Marine Pollution Bulletin 87 (2014) 22–28
Contents lists available at ScienceDirect
Marine Pollution Bulletin journal homepage: www.elsevier.com/locate/marpolbul
Focus
Environmental accounting for Arctic shipping – A framework building on ship tracking data from satellites A. Mjelde ⇑, K. Martinsen, M. Eide, Ø. Endresen DNV GL, Veritasveien 1, Høvik, Norway
a r t i c l e
i n f o
Article history: Available online 29 August 2014 Keywords: Arctic shipping AIS (Automatic Identification System) Environmental accounting Black carbon emissions Heavy fuel oil
a b s t r a c t Arctic shipping is on the rise, leading to increased concern over the potential environmental impacts. To better understand the magnitude of influence to the Arctic environment, detailed modelling of emissions and environmental risks are essential. This paper describes a framework for environmental accounting. A cornerstone in the framework is the use of Automatic Identification System (AIS) ship tracking data from satellites. When merged with ship registers and other data sources, it enables unprecedented accuracy in modelling and geographical allocation of emissions and discharges. This paper presents results using two of the models in the framework; emissions of black carbon (BC) in the Arctic, which is of particular concern for climate change, and; bunker fuels and wet bulk carriage in the Arctic, of particular concern for oil spill to the environment. Using the framework, a detailed footprint from Arctic shipping with regards to operational emissions and potential discharges is established. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction The melting of Arctic sea-ice will effectively unlock greater parts of the Arctic Ocean, leaving it increasingly attractive to shipping and oil and gas related activities (e.g. Stephenson et al., 2011, 2013). A significant increase in the maritime activity in the Arctic waters is expected (e.g. AMSA, 2009; Paxian et al., 2010; Peters et al., 2011), leading to increased concern over the potential environmental impacts from this activity (e.g. Corbett et al., 2010; Browse et al., 2013; Dalsøren et al., 2007, 2013; Peters et al., 2011). Accurate, up-to-date inventories of emissions and discharges from ships are essential for scientists quantifying the environmental impact of the shipping activities and for policy-makers to find cost effective regulative measures. Modelling studies have produced Arctic emission inventories, based on fleet and movement data (e.g. from AMSA in Corbett et al., 2010), as well as on ship observations in Arctic (e.g. Peters et al., 2011; Dalsøren et al., 2007). These data includes a range of assumptions and limitations (Endresen et al., 2003; Dalsøren et al., 2007; Corbett et al., 2010). The Automatic Identification System (AIS) ship tracking data from satellites opens new possibilities for expedient and accurate environmental accounting for Arctic shipping. An AIS transponder, which is mandatory on board almost all ships above 300 gross tons, automatically transmits a unique ship identity code, a precise ⇑ Corresponding author. E-mail address: [email protected] (A. Mjelde). http://dx.doi.org/10.1016/j.marpolbul.2014.07.013 0025-326X/Ó 2014 Elsevier Ltd. All rights reserved.
position reference, and the ship’s heading and speed down to seconds intervals. AIS data coupled with ship registers and other databases, has already produced valuable insights into the operational risks (Eide et al., 2006, 2007, 2008; Brude et al., 2012), and the technical and environmental performance of ships (e.g. Endresen et al., 2007; Buhaug et al., 2009; Pitana et al., 2010; Jalkanen et al., 2009; Smith et al., 2013). Winther et al. (2014) were the first to publish in a journal an emission inventory for ships in the Arctic area based on AIS data. In addition, studies carried out on behalf of PAME (The Protection of the Arctic Marine Environment Working Group under the Arctic Council) and Norwegian authorities, presented in 2011 an inventory of ships carrying heavy fuel oil (HFO) in the Arctic (DNV, 2012). This paper describes a framework for environmental accounting for Arctic shipping based on AIS ship tracking data from satellites. Results from two cases are presented; Emissions of black carbon (BC) in the Arctic, and; Bunker fuels and wet bulk carriage in the Arctic.
