Estimation of CO2 Emission from Marine Traffic in Singapore Straits

Using Automatic Identification Systems Data

See Chuan Leong
Institute of Systems Science, National University of Singapore, Singapore
Carol Anne Hargreaves (PhD)
Institute of Systems Science, National University of Singapore, Singapore
Prateek Singhal
Institute of Systems Science, National University of Singapore, Singapore
Jun Yuan (PhD)
Centre for Maritime Studies, National University of Singapore, Singapore

ABSTRACT: The Singapore Straits area is one of the busiest marine traffic regions in the world. Numerous
ships arrive, depart and travel through this region, and affect the air quality of the port environment. Our goal
is to estimate carbon dioxide (CO2) emission from Marine Traffic in the Singapore Straits, using Automatic
Identification Systems (AIS) data. Key questions such as, how much total CO2 emissions are emitted daily by
ships, which types of ships should be preferred and rewarded because of their low carbon emissions, what is
the likely impact of the long term carbon emission forecasts on Singapore Port will be analyzed. This research
will be of value to Singapore Port Authorities responsible for managing CO2 emissions at the Singapore Ports
as they can better understand how well they are progressing towards the International Maritime Organization
(IMO) green house gases (GHG) 2020 target and secondly, can make smart decisions on which type of ship or
country of ship may enter the Singapore port and thereby controlling CO2 emissions.

