International Shipping

Historical emissions are difficult to estimate for international shipping as there has not been any long-term monitoring and reporting of emissions. The International Maritime Organization (IMO) has published several reports estimating emissions, albeit using different scopes and methodologies. Additionally, other studies have also attempted to measure shipping measures. Those studies and this assessment differ with respect to three components:

1. CO₂ only vs all GHGs,
2. voyage-based or vessel-based approach, and
3. well-to-wake or tank-to-wake.

The IMO’s Fourth GHG Study (2020) calculated both CO₂ and GHG emissions using a voyage-based approach covering the tank-to-wake emissions (TtW), while the previous Third GHG Study applied a vessel-based approach. Following the adoption of the Revised GHG Reduction Strategy in 2023, studies estimate emissions on a well-to-wake (WtW) GHG emissions aligned with the new scope covered by the strategy.

Our analysis and rating methodology is currently limited to CO₂ emissions.

The IPCC GHG inventory guidelines divide shipping into international navigation, domestic navigation and fishing (IPCC, 2006). It defines international shipping as “journeys that depart in one country and arrive in a different country,” while domestic shipping includes “journeys from ports to ports within the same country.” International shipping is responsible for most emissions (87% in 2017) (IMO, 2020). Data incorporating both international, domestic shipping and fishing is often referred to as “total shipping.”

Emissions from domestic shipping and fishing are considered as part of national totals and are included in the assessment of individual CAT countries.

The voyage-based approach calculates emissions beyond national boundaries and therefore excludes domestic navigation emissions. The vessel-based approach covers the entire global fleet giving a more complete estimate of emissions from total shipping. However, GHG inventories are defined by national jurisdictions. To avoid double counting of emissions across all the CAT’s country assessments. For this reason, we select the voyage-based approach for our analysis of international shipping. It should be noted that vessel-based emissions data show higher emissions than voyage-based emissions.

Our assessment is of international shipping only and all references to shipping should be taken to mean ‘international shipping’ unless otherwise stated.

The IMO GHG Reduction Strategy covers vessels greater than 5,000 gross tonnes (GT). It estimated that the remaining smaller vessels account for 15% of the total GHG emissions. However recent AIS based studies estimate to be closer to 5%.

Our analysis covers only ships greater than 5,000 GT, in line with the IMO. Using the voyage-based approach, it is not possible to estimate the additional 5% emission from vessels smaller than 5,000 GT because it is not possible to determine how many of these smaller vessels are operating outside of national boundaries.

1990-2018 estimates

The CAT derived its historical emissions data for the period 1990 to 2018 from the different IMO GHG reports (IMO, 2009, 2014, 2020).

We use data from the IMO Fourth GHG study for years 2012-2018 (IMO, 2020) (which provided both CO₂ and GHG emissions using a voyage-based approach), extended back to 1990 using growth rates from the IMO Third GHG study (2007-2011) (IMO, 2014) and Second GHG study (1990-2006) (IMO, 2009) The 2^nd and 3^rd GHG studies apply a vessel-based approach.

The Fourth GHG Study also provides a direct estimate with for the 2008 which is IMO’s reference year. We use the estimate based on the voyage-based approach from the IMO Fourth GHG study for the year 2008 provided to be 776 MtCO₂.

We take the year-on-year growth rates of the emission timeseries from the Second and Third IMO GHG studies and harmonise to the 2008 voyage-based emissions from the Fourth GHG Study to back calculated the historical emissions from 1990 to 2008. We do the same from the for the period 2009 to 2011 by back calculating from the 2012 voyage-based value. We assume a constant ratio between vessel and voyage- based approaches.

2019-2023 estimates

We use growth rates for international shipping CO₂ emissions in EDGAR total GHG emissions dataset to extend the timeseries to 2023 (European Commission et al., 2024) by harmonising the last historic data point (2018) from the IMO Fourth GHG Study (IMO - MEPC, 2020) emissions with the growth rates from the EDGAR timeseries. This gives an estimate for TtW CO₂ emissions in 2023 of 736 MtCO₂.

