2568.7 744.6 2318.7
Year ≥ 2020
75 618.6 1926.3 558.4 1738.9
Year < 2020
75 824.9 2568.7 744.6 2318.7
Year ≥ 2020
50 412.5 1284.5 372.3 1159.3
Year < 2020
50 824.9 2568.7 744.6 2318.7
Year ≥ 2020
25 206.3 642.4 186.2 579.8
Year < 2020
25 824.9 2568.7 744.6 2318.7
Year ≥ 2020
0 0.0 0.0 0.0 0.0
Year < 2020
0 824.9 2568.7 744.6 2318.7
Year ≥ 2020

Table 15
Net impact ($) with scrubbers and emissions (80 per cent at sea)

Percentage in ECA Remaining Lifetime Differences in Changing to Scrubbers

(years)
High-Sulphur IFO High-Sulphur IFO Emission Cost Net with High-Sulphur Net with High-Sulphur
180 360 IFO 180 IFO 360

100 10 9,529,428 11,933,111 176,106 9,353,322 11,757,005

20 17,000,680 21,321,656 352,353 16,648,327 20,969,303
30 22,960,063 28,810,498 528,741 22,431,322 28,281,757

50 10 7,946,440 9,960,576 149,701 7,796,739 9,810,875

20 15,417,693 19,349,121 325,948 15,091,745 19,023,173
30 21,377,076 26,837,964 502,336 20,874,740 26,335,628

0 10 6,363,580 7,988,168 123,289 6,240,291 7,864,879

20 13,834,832 17,376,713 299,536 13,535,296 17,077,177
30 19,794,215 24,865,556 475,924 19,318,291 24,389,632

new vessels and the other for existing ships. Logically, the options may differ substantially from one approach to the other due mainly
to differences in the remaining lifetimes of vessels, technological possibilities and costs. It this paper we focus on the existing fleet and
therefore consider two alternatives: (1) switching to low-sulphur marine fuels; and (2) installing scrubbers.

248
L.M. Abadie et al. Transportation Research Part D 57 (2017) 237–250

Table 16
Net impact with scrubbers and emissions (70 per cent at sea)

Percentage in ECA Remaining Lifetime Differences in Changing to Scrubbers

(years)
High-Sulphur IFO High-Sulphur IFO Emission Cost Net with High-Sulphur Net with High-Sulphur
180 360 IFO 180 IFO 360

100 10 8,514,169 10,667,687 158,963 8,355,206 10,508,724

20 15,188,166 19,059,433 318,053 14,870,113 18,741,380
30 20,511,620 25,753,166 477,271 20,034,349 25,275,895

5 10 7,099,521 8,904,038 135,125 6,964,396 8,768,913

20 13,773,519 17,295,785 294,215 13,479,304 17,001,570
30 19,096,972 23,989,518 453,433 18,643,539 23,536,085

0 10 5,684,832 7,140,341 111,287 5,573,545 7,029,054

20 12,358,830 15,532,087 270,378 12,088,452 15,261,709
30 17,682,283 22,225,820 429,595 17,252,688 21,796,225

We show that there several factor that can affect this decision, including (a) fuel prices; (b) the area in which the ship usually
operates and the regulation by which it is affected; (c) the number of days at sea; and (d) the remaining useful lifetime of the ship. We
also argue that as some of these factors a certain degree of uncertainty sophisticated methods are needed to account for that un-
certainty in the financial assessment of investment decisions. Stochastic modelling is proposed in this paper as a way of dealing with
uncertainties in fuel prices, areas of navigation or even a hypothetical cost for incremental CO2 emissions because of the use of
scrubbers.
We apply a stochastic model that enables future marine fuel distributions to be obtained based on their spot prices and the futures
market for crude oil. Our method is consistent with current and future crude oil and marine diesel prices. And these prices can be
hedged, at least partially, in the market. This stochastic diffusion model is calibrated with marked data.
As stated earlier, this approach overcomes the problem of having to select scenarios. In the context of the future 2020 sulphur cap
many decisions have to be made under uncertainty concerning several factors. These decisions involve serious investments with
considerable fixed initial costs. Using stochastic modelling makes for optimal decision making under uncertainty. In addition, the
method enables the incremental emissions of installing a scrubber and the incremental cost if those emissions are subject to EU ETS to
be calculated. As mentioned, other alternatives such as slow steaming or reducing speed are not considered.
We show calculations to compare the possibility of investing in the installation of scrubbers with investing in fuel switching under
different assumptions. The numbers show that the remaining lifetime of the vessel is a determinant factor in opting for one alternative
or the other, as indeed are time spent in ECAs and time spent at sea. In particular, the longer the remaining lifetime of a ship is, the
longer it sails in ECAs, and the longer it spends at sea, the more attractive investing in scrubbers becomes under current regulations.
However, we also note that when scrubbers are installed both fuel consumption and emissions increase slightly, so if CO2 emissions
have to be paid for (e.g. if the shipping sector is included in the EU ETS system or a similar CO2 pricing system) then the numbers may
change.
The method presented in this paper can be applied to many other sources of uncertainty and assumptions that can affect in-
vestment decisions. In this case we illustrate the case of existing vessels but the method could also be extended to the analysis of new
ships. Furthermore, alternative policy options to regulate emissions of CO2 and other pollutants could also be tested with this method,
as it offers a highly detailed illustration of the impacts that may result for alternative investment decisions. Among those alternatives
carbon taxing and global emission trading schemes could be considered. And finally, the method can easily be adapted for analysing
investment decisions in other economic sectors.