2. The framework An AIS-based modelling approach has been developed by DNV GL, which models the ship activities, emissions and potential discharges, and allocates the results geographically. The framework (outlined in Fig. 1) combines AIS data and ship particulars from ship registers, to form an activity basis for the Arctic fleet. The AIS-based modelling approach builds on the individual satellite
A. Mjelde et al. / Marine Pollution Bulletin 87 (2014) 22–28
23
Fig. 1. Framework for AIS-based environmental accounting and impact modelling.
tracing of ships (each position record). In principle, any relevant quantity, Q, in a given area and time period, can be calculated in the model by applying the following formula;
Q¼
XX qi;x i
ð1Þ
x
where qi,x denotes contribution from an individual ship i, in the time and space interval associated with an individual AIS position message x. The quantification of qi,x is made by combining individual AIS satellite positions, vessel particulars and supporting data, see Fig. 1. The calculated inventories based on Eq. (1) can be aggregated in groups based on ship characteristics, typically 13 ship types and 7 size segments. Results covering air emissions, potentially discharges to sea and voyage based risk are presented. The inventory is automatically updated, using a continuous download of AIS data, to track the development over time. Statistics are presented on different time scales (e.g. hourly, daily, monthly, yearly), and are integrated in an environmental accounting system, which is operated by the Norwegian Coastal Administration (NCA). Spatial distribution is inherent in the AIS data, and can be aggregated at any relevant resolution, e.g. 10 10 km grids. Having created an inventory, the concentration and impacts of the pollutant in question can be assessed using relevant atmospheric models or risk models (Fig. 1). Within the framework,
databases of mitigation options can be linked to the inventory to assess the costs and effects of abatement options in ‘‘what-if’’ analysis. As an example, the effect of using alternative marine fuel such as Liquid Natural Gas (LNG) can be assessed, capturing the benefit from oil spill reductions, as well as the reduction in emissions such as CO2, NOX, SOX and BC. The AIS based modelling approach is by no means restricted to the Arctic, but may equally well be applied globally or to any specific region or area of interest. The framework has been used to model emissions to air, navigation risk and potential discharges to sea (grey water, black water, oils, etc.) and can be extended to other ship activity impacts including the risks related to invasive species from ballast water discharges. The high resolution of the inventories, combined with the ease of updating with new inputs (e.g. emission factors and emission reduction potentials from appropriate measures like Selective Catalytic Reduction (SCR), exhaust gas scrubbers, introduction of LNG/LPG, etc.), makes the framework ideally suited to providing policymakers and regulators, ship-owners and other stakeholders with the best possible decision basis.
3. Black carbon emissions in the Arctic Emissions to air from shipping are of global concern. In the Arctic, particular focus has been on emissions of black carbon (BC) (e.g. Corbett et al., 2010; DNV, 2012; Winther et al., 2014).
24
A. Mjelde et al. / Marine Pollution Bulletin 87 (2014) 22–28
Table 1 The share of fuel consumption and BC emissions in the Arctic region in 2012. Ship segment
Ship type
Distillate (%)
HFO (%)
BC emissions (%)
Wet bulk
Oil tanker Chemical/Prod tanker Gas tanker
2.0 1.8 0.0
13.5 7.7 0.8
7.1 4.4 0.4
Dry bulk
Bulk carrier General cargo Container vessel RoRo Reefer Passenger
0.0 6.6 0.3 0.1 0.8 4.4
9.6 5.8 26.5 0.4 2.7 10.5
4.3 6.3 12.0 0.2 1.6 7.1
Offshore supply vessel Other offshore service vessel Other activities Fishing vessel
4.5 0.6 29.9 48.9 100.0
4.2 0.0 12.2 6.1 100.0
4.4 0.3 22.0 29.7 100.0
Passenger Other activities
Fishing Totals
BC contributes to reduce air quality as well as influencing climate change. In the Arctic, atmospheric BC absorb radiation both from incident sunlight and sunlight reflected from snow and ice. In addition, BC deposited on snow or ice reduce surface reflectivity (i.e. albedo), thus accelerating the ice melting process (Flanner et al., 2007; Hansen and Nazarenko, 2004). The BC emissions in the Arctic have been modelled utilising the AIS-based modelling framework. The model is similar to the activity based model described by Dalsøren et al. (2009), but with engine load factors and time at sea derived for each ship position record from the AIS source. For a given geographical area and time span, the total exhaust gas emission for a given emission component, E (CO2, SOx, NOx, etc.), is calculated by:
E¼
X Me ðM i þ MAUX Þ i
ð2Þ
i
where MMe,AUX denotes the amount of pollutant emitted by ship i for i main (Me) and auxiliary (AUX) engines respectively (kg pollutant).
