Key findings from the Third IMO GHG Study 2014
1. Shipping emissions during the period 2007–2012 and their significance relative to other anthropogenic emissions.
1.1. For the year 2012, total shipping emissions were approximately 938 million tonnes CO2 and 961 million tonnes CO2e for GHGs combining CO2, CH4 and N2O. International shipping emissions for 2012 are estimated to be 796 million tonnes CO2 and 816 million tonnes CO2e for GHGs combining CO2, CH4 and N2O. International shipping accounts for approximately 2.2% and 2.1% of global CO2 and GHG emissions on a CO2 equivalent (CO2e) basis, respectively. Table 1 presents the full time series of shipping CO2 and CO2e emissions compared with global total CO2 and CO2e emissions.
For the period 2007–2012, on average, shipping accounted for approximately 3.1% of annual global CO2 and approximately 2.8% of annual GHGs on a CO2e basis using 100-year global warming potential conversions from the IPCC Fifth Assessment Report (AR5). A multi-year average estimate for all shipping using bottom-up totals for 2007–2012 is 1,015 million tonnes CO2 and 1,036 million tonnes CO2e for GHGs combining CO2, CH4 and N2O. International shipping accounts for approximately 2.6% and 2.4% of CO2and GHGs on a CO2e basis, respectively. A multi-year average estimate for international shipping using bottom-up totals for 2007–2012 is 846 million tonnes CO2 and 866 million tonnes CO2e for GHGs combining CO2, CH4 and N2O. These multi-year CO2 and CO2e comparisons are similar to, but slightly smaller than, the 3.3% and 2.7% of global CO2 emissions reported by the Second IMO GHG Study 2009 for total shipping and international shipping in the year 2007, respectively.
Table 1 a) Shipping CO2 emissions compared with global CO2 (values in million tonnes CO2); and b) Shipping GHGs (in CO2e) compared with global GHGs (values in million tonnes CO2e).
Third IMO GHG Study 2014 CO2
% of global
% of global
Third IMO GHG Study 2014 CO2e
1.2. This study estimates multi-year (2007–2012) average annual totals of 20.9 million and 11.3 million tonnes for NOx (as NO2) and SOx (as SO2) from all shipping, respectively (corresponding to 6.3 million and 5.6 million tonnes converted to elemental weights for nitrogen and sulphur, respectively). NOxand SOx play indirect roles in tropospheric ozone formation and indirect aerosol warming at regional scales. Annually, international shipping is estimated to produce approximately 18.6 million and 10.6 million tonnes of NOx(as NO2) and SOx (as SO2), respectively; this converts to totals of 5.6 million and 5.3 million tonnes of NOx and SOx (as elemental nitrogen and sulphur, respectively). Global NOxand SOx emissions from all shipping represent about 15% and 13% of global NOx and SOx from anthropogenic sources reported in the IPCC Fifth Assessment Report (AR5), respectively; international shipping NOx and SOx represent approximately 13% and 12% of global NOx and SOx totals, respectively.
1.3. Over the period 2007–2012, average annual fuel consumption ranged between approximately 247 million and 325 million tonnes of fuel consumed by all ships within this study, reflecting top-down and bottom-up methods, respectively. Of that total, international shipping fuel consumption ranged between approximately 201 million and 272 million tonnes per year, depending on whether consumption was defined as fuel allocated to international voyages (top-down) or fuel used by ships engaged in international shipping (bottom-up), respectively.
1.4. Correlated with fuel consumption, CO2emissions from shipping are estimated to range between approximately 739 million and 795 million tonnes per year in top-down results, and to range between approximately 915 million and 1135 million tonnes per year in bottom-up results. Both the top-down and the bottom-up methods indicate limited growth in energy and CO2 emissions from ships during 2007–2012, as suggested both by the IEA data and the bottom-up model. Nitrous oxide (N2O) emission patterns over 2007–2012 are similar to the fuel consumption and CO2patterns, while methane (CH4) emissions from ships increased due to increased activity associated with the transport of gaseous cargoes by liquefied gas tankers, particularly over 2009–2012.