1 INTRODUCTION plans to improve the sectors carbon footprint by

stressing the importance of a global approach in fur-
1.1 Background thering improvements to energy efficiency and emis-
The world is dependent on the shipping industry for sion reduction.
the transportation of goods from one country to an- As the global economy slowly returns to growth,
other. Shipping contributes 90% of the transporta- the Port of Singapore has maintained its global lead
tion of world trades. Although ships are considered in bunker sales and has achieved consistent growth
to be the cleanest mode of transport when compared (at least 5% on average per year since 2009) in an-
to rail and air, they still contributes to more than 1 nual vessel arrival tonnage, container and cargo
gigatonnes of CO2 emissions worldwide yearly. throughput in 2013. Figuratively, in 2013, annual
(IMO 2009a) In addition, ships also emit various vessel arrival tonnage reached 2.33 billion gross
other pollutants such as CH4, N2O, NOx, SOx and tonnes (GT), container throughput hit 32.6 million
PM. In a recent report by World Meteorological Or- twenty foot equivalents (TEU), volume of bunkers
ganization, greenhouse gases have surged to its fast- sold was 42.5 million tonnes and total tonnage of
est rate for 30 years in 2013. (WMO 2014) GHG and ships under the Singapore flag was 73.6 million GT.
non-GHG have been known to cause climate warm- (MPA 2014)
ing and climate forcing effects that are detrimental to
a sustainable future.
Sustainable development is therefore a priority of 1.2 Literature Review
todays modern and technologically advanced world. In the literature review, previous studies have rarely
The international shipping industry is firmly on track involved the theoretical support of low-carbon port
to reduce its CO2 emissions by 20% by 2020, with development, more particularly the lack of quantita-
significant further reductions thereafter. However, tive research methods. For the few typical research
the next United Nations Climate Change Conference achievements, (Dong 2009) pointed out that more at-
(COP 19) needs to support the International Mari- tention to the economic development of low-carbon
time Organization (IMO) as it continues its vital ports has to be given. (Chen 2009) proposed that the
work to help deliver meaningful CO2 emission re- fifth generation port is the low-carbon port, which
ductions by international shipping, including the de- marked the development of ports in the world has
velopment of Market Based Measures. (UNFCCC entered a new phase. In July 2011, governments at
2013) The shipping industry is regulated by the In- IMO agreed on a comprehensive package of tech-
ternational Maritime Organization (IMO), which nical regulations for reducing shippings CO2 emis-
sions which entered into force in January 2013. As emission control area along the Singapore Straits
an interim measure, consideration is also being given that contains the shipping lanes and also including
by IMO to measure, monitor and report CO2 emis- the Singapore port limit area jurisdiction. Refer to
sions from individual ships, with detailed proposals Figure 1 for the defined rectangular boundary area
by governments scheduled to be discussed during with WGS84 corner coordinates [1.11N, 1.57N;
2014, including proposals by International Chamber 104.26E, 103.43E]. The ships that operate in this
of Shipping (ICS). (UNFCCC 2013) IMO introduced area will be collected from Automatic Identification
technical measures, operational measures, and eco- System (AIS) from the period January to June 2014.
nomic measures as instruments for reducing CO2. CO2 emissions are largely dependent on the number
Technical measures include improvements in the of ships, type of ships, speed of the ship, within a
Energy Efficiency Design Index (EEDI) involving specific area. Assuming that the number of ships is a
the use of modified hull form, modified propeller, predominant factor in the overall emissions, the
exhaust heat recovery and other methods. Operation- maximum number of ships in a single day will be
al measures include the Ship Energy Efficiency determined and set as the baseline for the emission
Management Plan (SEEMP) that investigates the ef- calculation and investigation. Visual displays of sta-
ficiency based on ship types and Energy Efficiency tistical data will be used for the analysis of CO2
Operational Indictor (EEOI). Economic measures emissions.
may include financial incentives to promote the use
of technical and operational measures. By 2020, new
2.2 Overview and Analysis of AIS Data
ships of various types must improve EEDI efficiency
by 10% with the reference emission set in 2013. By The AIS is an automatic tracking system used on
2025, the EEDI reduction must be 20% and by 2030, ships to identify and locate vessels by electronically
the EEDI reduction must be 30%. (MARPOL 2011) exchanging data with other nearby ships and receiver
The reduction of CO2 measures by technical base stations at regular intervals. The signal is
measures is estimated to be 28.2% by 2020 while transmitted using an onboard transponder operating
operational measures, e.g. 10% speed reduction, is in the very high frequency band. (AIS 2014) AIS
estimated to be 23.3%. (IMO 2000) Shipping com- contain information such as the unique ship identifi-
panies therefore have a very strong incentive to re- er, position, course, speed, and navigation status.
duce their fuel consumption and thus reduce their Ship identifier in AIS can be cross referenced with
CO2 emissions. Shipping ports like Singapore may other ship information from the World Shipping
even be included as a new emission control area by Register (WSR 2014) or Lloyds database (Lloyd
IMO in the future. (Jean-Florent 2013) 2014) to provide the ships physical dimension and
tonnage information. The fused information is often
presented in open source web based maps (Vessel
1.3 Purpose of Study Finder 2014) where ships details and ship move-
With the projected growth in shipping traffic and the ment are shown. For example, in Figure 1, the out-
mandatory emission limits set by IMO, there are line map depicts the traffic density of ships in vari-
many questions to be answered, namely, what is the ous shipping lanes along the Singapore Straits.
current CO2 emission level of all ships entering and (Marine Traffic 2014) What is also shown are the
leaving Singapore? Which type of ship emits more major ports in Singapore indicated as black dots, the
CO2? Singapore port limit polygon boundary and also the
The aim of this paper is to assess the operational rectangular boundary for analysis of ships and CO2
emission impact factors through the quantification of emissions.
CO2 emissions breakdown by area, ship quantity,
ship type and country of registration. With this in-
2.3 Processing of AIS and Ship Data
formation, it is expected that Singapore Port Author-
ities can use the baseline of CO2 emissions released The AIS data, received by various base receiver sta-
from ships in the Singapore Straits/Port to determine tions located in Singapore, fused with other vessel
the various technical, operational and economic information, is obtained from the Maritime and Port
measures so as to meet the required emission limits Authority of Singapore (MPA). Ships that are locat-
set by IMO. ed within the rectangular bounding box area will be
analyzed. Erroneous vessel locations in the AIS data
may be due to equipment malfunction or invalid op-
2 METHODOLOGY erator inputs. For example, locations that are inland
are removed. The inland land boundaries are ob-
2.1 Framework and Data Collection tained from an OpenStreetMap. (OpenStreetMap
CO2 emissions depend on the number and type 2014) Vessel movement speeds, derived from delta
ships operating in the region. We established a rec- locations in consecutive time periods, are analyzed
tangular boundary area that serves as a possible to determine if they are significantly higher than the
vessels maximum cruising speeds. For such move- Table 1. Consumption of full power fuel, Cjk, (tonnes/day) ver-
ments where the speed is more than 5 times its max- sus Gross Tonnage (GT).
imum speed, the location information is removed. Ship Type Consumption (tonnes/day)
Solid bulk Cjk = 20.186 + .00049 GT
Liquid bulk Cjk = 14.685 + .00079 GT
Gen.Cargo Cjk = 9.8197 + .00143 GT
1.57 N, 104.26 E Container Cjk = 8.0552 + .00235 GT
Passenger/RoRo/Cargo Cjk = 12.834 + .00156 GT
Passenger Cjk = 16.904 + .00198 GT
High speed ferry Cjk = 39.483 + .00972 GT
Inland cargo Cjk = 9.8197 + .00143 GT
Sail ships Cjk = 0.4268 + .00100 GT
Tugs Cjk = 5.6511 + .01048 GT
Fishing Cjk = 1.9387 + .00448 GT
Singapore Port Limit
Area of Study
(Singapore Straits)
Other Ships Cjk = 9.7126 + .00091 GT
1.11 N, 103.43 E