MMe ¼ i
X x
MMe i;x ¼
X
Me cMe i F i;x
ð3Þ
x
where MMe i,x denotes the amount of pollutant emitted by ship i for main engines associated with AIS position message x (kg pollutant). cMe denotes the main engine emission factor for the given poli lution type for the given ship i (kg pollution/kg fuel). FMe i,x denotes the main engine fuel consumption for ship i associated with AIS position message x (kg fuel).
Fig. 2. Spatial distribution of BC emissions for Arctic region 2012.
25
A. Mjelde et al. / Marine Pollution Bulletin 87 (2014) 22–28
and
LF Me i;x ¼
Me
Ddx =Dtx SSi
3 ð5Þ
where Ddx is the great circle distance between AIS position message x and message x 1 (nm). SSi denotes the service speed of the individual ship i (knots). MAUX is calculated in a similar manner with auxiliary engine i emission factors, SFOC, and load factor based on ship operation mode (Open sea and Harbor mode), as determined based on ship speed over a period of time.
Fig. 3. Monthly break-down of BC emissions for Arctic region 2012 (tons).
Me Me F Me i;x ¼ pi bi LF i;x Dt x
ð4Þ
pMe denotes installed main engine power installed for ship i i (kW). bi Me denotes the specific fuel consumption (SFOC) for main engines for ship i (kg fuel/kW h). LFMe i,x denotes the main engine load factor of the individual ship i in the time interval associated with AIS position message x (–). Dtx is the time between AIS position message x and message x 1 (h).
The results are calculated (Eq. (2)) and presented for the Arctic region, as defined in the proposed International Maritime Organization (IMO) guideline for ships operating in ice-covered waters (IMO, 2010). AIS data from the Norwegian satellite AISsat-1 has been made available by the Norwegian Coastal Administration (NCA). The ship particulars are collected from DNV GL vessels and IHS Fairplay World Register of Ships (IHS Fairplay, 2012). The BC emission factor of 0.35 kg/ton fuel is from the IMO GHG study (Buhaug et al., 2009). AIS data for 2012 is used in the emission modelling. A total of 1347 unique vessels were found to operate in the Arctic region throughout 2012, having collectively sailed distance of
Table 2 Black carbon emissions from shipping in the Arctic, using three different Arctic definitions.
This study Winther et al. (2014) Corbett et al. (2010) Peters et al. (2011)
Reference year
Primary activity data source
Arctic region* (tons)
AMAP area** (tons)
2012 2012 2004 2004
AIS AIS AMSA shipping database COADS/AMVER
105 – – –
1165 – 1230 1150
***
Above 58,95° north (tons) 1330 1584 – –
*
According to the definition given by IMO (2010). According to the definition given by AMAP (1998). An approximation to the actual AMAP area is used. The area used includes parts of the Norwegian and Baltic coastline between 60° and 62° which are not part of the AMAP area, and excludes parts of Canadian waters and the North Sea below 60° which are a part of the AMAP area. A best judgment estimate of the discrepancy introduced by this approximation is 10%. **
***
Table 3 Number of unique vessels on the Arctic region in 2012. Numbers of vessels using HFO are given in brackets. Colour coding indicates the dominating type of fuel within each segment. Full red and green means 100% and 0% HFO, respectively.