Conversion factor to switch from WtW GHG to TtW CO₂

DNV’s Comprehensive Impact Assessment (MEPC 82/INF.8/Add.1) constructs a 2023 well-to-wake (WtW) GHG inventory for the MARPOL Annex VI Chapter 4 fleet (vessel-based scope) and maps it to the Fourth IMO GHG Study’s voyage-based scope for international shipping, explicitly following the Fourth Study’s method to estimate the 2023 emissions. Methodologically, DNV decomposes WtW into well-to-tank (WtT) upstream emissions and on-board TtW components (CO₂ from fuel carbon content; CH₄ and N₂O from engine-type TtW factors), using the Fourth Study’s TtW emission factors for combustion species.

For 2023, DNV estimates WtW GHG emissions to be 928 MtCO₂e, of which the WtT upstream emissions are 153 MtCO₂e, the TtW methane (CH₄) emissions are 7.9 MtCO₂e and emissions from nitrous oxides (N₂O) are 11.2 MtCO₂e. Subtracting these from the 928 MtCO₂e provides the residual TtW CO₂ emissions of 756 MtCO₂, implying a conversion factor from WtW GHG to TtW CO₂ of 0.8145. Applying this conversion factor to DNV’s 2023 estimate for voyage-based international shipping WtW GHG emissions of 900 MtCO₂e yields a TtW CO₂ value of 733 MtCO₂. This leaves a discrepancy of 3MtCO₂ compared with the 2023 estimate of 736 MtCO₂ from the method above which we will assume to be negligible.

We use this conversion to estimate the current policy projections as outlined below.

This data appears as the solid black line in our graph.

For comparison purposes, we also include emissions calculated using the vessel-based approach (light grey line). These emissions are calculated using the same sources and methods as above. For more on the different approaches, see here.

2030 Carbon intensity target

We follow the methodology outlined by the ICCT in 2021 (ICCT, 2021) which estimates the 40% carbon intensity target using the 2008 baseline emission value and BAU derived from the IMO’s Fourth GHG Study (IMO, 2020). The ICCT (2021) study quantifies the resulting emissions in 2030 using a vessel-based approach for using first the EEOI method and then converting to AER. We replicate their works and then convert the EEOI vessel based into EEOI voyage based using a conversion factor of 0.81. This provides the emissions in 2030 which represents our CI target quantification. Note we only use the quantification from the EEOI method.

2030 and 2040 ‘indicative checkpoints’

The indicative goals apply to all GHG emissions (excluding black carbon); however, we calculate CO₂ emissions only as this is the basis of our rating system.

For 2030, we apply the 20% and 30% reduction goals to the 2008 value (voyage-based approach) for CO₂ emissions provided in the IMO’s Fourth GHG study. We apply the same approach to the 2040 reduction goals.

Our analysis uses AR5 Global Warming Potential. (GWP) values in line with the IMO’s Fourth GHG Study.

We are not able to calculate the net zero target because no information is provided on the level of remaining emissions from international shipping by 2050 and beyond. Furthermore, with a deadline of “by or around 2050”, it is uncertain when exactly shipping will exactly need to reach net zero emissions.

For the purposes of our global temperature estimate, we have used a value of 0 in 2070 to account for these uncertainties.

The top end of our current policy projection range is based on a scenario from UMAS (Smith, Perico, et al., 2025). The study bases its projections on the SSP2_RCP2.6 L scenario, which corresponds to part of our previous upper bound in our 2021 assessment, but includes the impacts of the pandemic and the latest energy efficient measures. We use the growth rates from this projection and apply these to the last historic year.

The bottom end of our range is based on the minimum values of the three scenarios included in the IMO’s Fourth Study logistics model method (OECD_RCP26_L) (IMO, 2020). To estimate the impact of the latest energy efficiency measures, we assume that this scenario will fall in equal proportion to difference between our previous top and bottom ranges and the UMAS study to scale this projection downwards.

Emissions estimates in the 4^th GHG Study are for all shipping. We take the ratio between all shipping and the voyage-based approach in 2018 to derive voyage-based estimates. The UMAS study uses the vessel-based approach, we assume a constant ratio between vessel and voyage.