Other theoretical formulations using engine type

Figure 1. Outline map of ship traffic density at Singapore and engine power exist (IMO 2009b) but these for-
Straits. mulae require the ship and the operational infor-
mation that is currently not available for this study.
In our analysis, there will be modifications and adap-
2.4 Emission Calculation tations to equation (1). The time interval between
consecutive AIS reports is typically 5 minutes, there-
A complete methodology, using the European fore, Cjk is modified as a fraction of the time interval
(MEET) framework, for the estimate of air pollutant per day, i.e. 5/(6024)Cjk. Gross tonnage is defined
emissions from ships in port environment was estab- in accordance to the 1969 International Tonnage
lished by (Trozzi 1999). Trozzi used a database of Convention. (IMO 1969) The CO2 emissions factor,
15,000 ships with daily engine fuel consumption, C, Fjlm, defined in equation (1) is assumed to be inde-
in tonnes/day, to construct a linear fit fuel usage pendent of the fuel type, ship type, and engine type
based on the ship type, k and fuel type, j at full pow- but dependent on the operational mode. Fjl is taken
er. There is no indication whether the ships fuel to be 3173kg/tonne of fuel used based on a maxi-
consumption is dependent on the ship cargo load, mum operating mode during cruising. (Oonk, H.
whether the ship is sailing upstream or downstream 2003) The emission factor as a function of the oper-
to the water current or whether is in sailing in differ- ating mode is as shown in Table 2. The operating
ent weather conditions. The results are as shown in mode is classified based on the ship maximum
Table 1 with fuel type, j, assumed to be marine die- speed. (Trika Pitana 2010) The ships maximum
sel oil. The total emission, Ejklm, in kg/interval t, is as speed is derived based on the AIS information ex-
shown in Equation (1) below (Trozzi 1999) :- tracted during the entire duration of 6 months. Using
the AIS information, the maximum speed is set at 3
Ejklm|t = Cjk|t Fjlm (1) standard deviations from the mean speed. Speeds
above the maximum speed are identified as outliers
where, and are removed. The fraction of maximum fuel
j is the fuel type (Bunker fuel oil, Marine diesel oil, used is based on the assumption that the full power
Marine gas oil, Gasoline fuel) fuel used corresponds to the maximum speed of the
k is the ship type for use in consumption classifica- ship. The reduction of fuel used will be based on a
tion (Solid bulk, Liquid bulk, General cargo, Con- cubic fitted relationship using estimates of a 10,000
tainer, Passenger/Ro-Ro/Cargo, Passenger, High TEU ship size operating at 25 knots for maximum
speed ferries, Inland cargo, Sail ships, Tugs, Fishing, speed. (Notteboom T. 2009) For example, a ship
Other ships) having a speed of 0.6 of its maximum speed is de-
l is the engine type class for use in emission factors fined to be operating in the maneuvering mode with
characterization the emission factor calculated as 0.483173=1523kg
m is the operation mode (Cruising, Maneuvering, per tonne of daily fuel used. No fuel is used when
Hotelling, Tanker offloading) the speed is at 0 km/hr. The total emissions per day,
Fjlm is the average emission factors of CO2 in Ejklm, can be obtained by summing all the time inter-
kg/tonne/day vals that have occurred for each ship throughout the
t is the time interval between a AIS report pair day period within the boundary area.
Table 2 Operating mode and emission factor used.
Operating Speed Classification Fraction of Maximum
Mode Fuel Used
Cruising (0.8-1] of max speed 1.0
Maneuvering (0.2-0.8] of max speed 0.48
Hotelling (0-0.2] of max speed 0.03