1.5 International shipping CO2estimates range between approximately 596 million and 649 million tonnes calculated from top-down fuel statistics, and between approximately 771 million and 921 million tonnes according to bottom-up results. International shipping is the dominant source of the total shipping emissions of other GHGs: nitrous oxide (N2O) emissions from international shipping account for the majority (approximately 85%) of total shipping N2O emissions, and methane (CH4) emissions from international ships account for nearly all (approximately 99%) of total shipping emissions of CH4.
1.6. Refrigerant and air conditioning gas releases account for the majority of HFC (and HCFC) emissions from ships. For older vessels, HCFCs (R-22) are still in service, whereas new vessels use HCFs (R134a/R404a). Use of SF6 and PFCs in ships is documented as rarely used in large enough quantities to be significant and is not estimated in this report.
1.7. Refrigerant and air conditioning gas releases from shipping contribute an additional 15 million tons (range 10.8 million–19.1 million tons) in CO2 equivalent emissions. Inclusion of reefer container refrigerant emissions yields 13.5 million tons (low) and 21.8 million tons (high) of CO2 emissions.
1.8. Combustion emissions of SOx, NOx, PM, CO and NMVOCs are also correlated with fuel consumption patterns, with some variability according to properties of combustion across engine types, fuel properties, etc., which affect emissions substances differently.
2. Resolution, quality and uncertainty of the emissions inventories.
2.1. The bottom-up method used in this study applies a similar approach to the Second IMO GHG Study 2009 in order to estimate emissions from activity. However, instead of analysis carried out using ship type, size and annual average activity, calculations of activity, fuel consumption (per engine) and emissions (per GHG and pollutant substances) are performed for each in-service ship during each hour of each of the years 2007–2012, before aggregation to find the totals of each fleet and then of total shipping (international, domestic and fishing) and international shipping. This removes any uncertainty attributable to the use of average values and represents a substantial improvement in the resolution of shipping activity, energy demand and emissions data.
2.2. This study clearly demonstrates the confidence that can be placed in the detailed findings of the bottom-up method of analysis through both quality analysis and uncertainty analysis. Quality analysis includes rigorous testing of bottom-up results against noon reports and LRIT data. Uncertainty analysis quantifies, for the first time, the uncertainties in the top-down and the bottom-up estimates.
2.3. These analyses show that high-quality inventories of shipping emissions can be produced through the analysis of AIS data using models. Furthermore, the advancement in the state-of-the-art methods used in this study provides insight and produces new knowledge and understanding of the drivers of emissions within sub-sectors of shipping (ships of common type and size).
2.4. The quality analysis shows that the availability of improved data (particularly AIS data) since 2010 has enabled the uncertainty of inventory estimates to be reduced (relative to previous years' estimates). However, uncertainties remain, particularly in the estimation of the total number of active ships and the allocation of ships or ship voyages between domestic and international shipping.
2.5. For both the top-down and the bottom-up inventory estimates in this study, the uncertainties relative to the best estimate are not symmetrical (the likelihood of an overestimate is not the same as that of an underestimate). The top-down estimate is most likely to be an underestimate (for both total shipping and international shipping), for reasons discussed in the main report. The bottom-up uncertainty analysis shows that while the best estimate is higher than top-down totals, uncertainty is more likely to lower estimated values from the best estimate (again, for both total shipping and international shipping).
2.6. There is an overlap between the estimated uncertainty ranges of the bottom-up and the top-down estimates of fuel consumption in each year and for both total shipping and international shipping. This provides evidence that the discrepancy between the top-down and the bottom-up best estimate value is resolvable through the respective methods' uncertainties.
2.7. Estimates of CO2 emissions from the top-down and bottom-up methods converge over the period of the study as the source data of both methods improve in quality. This provides increased confidence in the quality of the methodologies and indicates the importance of improved AIS coverage from the increased use of satellite and shore-based receivers to the accuracy of the bottom-up method.
2.8. All previous IMO GHG studies have preferred activity-based (bottom-up) inventories. In accordance with IPCC guidance, the statements from the MEPC Expert Workshop and the Second IMO GHG Study 2009, the Third IMO GHG Study 2014 consortium specifies the bottom-up best estimate as the consensus estimate for all years' emissions for GHGs and all pollutants.
3. Comparison of the inventories calculated in this study with the inventories of the Second 2009 IMO GHG study.
3.1. Best estimates for 2007 fuel use and CO2emissions in this study agree with the "consensus estimates" of the Second IMO GHG Study 2009 as they are within approximately 5% and approximately 4%, respectively.
