Third IMO Greenhouse Gas Study 2014

Inventories of emissions of GHGs and other relevant substances 123
2.5 Other relevant substances emissions inventory uncertainty analysis
The uncertainties involved with missing technical data for ships, incomplete geographical/temporal coverage
of activity data and resistance/powering prediction are described in Section 1. Other sources of uncertainty
include estimates of fuel consumption, allocation of fuel types consumed versus actual fuels consumed,
auxiliary engine and boiler loads by mode, assignment of modes based on AIS data, IMO sulphur survey
annual averages, and the factors used to estimate emissions. Uncertainty associated with these, with the
exception of the emissions factors, is discussed in Section 1.5. The uncertainties associated with emissions
factors include the vessels tested compared to the fleet modelled and robustness of the number of ships tested
in each subclass. While some emissions factors have remained within the same ranges since the Second
IMO GHG Study 2009, there were several pollutants that had moderate to significant changes, as detailed in
Section 2.4.1.
2.6 Other relevant substances emissions inventory comparison
against Second IMO GHG Study 2009
Figure 75 presents the time series results for non-CO 2 relevant substances estimated in this study using
bottom-up methods, with explicit comparison with the Second IMO GHG Study 2009 results. Section 1.6
shows that for ship types that could be directly compared, fuel consumption and CO 2 emissions totals
estimated by the methods used in this study compare very well with methods used in the earlier study. As
reported in Section 1.6, the additional precision in observing vessel activity patterns in the Third IMO GHG
Study 2014 largely match general vessel activity assumptions in the 2009 study, at least for the inventory year
2007. (The updated methodology provides greatest value in the ability to observe year-on-year changes in
shipping patterns, which the Second IMO GHG Study 2009 methods were less able to do.)
Given that the Second IMO GHG Study 2009 “concluded that activity-based estimates provide a more correct
representation of the total emissions from shipping”, only bottom-up emissions for other relevant substances
can be compared. The Third IMO GHG Study 2014 estimates of non-CO 2 GHGs and some air pollutant
substances differ substantially from the Second IMO GHG Study 2009 results for the common year 2007. The
Third IMO GHG Study 2014 produces higher estimates of CH 4 and N 2 O than the Second IMO GHG Study
2009, higher by 43% and 40% respectively (approximate values). The Third IMO GHG Study 2014 estimates
lower emissions of SO x (approximately 30% lower) and approximately 40% of the CO emissions estimated in
the Second IMO GHG Study 2009. Estimates for NO x , PM and NMVOC in both studies are similar for 2007,
within 10%, 11% and 3% respectively (approximate values).
These underlying activity similarities essentially reduce the comparisons of other relevant substances
estimated in the Second IMO GHG Study 2009 to a description of differences in EFs, as illustrated in Table 68.
Differences in EFs essentially relate to static values in the 2009 study, which assumed an average MCR and
EFs representative of the average engine’s actual duty cycle. The consortium made more detailed calculations
in the Third IMO GHG Study 2014, which computes hourly fuel consumption and engine load factors and
applies a load-factor specific EF. In theory, if the average EFs across a duty cycle in the earlier study were
computed for the same or similar activity, then the average EFs would mathematically represent the weighted
average of the hourly load-dependent calculations. The crossplots presented in Section 1.6 provide evidence
that the duty cycle assumptions in the Second IMO GHG Study 2009 were generally consistent with the more
detailed analyses presented in this study.
Another source of differences may be related to fuel quality or engine parameter data representing a different
understanding of the fleet technology. For example, the 2009 study assumed fuel sulphur content was 2.7%,
while the current study documents that the typical fuel sulphur content in 2007 was closer to 2.4%. Moreover,
the Third IMO GHG Study 2014 updates these sulphur contents for later years.
Another example is natural-gas-fuelled engines, which are observed in the Third IMO GHG Study 2014
fleet but were not addressed in the 2009 study. This enables better characterization of methane emissions
(sometimes called methane slip), which has been significantly reduced through engine innovations. The
Second IMO GHG Study 2009 characterized methane losses due to evaporation during transport of fuels
as cargo (Second IMO GHG Study 2009, non-exhaust emissions, paragraph 3.47), and used a top-down
methodology to evaluate methane emissions from engine exhaust (Second IMO GHG Study 2009, Tables 3.6
and 3.7). The 2009 study reported total methane emissions (combining exhaust and cargo transport estimates
in Table 3.11), but did not determine any value for international shipping (Second IMO GHG Study 2009,