3 RESULTS AND DISCUSSION

3.1 Traffic Density
Figure 2. Average numbers of ships per hour during each
Figure 1, shows the traffic density at the Singapore
month of 2014.
Straits, indicating that there are a lot more ships that
are berthed within the Singapore port limit as com-
pared to those that are sailing along the shipping
lanes. Figure 2, shows the traffic density data col-
lected between 01st January 2014 and 30th June
2014. The hourly number of ships is aggregated to a
day and a month. For example there were about 1660
ships on average per hour in the month of May 2014
while the densest traffic was recorded during Febru-
ary 2014 which has about 1730 ships on average per
hour within the bounding area. The average number
of unique ships in an hour, a day and a month from Figure 3. Average number of ships per hour during each day in
January to Jun is 1708, 2191, 8009 respectively. February 2014.
Considering February 2014 as the period with the
highest traffic, Figure 3 shows that February 12,
2014 was the peak day for marine traffic density. In
contrast, the lowest traffic density was recorded on
February 27, 2014 with 1665 ships. Taking February
12, 2014 as the densest day, the number of ships is
further broken down by each hour as shown in Fig-
ure 4. On the selected date during the time interval
1300H to 1400H, the highest traffic of 1839 ships
was recorded. It is also shown in Figure 1 that the
ship traffic consist of various ship types, so it is nec-
essary to include ship types into the analysis as it
will affect the emission factors. Figure 5 shows that
the majority proportion of ships recorded on Febru- Figure 4. Average number of ships during each hour on Febru-
ary 12, 2014 were Liquid bulk, Tugs, General Cargo, ary 12, 2014.
Container followed by Passenger/Ro-Ro/Cargo and
Solid bulk ship types. About 83% of Liquid bulk
ship types are contributed by tankers. The ship types,
obtained from MPA, are remapped to the Acurex
classification. (EPA 2000) The distribution of ships
registered country of origin is as shown in Figure 6.
Majority of the ships come from Singapore (40.4%),
mixed (23.7%), Indonesia (8.3%) and Panama
(8.1%). Figure 7 shows a box-plot of gross tonnage
(1000) categorized by the different ship types. Con-
tainer and Liquid bulk ship types have large variance
in gross tonnage. Solid bulk ship type has the highest
median gross tonnage across all ship types.