Acknowledgements

Luis M. Abadie and Nestor Goicoechea are grateful for financial support from the Spanish Ministry of Science and Innovation
(ECO2015-68023). Luis M. Abadie and Ibon Galarraga are grateful for financial support received from the Basque Government via
project GIC12/177-IT-399-13. Ibon Galarraga acknowledges the support of the Spanish Ministry of Economics and Competitiveness
(Project ECO2013-41183-P: Economía de la Eficiencia Energética en el Sector Residencial y de Transporte en España.

References

Abadie, L.M., Chamorro, J.M., 2013. Investments in Energy Assets Under Uncertainty. Springer, London, UK.
Abadie, L.M., Goicoechea, N., Galarraga, I., 2017. Carbon risk and optimal retrofitting in cement plants: an application of stochastic modelling, Monte Carlo simulation
and Real Options Analysis. J. Cleaner Prod. 142, 3117–3130. http://dx.doi.org/10.1016/j.jclepro.2016.10.155.
Abadie, L.M., Galarraga, I., Rübbelke, D., 2014. Evaluation of two alternative carbon capture and storage technologies: a stochastic model. Environ. Model. Softw. 54
(2014), 182–195.
Andersson, K., Brynolf, S., 2015. Marine fuel alternatives for a low carbon future – market influence on the pathways selected, low carbon shipping & shipping in
changing climates. In: A Research Led Consortium on Sustainable shipping, SCC Conference 2015 (last accessed October 2016).
Bengtsson, S., Andersson, K., Fridell, E., 2011.A comparative life cycle assessment of marine fuels liquefied natural gas and three other fossil fuels. Proc. Inst. Mech.
Eng. Part M: J. Eng. Marit. Environ. 225(2), 97–110. http://dx.doi.org/10.1177/1475090211402136.