3.2. Differences with the Second IMO GHG Study 2009 can be attributed to improved activity data, better precision of individual vessel estimation and aggregation and updated knowledge of technology, emissions rates and vessel conditions. Quantification of uncertainties enables a fuller comparison of this study with previous work and future studies.
3.3. The estimates in this study of non-CO2GHGs and some air pollutant substances differ substantially from the 2009 results for the common year 2007. This study produces higher estimates of CH4and N2O than the earlier study, higher by 43% and 40%, respectively (approximate values). The new study estimates lower emissions of SOx(approximately 30% lower) and approximately 40% of the CO emissions estimated in the 2009 study.
3.4. Estimates for NOx, PM and NMVOC in both studies are similar for 2007, within 10%, 11% and 3%, respectively (approximate values).
4. Fuel use trends and drivers in fuel use (2007–2012), in specific ship types.
4.1. The total fuel consumption of shipping is dominated by three ship types: oil tankers, containerships and bulk carriers. Consistently for all ship types, the main engines (propulsion) are the dominant fuel consumers.
4.2. Allocating top-down fuel consumption to international shipping can be done explicitly, according to definitions for international marine bunkers. Allocating bottom-up fuel consumption to international shipping required application of a heuristic approach. The Third IMO GHG Study 2014 used qualitative information from AIS to designate larger passenger ferries (both passenger-only Pax ferries and vehicle-and-passenger RoPax ferries) as international cargo transport vessels. Both methods are unable to fully evaluate global domestic fuel consumption.
4.3. The three most significant sectors of the shipping industry from a CO2 perspective (oil tankers, containerships and bulk carriers) have experienced different trends over the period of this study (2007–2012). All three contain latent emissions increases (suppressed by slow steaming and historically low activity and productivity) that could return to activity levels that create emissions increases if the market dynamics that informed those trends revert to their previous levels.
4.4. Fleet activity during the period 2007–2012 demonstrates widespread adoption of slow steaming. The average reduction in at-sea speed relative to design speed was 12% and the average reduction in daily fuel consumption was 27%. Many ship type and size categories exceeded this average. Reductions in daily fuel consumption in some oil tanker size categories was approximately 50% and some container ship size categories reduced energy use by more than 70%. Generally, smaller ship size categories operated without significant change over the period, also evidenced by more consistent fuel consumption and voyage speeds.
4.5. A reduction in speed and the associated reduction in fuel consumption do not relate to an equivalent percentage increase in efficiency, because a greater number of ships (or more days at sea) are required to do the same amount of transport work.
5. Future scenarios (2012–2050).
5.1. Maritime CO2 emissions are projected to increase significantly in the coming decades. Depending on future economic and energy developments, this study's BAU scenarios project an increase by 50% to 250% in the period to 2050. Further action on efficiency and emissions can mitigate the emissions growth, although all scenarios but one project emissions in 2050 to be higher than in 2012.
5.2. Among the different cargo categories, demand for transport of unitized cargoes is projected to increase most rapidly in all scenarios.
5.3. Emissions projections demonstrate that improvements in efficiency are important in mitigating emissions increase. However, even modelled improvements with the greatest energy savings could not yield a downward trend. Compared to regulatory or market-driven improvements in efficiency, changes in the fuel mix have a limited impact on GHG emissions, assuming that fossil fuels remain dominant.
5.4. Most other emissions increase in parallel with CO2 and fuel, with some notable exceptions. Methane emissions are projected to increase rapidly (albeit from a low base) as the share of LNG in the fuel mix increases. Emissions of nitrogen oxides increase at a lower rate than CO2 emissions as a result of Tier II and Tier III engines entering the fleet. Emissions of particulate matter show an absolute decrease until 2020, and sulphurous oxides continue to decline through 2050, mainly because of MARPOL Annex VI requirements on the sulphur content of fuels.
 Global comparator represents CO2from fossil fuel consumption and cement production, converted from Tg C y-1to million metric tonnes CO2. Sources: Boden et al. 2013 for years
2007–2010; Peters et al. 2013 for years 2011–2012, as referenced in IPCC (2013).
 Global comparator represents N2O from fossil fuels consumption and cement production. Source: IPCC (2013, Table 6.9).