Figure 5. Percentage of ship type distribution on February 12,

2014.
 baseline, the ratio of the number of ships for Con-
Marshall Islands
4.7%
tainer type is obtained from Figure 5. Their relative
GT ratio is obtained from the median line in Figure
7. The relative fuel consumption is taken from the
Liberia GT coefficient in Table 1. The overall relative emis-
4.7%
Malaysia
sion is obtained from Table 3. Relative speed is de-
5.3% rived from the rest of the factors. The low relative
Singapore speed (0.13) for Container ship type indicates that
Panama 40.4% most of them are stationary, possibly involved in
8.1%
loading/unloading activities. GT for Container has a
Indonesia
larger variance and median value as compared to
8.3% Liquid bulk ship type. Multiplied with the relative
fuel consumption and speed, the variance is at 1.74,
Others
23.7%
and this explains the large variance. However, be-
cause the proportion of Container ships is only 0.3
times of Liquid bulk, the overall emission is lower.
Table 3. Estimates for CO2 emission in the Singapore Straits on
Figure 6. Percentage of ship registered country distribution on February 12, 2014 categorized by ship type.
February 12, 2014. Ship Type CO2 emissions %age of CO2 emissions
(tonnes/day) emissions /GT (kg/day)
Liquid bulk 5,814 40.3% 0.36
Container 4,340 30.1% 0.60
Solid bulk 1,155 8.0% 0.27
Tugs 993 6.9% 7.68
General cargo 816 5.7% 0.88
Passenger 387 2.7% 1.05
Passenger/Ro- 582 4.0% 2.19
Ro/Cargo
Others 268 1.9% 0.43
Inland cargo 72 0.5% 3.33
Sail ship 4 0.0% 0.44
Fishing 2 0.0% 1.06
Total 14,432 100.00%