Ship type
Contents lists available at ScienceDirect
Marine Pollution Bulletin journal homepage: www.elsevier.com/locate/marpolbul
Focus
Environmental accounting for Arctic shipping – A framework building on ship tracking data from satellites A. Mjelde ⇑, K. Martinsen, M. Eide, Ø. Endresen DNV GL, Veritasveien 1, Høvik, Norway
a r t i c l e
i n f o
Article history: Available online 29 August 2014 Keywords: Arctic shipping AIS (Automatic Identification System) Environmental accounting Black carbon emissions Heavy fuel oil
a b s t r a c t Arctic shipping is on the rise, leading to increased concern over the potential environmental impacts. To better understand the magnitude of influence to the Arctic environment, detailed modelling of emissions and environmental risks are essential. This paper describes a framework for environmental accounting. A cornerstone in the framework is the use of Automatic Identification System (AIS) ship tracking data from satellites. When merged with ship registers and other data sources, it enables unprecedented accuracy in modelling and geographical allocation of emissions and discharges. This paper presents results using two of the models in the framework; emissions of black carbon (BC) in the Arctic, which is of particular concern for climate change, and; bunker fuels and wet bulk carriage in the Arctic, of particular concern for oil spill to the environment. Using the framework, a detailed footprint from Arctic shipping with regards to operational emissions and potential discharges is established. Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction The melting of Arctic sea-ice will effectively unlock greater parts of the Arctic Ocean, leaving it increasingly attractive to shipping and oil and gas related activities (e.g. Stephenson et al., 2011, 2013). A significant increase in the maritime activity in the Arctic waters is expected (e.g. AMSA, 2009; Paxian et al., 2010; Peters et al., 2011), leading to increased concern over the potential environmental impacts from this activity (e.g. Corbett et al., 2010; Browse et al., 2013; Dalsøren et al., 2007, 2013; Peters et al., 2011). Accurate, up-to-date inventories of emissions and discharges from ships are essential for scientists quantifying the environmental impact of the shipping activities and for policy-makers to find cost effective regulative measures. Modelling studies have produced Arctic emission inventories, based on fleet and movement data (e.g. from AMSA in Corbett et al., 2010), as well as on ship observations in Arctic (e.g. Peters et al., 2011; Dalsøren et al., 2007). These data includes a range of assumptions and limitations (Endresen et al., 2003; Dalsøren et al., 2007; Corbett et al., 2010). The Automatic Identification System (AIS) ship tracking data from satellites opens new possibilities for expedient and accurate environmental accounting for Arctic shipping. An AIS transponder, which is mandatory on board almost all ships above 300 gross tons, automatically transmits a unique ship identity code, a precise ⇑ Corresponding author. E-mail address: [email protected] (A. Mjelde). http://dx.doi.org/10.1016/j.marpolbul.2014.07.013 0025-326X/Ó 2014 Elsevier Ltd. All rights reserved.
position reference, and the ship’s heading and speed down to seconds intervals. AIS data coupled with ship registers and other databases, has already produced valuable insights into the operational risks (Eide et al., 2006, 2007, 2008; Brude et al., 2012), and the technical and environmental performance of ships (e.g. Endresen et al., 2007; Buhaug et al., 2009; Pitana et al., 2010; Jalkanen et al., 2009; Smith et al., 2013). Winther et al. (2014) were the first to publish in a journal an emission inventory for ships in the Arctic area based on AIS data. In addition, studies carried out on behalf of PAME (The Protection of the Arctic Marine Environment Working Group under the Arctic Council) and Norwegian authorities, presented in 2011 an inventory of ships carrying heavy fuel oil (HFO) in the Arctic (DNV, 2012). This paper describes a framework for environmental accounting for Arctic shipping based on AIS ship tracking data from satellites. Results from two cases are presented; Emissions of black carbon (BC) in the Arctic, and; Bunker fuels and wet bulk carriage in the Arctic.