249
L.M. Abadie et al. Transportation Research Part D 57 (2017) 237–250

Bialystocki, N., Konovessis, D., 2016. On the estimation of shiṕs fuel consumption and speed curve: a statistical approach. J. Ocean Eng. Sci. 1 (2016), 157–166. http://
dx.doi.org/10.1016/j.joes.2016.02.001.
Boer, E., Hoen, M., 2015. Scrubbers–an economic and ecological assessment. Report Delft, March 2015. Publication code: 15.4F41.20.
Brynolf, S., Fridell, E., Andersson, K., 2014. Environmental assessment of marine fuels: liquefied natural gas, liquefied biogas, methanol and bio-methanol. J. Cleaner
Prod. 74 (2014), 86–95. http://dx.doi.org/10.1016/j.jclepro.2014.03.052.
Cariou, P., 2011. Is slow steaming a sustainable means of reducing CO2 emissions from container shipping?. Transp. Res. Part D: Transp. Environ. 16 (3), 260–264.
Chamorro, J.M., Abadie, L.M., de Neufville, R., Ilic, M., 2012. Market-based valuation of transmission network expansion. A heuristic application in GB. Energy 44 (1),
302–320. http://dx.doi.org/10.1016/j.energy.2012.06.028.
Corbett, J.J., Wang, H., Winebrake, J.J., 2009. The effectiveness and costs of speed reductions on emissions from international shipping. Transp. Res. Part D: Transp.
Environ. 14 (8), 593–598.
Corbett, J.J., Köhler, H.W., 2003. Updated emissions from ocean shipping. J. Geophys. Res. 108, D204650. http://dx.doi.org/10.1029/2003JD003751.
Cortazar, G., Milla, C., Severino, F., 2008. Multicommodity model of futures prices: using futures prices of one commodity to estimate the stochastic process of another.
J. Futures Markets 28 (6).
Cullinane, K., Bergqvist, R., 2014. Emission control areas and their impact on maritime transport. Transp. Res. Part D 0. http://dx.doi.org/10.1016/j.trd.2013.12.004.
Directive 2009/29/EC of The European Parliament and of the Council of April 2009 amending Directive 2003/87/EC so as to improve and extend the greenhouse gas
emission allowance trading scheme of the community.
Directive 2005/33/EC of The European Parliament and of the Council of 2005 amending Directive 1999/32/EC as regards the sulphur content of marine fuels.
Dixit, A.K., Pindyck, R.S., 1994. Investment Under Uncertainty. Princeton University Press.
Endresen, O., Sorgard, E., Sundet, J.K., Dalsoren, S.B., Isaksen, I.S.A., Berglen, T.F., Gravir, G., 2003. Emission from international sea transportation and environmental
impact. J. Geophys. Res. 108 (D17), 4560. http://dx.doi.org/10.1029/2002JD002898.
European Commission, 2016.Reducing emissions from transport: A European Strategy for low-emission mobility. < https://ec.europa.eu/clima/policies/
transport_en > .
Eyring, V., Köhler, H.W., Van Aardenne, J., Lauer, A., 2005. Emissions from International Shipping: the last 50. Years. J. Geophys. Res. 110, D17305. http://dx.doi.
org/10.1029/2004JD005619.
Fuglestvedt, J., Eyring, V., Isaksen, I., Lee, D., Sausen, R., 2009. Shipping emissions: from cooling to warming of climate-and reducing impacts on health. Environ. Sci.
Technol. 43 (24), 9057–9062.
Global Marine Trends 2030. Lloyd’s Register Group Limited, QuinetiQ and University of Strathclyde Engineering ISBN 978-6-9575904-0-3.
Green Ship of the Future 2012. < http://www.greenship.org/fpublic/greenship/dokumenter/Downloadspercent20-percent20maga/ECApercent20study/
GSFpercent20ECApercent20paper.pdf > .
Hassellöv, I., Turner, D., Lauer, A., Corbett, J., 2013. Shipping contributes to ocean acidification. Geophys. Res. Lett. 40, 2731–2736.
ICS, (International Chamber of Shipping), 2014. Shipping, World Trade and the Reduction of CO2 Emissions. United Nations Framework Convention on Climate
Change (UNFCC).
IMO 2009, Second IMO GHG Study 2009, International Maritime Organization (IMO), London, UK, April 2009; Buhaug, O., Corbett, J.J., Endersen, O, Eyring, V.
Faber, J., Hanayama, S., Lee, D.S., Lee, D., Lindstad, H., Markowska, A.Z., Mjelde, A., Nelissen, J., Palsonnon, C., Winebrake, J.J., Wu, W., Yoshida, K., < http://
www.imo.org/en/OurWork/Environment/PollutionPrevention/AirPollution/Pages/Greenhouse-Gas-Study-2009.aspxlast accessedJanuary2017 > . (last accessed
September 2016).
IMO 2014, Third IMO GHG Study 2014; International Maritime Organization (IMO) London, UK, April 2015; Smith, T.W.P., Jalkanen, J.P., Anderson, B.A., Corbett, J.