Figure 7. Boxplot of GT categorized by ship type on February Table 4. Comparison of relative factors contributing to overall
12, 2014. emission for Container and Liquid bulk ship types.
Ship Type Rel. Rel. Rel. fuel Rel. Rel.
no. of GT consump- Speed Emission
3.2 CO2 Emissions
ships tion (Derived)
With the highest ship traffic recorded on February Liquid 1 1 1 1 1
12, 2014, the estimated CO2 emission within the bulk
bounding area was therefore determined. Out of the Container 0.3 6.5 2.97 0.13 0.75
1725 average number of ships recorded on that day,
the proportion of liquid bulk and container ships
amount to 46% of the total ships. However, in terms
of CO2 emission, they contributed a higher percent-
age, 70.4%, of the total CO2 emissions. Refer to Ta-
ble 3 for the emissions categorized by ship type. The
CO2 emissions per GT in each ship type category is
also computed. Tugs emit 28.4 times more CO2 as
compared to Solid bulk per GT volume.
Figure 8 shows the box plot distribution of CO2
emissions categorized by ship type. The Container
ship type has higher emissions variance across the
ships as compared to Liquid bulk ship type but the
overall contribution to CO2 emission is lower. This
can be explained by comparing the relative factors
that differentiates the overall emission as shown in Figure 8. Boxplot of CO2 emissions on February 12, 2014 cate-
Table 4. Setting Liquid bulk factors with 1 as the gorized by ship type.
 The CO2 emissions are also analyzed against the
ship registered country of origin. A majority of the
emissions, 28.6%, are contributed by local Singapore
registered ships although the proportion of Singa-
pore ships is much higher at 40.4%. An index of 1.0
for Singapore is created based on the ratio of per-
centage emissions and proportion of ships registered
by the country. Bahamas and Greece registered ships
emit 4.2 and 3.8 times more CO2 per ship respective-
ly as compared to Singapore registered ships. The
proportion of Singapore and Panama ships in each
ship type for the top 6 total emission ship type is as
shown in Table 6. The larger percentage of Singa-
pore ships for liquid bulk, tugs as compared to Pan-
ama explains the larger proportion of total emis- Figure 9. Boxplot of CO2 emissions on February 12, 2014, cat-
sions. Figure 9 shows the box plot distribution of egorized by country type.
CO2 emissions categorized by country of origin.
There is significant difference (p < 0.001) for CO2
emissions across the country types based on a one 3.3 Emissions Reduction
way ANOVA test. The significant emission contrib-
In this section, 3 factors, namely, 1) reduction of
utors, Singapore and Panama, seem to have a smaller Liquid and Container ship types, 2) slow stemming
variance in CO2 emission as compared to the other
of Liquid and Container ship types, 3) use of larger
countries. The exception to this observation is Indo-
GT Liquid and Container ship types, are investigated
nesia and Malaysia country types.
for their contribution to total emissions reduction.
These ship types are targeted because they contribut-
Table 5. Estimates of CO2 emission in Singapore Straits on
ed the most to the CO2 emissions.
February 12, 2014, categorized by country type.
As described in the previous section, the highest
Ship CO2 emis- Proportion %age of Rel
CO2 emissions (70.4%) come from the large propor-
Country sions of ships in emissions Index
tion of Liquid bulk and Container ship types. A 10%
(tonnes/day) %age
reduction in the total number of Liquid bulk ships
Singapore 4,123 40.4% 28.6% 1.0
will reduce total emissions by 4%. Comparatively, a
Others 3,302 23.7% 22.9% 1.4
10% reduction in Container ships will only reduce
Panama 1594 8.1% 11.0% 1.9
the total emissions by 3%.
Liberia 1391 4.7% 9.6% 2.9
By adopting slow steaming and extra slow steam-
Marshall 1172 4.0% 8.1% 2.9
ing operations, it will reduce the amount of emis-
Islands
sions. If Liquid bulk ships reduce their speed by
Hong 592 2.3% 4.1% 2.5
10%, carbon emissions will be reduced by 23.5%.
Kong
(IMO 2000), resulting in the overall emissions being
Malaysia 542 5.3% 3.8% 1.0
reduced by 9.5% using the figures (76.5% 5,814 =
Bahamas 522 1.2% 3.6% 4.2
4,448 tonnes/day) from Table 3. However, if we re-
Indonesia 514 8.3% 3.6% 0.6
simulate all the new Liquid bulk ship type speeds de-
Greece 343 0.9% 2.4% 3.8
rived from the AIS information, there will be more
Malta 337 1.1% 2.3% 3.0
ships crossing into a lower operating mode. (Table
Total 14,432 100% 100% 2) The total emission is reduced by 3.55%. For the
Container ship type, the total emission reduction is
Table 6. Proportion of Singapore and Panama ships within each 2.75% based on simulation of new speed reduction
ship type for the top 6 total emission ship types. of 10%. More intermediate operating modes and
Ship Type sort- Proportion of Sin- Proportion of Pan- ship speeds categories will increase the fuel reduc-