2. The framework An AIS-based modelling approach has been developed by DNV GL, which models the ship activities, emissions and potential discharges, and allocates the results geographically. The framework (outlined in Fig. 1) combines AIS data and ship particulars from ship registers, to form an activity basis for the Arctic fleet. The AIS-based modelling approach builds on the individual satellite
A. Mjelde et al. / Marine Pollution Bulletin 87 (2014) 22–28
23
Fig. 1. Framework for AIS-based environmental accounting and impact modelling.
tracing of ships (each position record). In principle, any relevant quantity, Q, in a given area and time period, can be calculated in the model by applying the following formula;
Q¼
XX qi;x i
ð1Þ
x
where qi,x denotes contribution from an individual ship i, in the time and space interval associated with an individual AIS position message x. The quantification of qi,x is made by combining individual AIS satellite positions, vessel particulars and supporting data, see Fig. 1. The calculated inventories based on Eq. (1) can be aggregated in groups based on ship characteristics, typically 13 ship types and 7 size segments. Results covering air emissions, potentially discharges to sea and voyage based risk are presented. The inventory is automatically updated, using a continuous download of AIS data, to track the development over time. Statistics are presented on different time scales (e.g. hourly, daily, monthly, yearly), and are integrated in an environmental accounting system, which is operated by the Norwegian Coastal Administration (NCA). Spatial distribution is inherent in the AIS data, and can be aggregated at any relevant resolution, e.g. 10 10 km grids. Having created an inventory, the concentration and impacts of the pollutant in question can be assessed using relevant atmospheric models or risk models (Fig. 1). Within the framework,
databases of mitigation options can be linked to the inventory to assess the costs and effects of abatement options in ‘‘what-if’’ analysis. As an example, the effect of using alternative marine fuel such as Liquid Natural Gas (LNG) can be assessed, capturing the benefit from oil spill reductions, as well as the reduction in emissions such as CO2, NOX, SOX and BC. The AIS based modelling approach is by no means restricted to the Arctic, but may equally well be applied globally or to any specific region or area of interest. The framework has been used to model emissions to air, navigation risk and potential discharges to sea (grey water, black water, oils, etc.) and can be extended to other ship activity impacts including the risks related to invasive species from ballast water discharges. The high resolution of the inventories, combined with the ease of updating with new inputs (e.g. emission factors and emission reduction potentials from appropriate measures like Selective Catalytic Reduction (SCR), exhaust gas scrubbers, introduction of LNG/LPG, etc.), makes the framework ideally suited to providing policymakers and regulators, ship-owners and other stakeholders with the best possible decision basis.
3. Black carbon emissions in the Arctic Emissions to air from shipping are of global concern. In the Arctic, particular focus has been on emissions of black carbon (BC) (e.g. Corbett et al., 2010; DNV, 2012; Winther et al., 2014).
24
A. Mjelde et al. / Marine Pollution Bulletin 87 (2014) 22–28
Table 1 The share of fuel consumption and BC emissions in the Arctic region in 2012. Ship segment
Ship type
Distillate (%)
HFO (%)
BC emissions (%)
Wet bulk
Oil tanker Chemical/Prod tanker Gas tanker
2.0 1.8 0.0
13.5 7.7 0.8
7.1 4.4 0.4
Dry bulk
Bulk carrier General cargo Container vessel RoRo Reefer Passenger
0.0 6.6 0.3 0.1 0.8 4.4
9.6 5.8 26.5 0.4 2.7 10.5
4.3 6.3 12.0 0.2 1.6 7.1
Offshore supply vessel Other offshore service vessel Other activities Fishing vessel
4.5 0.6 29.9 48.9 100.0
4.2 0.0 12.2 6.1 100.0
4.4 0.3 22.0 29.7 100.0
Passenger Other activities
Fishing Totals
BC contributes to reduce air quality as well as influencing climate change. In the Arctic, atmospheric BC absorb radiation both from incident sunlight and sunlight reflected from snow and ice. In addition, BC deposited on snow or ice reduce surface reflectivity (i.e. albedo), thus accelerating the ice melting process (Flanner et al., 2007; Hansen and Nazarenko, 2004). The BC emissions in the Arctic have been modelled utilising the AIS-based modelling framework. The model is similar to the activity based model described by Dalsøren et al. (2009), but with engine load factors and time at sea derived for each ship position record from the AIS source. For a given geographical area and time span, the total exhaust gas emission for a given emission component, E (CO2, SOx, NOx, etc.), is calculated by:
E¼
X Me ðM i þ MAUX Þ i
ð2Þ
i
where MMe,AUX denotes the amount of pollutant emitted by ship i for i main (Me) and auxiliary (AUX) engines respectively (kg pollutant).