J., Faber, J., Hanayama, S., O’Keeffe, E., Parker, S., Johansson, L., Aldous, L., Raucci, C., Traut, M., Ettinger, S., Nelissen, D., Lee, D. S., Ng, S., Agrawal, A.,
Winebrake, J. J., Hoen, M., Chesworth, S., Pandey, A., < http://www.imo.org/en/OurWork/Environment/PollutionPrevention/AirPollution/Documents/Third
%20Greenhouse%20Gas%20Study/GHG3%20Executive%20Summary%20and%20Report.pdf > (last accessed September 2016).
Jiang, L., Kronbak, J., Christensen, L.P., 2014. The costs and benefits of sulphur reduction measures: Sulphur scrubbers versus marine gas oil. Transp. Res. Part D 28
(2014), 19–27. http://dx.doi.org/10.1016/j.trd.2013.12.005.
Kristensen, H.O., 2012. Energy Demand and Exhaust Gas Emissions of Ships. Work Package 2 of Project Emissionsbeslutningsstottesystem. Technical University of
Denmark, Copenhagen.
MARPOL Annex VI NOx Emission limits. Nitrogen Oxides (NOx) – Regulation 13. < http://www.imo.org/en/OurWork/environment/pollutionprevention/
airpollution/pages/nitrogen-oxides-%28nox%29-%E2%80%93-regulation-13.aspx > (last accessed 26/02/2017).
MARPOL Annex VI Fuel sulphur limits. Sulphur oxides (SOx) – Regulation 14. < http://www.imo.org/en/OurWork/environment/pollutionprevention/airpollution/
pages/sulphur-oxides-%28sox%29-%E2%80%93-regulation-14.aspx > (last accessed 26/02/2017).
MARPOL73/78. < http://www.imo.org/en/KnowledgeCentre/ReferencesAndArchives/HistoryofMARPOL/Documents/MARPOL%2073-78%20Brief%20History%20-
%20List%20of%20amendments%20and%20how%20to%20find%20them.htm > (last accessed 26/02/2017).
Moreno-Gutiérrez, J., Calderay, F., Saborido, N., Boile, M., Rodríguez Valero, R., Durán-Grados, V., 2015. Methodologies for estimating shipping emissions and energy
consumption: a comparative analysis of current methods. Energy 86 (2015), 603–616. http://dx.doi.org/10.1016/j.energy.2015.04.083.
Panagakos, G.P., Stamatopoulou, E.V., Psaraftis, H.N., 2014. The possible designation of the Mediterranean Sea as a SECA: a case study. Transp. Res. Part D 28, 74–90.
Panasiuk, I., Turkina, L., 2015. The evaluation of investments efficiency of SOx scrubber installation. Transp. Res. Part D 40, 87–96.
Psaraftis, H.N., Kontovas, C.A., 2010. Balancing the economic and environmental performance of maritime transportation. Transp. Res. Part D: Transp. Environ. 15 (8),
458–462.
Reynolds, G., 2009. The reduction of GHG emissions from shipping – a key challenge for the industry. In: Proceedings of the World Maritime Technology Conference
WMRC, Mumbai, WMRC, 21–24 Jan. 2009.
Ronen, D., 2011. The effect of oil price on containership speed and fleet size. J. Oper. Res. Soc. 62 (1), 211–216.
Schinas, O., Bani, J., 2012. The Impact of a Possible Extension at EU Level of SECAs to the Entire European Coastline. Note to the Committee on Transport and Tourism,
Directorate-General for Internal Policies, European Parliament, Brussels.
Silva, R.A., Adelman, Z., Fry, M.M., West, J.J., 2016. The impact of individual anthropogenic emissions sectors on the global burden of human mortality due to
ambient air pollution. Environ. Health Perspect. 124 (11). http://dx.doi.org/10.1289/EHP177.
Stopford, M., 2009. Maritime Economics, third ed. Routledge, London, pp. 225.
Tormos-Martínez, B., Miró-Mezquita, G., Pérez-Gutierrez, T., Diego-Pardo, J., 2016. Low viscosity engine oils: Fuel economy and performance in real-world fleet test.
Dyna Noviembre-Diciembre 2016 vol. 91 n°6 668/674. http://dx.doi.org/10.6036/7974.
Trigeorgis, L., 1996. Real Options: Managerial Flexibility and Strategy in Resource Allocation. The MIT Press, Cambridge.
UNCTAD, 2016, United Nations Conference on Trade and Development, Review of Maritime Transport, United Nations. ISBN 978-92-1-112904-5. < http://unctad.
org/rmt > (last accessed February 2017).
Yewen, Gu, Wallace, Stein W., Wang, Xin, 2016. The Impact of Bunker Risk Management on CO2 Emissions in Maritime Transportation under ECA Regulation
(November 16, 2016). NHH Dept. of Business and Management Science Discussion Paper No. 2016/17. Available at SSRN: < https://ssrn.com/abstract=
2870407 > or < http://dx.doi.org/10.2139/ssrn.2870407.http://dx.doi.org/10.2139/ssrn.2870407 > .

250