ed by top 6 total gapore ships ama ships tion and emissions.
emission From Table 1, due to the constant in the linear
Liquid bulk 43.0% 30.5% function, if the GT of each ship is increased and the
Container 5.2% 15.9% number of ships is then decreased, the overall fuel
Solid bulk 1.1% 17.2% consumption will reduce and therefore emissions
Tugs 42.9% 6.6% will be decreased. If each Liquid bulk ship GT is in-
General cargo 7.8% 14.6% creased by 10% (using average GT of a Liquid bulk
Passenger 0.5% 1.3% ship of 24,288), the number of ships required is de-
creased by 9.1% to maintain the overall GT of all
Liquid bulk ships. This will result in a 3.9% reduc-
tion in emission for Liquid bulk ships alone or Dong, Wu. 2009. Attach More Importance to the Economic
1.73% in overall emission reduction. Similarly for Development of Low-Carbon Ports. Port economy.
the Container ship type, the reduction is 0.73% in EPA. 2000. Analysis of Commercial Marine Vessels Emissions
emission for Container ship type or 0.17% of the and Fuel Consumption Data.
overall emission reduction. IMO. 1969. International Convention on Tonnage Measure-
In an emission reduction programme, the Singa- ment of Ships.
pore and Panama registered ships may be targeted IMO. 2000. Study of Greenhouse Gas Emissions from Ships.
for emissions reduction first, as they contribute al- IMO. 2009a. Second IMO GHG Study.
most 39.6% of all emissions. IMO. 2009b. Interim Guidelines on the Method of Calculation
Assuming the number of Liquid bulk and Con- of the Energy Efficiency Design Index for New Ships. Circu-
tainer ships are expected to grow 2% respectively lar MEPC.1/Circ. 681.
yearly, a combination of 10% reduction in speed and Jean-Florent Helfre. 2013. Emission Reduction in the Shipping
a 10% increase in GT will result in a 9.1% reduction Industry: Regulations, Exposure and Solutions. Sustainalyt-
number of ships for Liquid bulk and Container ship ics.
types and maintain the emissions for a period of Lloyds List Intelligence. 2014.
about 2 years. http://www.lloydslistintelligence.com/llint.
Marine Traffic. 2014. https://www.marinetraffic.com.
MARPOL Annex VI, Chapter 4. 2011.
4 CONCLUSION MPA. 2014. Singapores 2013 Maritime Performance Port of
Singapore Registers Good Growth in 2013.
The CO2 emission levels for the Singapore Straits Notteboom, T.. 2009. Fuel surcharge practices of container
area have been analyzed from January to June 2014 shipping lines: Is it about cost recovery or revenue making?
and the highest marine traffic occurred on February Proceedings of the 2009 International Association of Mari-
12, 2014. An average of 14432 tonnes of CO2 is time Economists (IAME) Conference.
emitted each day using ship gross tonnage and ship Oonk, H., Hulskotte. 2003. Emission Factors of Seagoing Ships
speed information derived from Automatic Identifi- on the Purpose of Yearly Emission Calculation. TNO-report
cation System data. The highest contributor (33%) of R 2003/438 version 2.
CO2 emissions are the tankers that is classified under OpenStreetMap. 2014. http://www.openstreetmap.org.
Liquid bulk type. Tugs also emit 28 times more CO2 Trika Pitana. 2010. Estimation of Exhaust Emissions of Marine
as compared to Solid bulk ship types per gross ton- Traffic Using Automatic Identification System Data (Case
nage volume. Singapore registered ships also con- Study: Madura Strait Area, Indonesia). OCEANS IEEE.
tributed 28.6% of CO2 emissions. Several operation- Trozzi, C. et al.. 1999, Actual and Future Air Pollutant Emis-
al factors such as the reduction in the number of sion from Ships. Proceeding of INRETS Conference, Aus
ships, a reduction in operation speed, and an increase tria.
in gross tonnage carrying capacity have been ana- UNFCCC. 2013. International Maritime Organization, Ship-
lyzed to determine the impact on the overall emis- ping World Trade and the Reduction of CO2. ICS.
sion reduction. A 10% reduction in the operational Vessel Finder. 2014. http://www.vesselfinder.com.
speed and a 10% increased in gross tonnage of all WMO. 2014. Record Greenhouse Gas Levels Impact Atmos-
Liquid bulk and Container ships can help to maintain phere and Oceans, Press Release No. 1002
the 2013 level of CO2 emissions for about 2 years World Shipping Register (WSR). 2014. http://www.world-
based on a 2% yearly projected growth rate of these ships.com.
ships. Only together with technical measures Singa-
pore Port will be able to meet the 2020 IMO CO2
emission limits. Simple analysis such as the average
number of ships per day, or per month and, infor-
mation such as the ship type, ship growth, will pro-
vide port authorities key information that they may
use when estimating, projecting and controlling the
carbon emissions along the Singapore straits.

5 REFERENCES

AIS - Automatic Identification System. 2014.

http://en.wikipedia.org/wiki/Automatic_Identification_Syste
m.
Chen, Yan. 2009. On the Fifth-generation Port. China Collec-
tive Economy.