MMe ¼ i
X x
MMe i;x ¼
X
Me cMe i F i;x
ð3Þ
x
where MMe i,x denotes the amount of pollutant emitted by ship i for main engines associated with AIS position message x (kg pollutant). cMe denotes the main engine emission factor for the given poli lution type for the given ship i (kg pollution/kg fuel). FMe i,x denotes the main engine fuel consumption for ship i associated with AIS position message x (kg fuel).
Fig. 2. Spatial distribution of BC emissions for Arctic region 2012.
25
A. Mjelde et al. / Marine Pollution Bulletin 87 (2014) 22–28
and
LF Me i;x ¼
Me
Ddx =Dtx SSi
3 ð5Þ
where Ddx is the great circle distance between AIS position message x and message x 1 (nm). SSi denotes the service speed of the individual ship i (knots). MAUX is calculated in a similar manner with auxiliary engine i emission factors, SFOC, and load factor based on ship operation mode (Open sea and Harbor mode), as determined based on ship speed over a period of time.
Fig. 3. Monthly break-down of BC emissions for Arctic region 2012 (tons).
Me Me F Me i;x ¼ pi bi LF i;x Dt x
ð4Þ
pMe denotes installed main engine power installed for ship i i (kW). bi Me denotes the specific fuel consumption (SFOC) for main engines for ship i (kg fuel/kW h). LFMe i,x denotes the main engine load factor of the individual ship i in the time interval associated with AIS position message x (–). Dtx is the time between AIS position message x and message x 1 (h).
The results are calculated (Eq. (2)) and presented for the Arctic region, as defined in the proposed International Maritime Organization (IMO) guideline for ships operating in ice-covered waters (IMO, 2010). AIS data from the Norwegian satellite AISsat-1 has been made available by the Norwegian Coastal Administration (NCA). The ship particulars are collected from DNV GL vessels and IHS Fairplay World Register of Ships (IHS Fairplay, 2012). The BC emission factor of 0.35 kg/ton fuel is from the IMO GHG study (Buhaug et al., 2009). AIS data for 2012 is used in the emission modelling. A total of 1347 unique vessels were found to operate in the Arctic region throughout 2012, having collectively sailed distance of
Table 2 Black carbon emissions from shipping in the Arctic, using three different Arctic definitions.
This study Winther et al. (2014) Corbett et al. (2010) Peters et al. (2011)
Reference year
Primary activity data source
Arctic region* (tons)
AMAP area** (tons)
2012 2012 2004 2004
AIS AIS AMSA shipping database COADS/AMVER
105 – – –
1165 – 1230 1150
***
Above 58,95° north (tons) 1330 1584 – –
*
According to the definition given by IMO (2010). According to the definition given by AMAP (1998). An approximation to the actual AMAP area is used. The area used includes parts of the Norwegian and Baltic coastline between 60° and 62° which are not part of the AMAP area, and excludes parts of Canadian waters and the North Sea below 60° which are a part of the AMAP area. A best judgment estimate of the discrepancy introduced by this approximation is 10%. **
***
Table 3 Number of unique vessels on the Arctic region in 2012. Numbers of vessels using HFO are given in brackets. Colour coding indicates the dominating type of fuel within each segment. Full red and green means 100% and 0% HFO, respectively.
Ship type
