A cargo's emissions figure should include the voyage home
Shippers reporting transport emissions are often given a factor-based figure for the loaded voyage alone. Carriers that can measure fuel should report what was burnt, allocate the empty return by a published rule and show the uncertainty, even when the result is less flattering.
More of our customers now ask for the greenhouse-gas emissions of the cargoes we carry. A steelmaker wants the figure for the iron ore we bring it, a grain company for the wheat we move to export elevators, a city for the road salt it spreads. Their reasons vary. Some report transport emissions to investors or lenders, some have their own reduction commitments and some are preparing for disclosure expectations they think will tighten. I believe the figure carriers give them is often too low, for a reason that has little to do with fuel efficiency. It leaves out the voyage home.
My position is that a carrier's emissions statement for a cargo should be built from fuel actually burnt, should include a share of the empty voyage the vessel or truck made to reach that cargo or return from it, allocated by a rule the carrier publishes, and should show its uncertainty. Shippers should prefer that figure to a smaller one built from an average factor, even when it means explaining why their reported transport emissions went up.
The voyage home is part of the delivery
A laker carrying iron ore down the lakes from a Lake Superior port to a steel mill on Lake Erie often returns with nothing in its holds, sailing in ballast. Some trades pair well: salt one way and stone the other, or grain down one leg and ore back up another. Many do not. Across our fleet in 2025 about 38% of the distance our vessels sailed was in ballast. A truck delivering road salt to a municipal depot in January almost always drives back empty.
The ballast voyage exists because of the cargo. The steel mill's ore could not have been delivered unless the vessel first reached the loading port, and the vessel would not have been there otherwise. Leaving that voyage out of the cargo's emissions is like costing a delivery without the return journey. A shipper that receives only the loaded leg will under-report, and two carriers serving the same trade can produce very different figures depending on whether they include it.
The recognised methods for transport emissions expect empty running to be counted, usually by averaging it across a defined set of operations. That is sound in principle. In practice the set is often drawn so broadly, across a whole fleet or a whole year, that a trade with a long ballast leg is blended with trades that have almost none. We allocate ballast by trade instead. The empty upbound voyages that exist to serve ore shippers are allocated to ore cargoes; a short repositioning between a salt discharge and a stone loading is shared between the two. The rule is written down, and a customer can see which voyages its cargoes carry a share of.
Measured fuel against an assumed factor
The other common shortcut is the factor. A carrier multiplies tonnes by kilometres and by an industry-average figure for the vessel or vehicle type, and sends the result. It is simple and comparable, and for a shipper dealing with hundreds of carriers it may be the only practical approach.
For a carrier that can measure its fuel, I think it is the wrong one. Every Morvaise vessel has flow meters on its main engines and generators, read continuously since 2019, and every tractor in our road division buys fuel on a card tied to that vehicle. We know what each voyage burnt. A laker's consumption varies with its age, its loading, the season, the hours spent in locks and at anchor, and the weather on the lakes. Two cargoes of the same tonnage over the same distance can differ by a fifth. An average factor erases those differences, including the ones a customer could influence, such as the hours a vessel waits for a berth with its generators running.
Measured fuel produces a figure that moves when operations change, which is what a shipper trying to cut its emissions needs. When we worked with one grain customer to reduce waiting at its elevator, its cargoes' emissions fell measurably the following season. A factor-based statement would have shown no change at all.
Publish the rule with the figure
Measured figures raise a problem that factors avoid: they need a method, and methods involve choices. How should fuel burnt in a lock queue be divided between cargoes on a part-loaded voyage? How much ballast belongs to which trade? What happens when a meter fails for a week? Each choice moves the result.
So we publish the method with every statement. For each cargo, customers receive the fuel attributed to the loaded voyage, the share of ballast allocated to it, the conversion factors used for each fuel type and a range reflecting meter accuracy and the gaps we filled by estimate. Our auditors review the method every year. When we change it, as we did in 2025 to treat time at anchor separately, we restate the previous season so that customers can compare like with like.
Not every customer has welcomed this. In 2025 our measured figures for ore cargoes came out about a quarter higher than the factor-based figures one steelmaker had been using. Its sustainability team asked whether we could also send the lower figure, for consistency with its earlier reports. We sent both, clearly labelled, with an explanation of the difference. I would not send only the lower one, because we could not reproduce it from anything we had measured.
My argument has honest limits. A shipper buying from many carriers cannot reconcile a different method from each, so some convergence on common rules is needed. Measurement costs money, and smaller carriers may not have meters on every engine. A measured figure can also be wrong, if meters drift or allocation rules are poorly chosen, and it may carry an authority it has not earned.
None of that persuades me the shortcut is better. I would ask a shipper's board to request three things from its larger carriers: figures built from measured fuel wherever the carrier can measure it; a written rule for empty running, showing what share of each cargo's figure it represents; and a range rather than a single number. Then I would ask the board to accept that the first honest figure may be higher than the last convenient one. A reduction plan built on the convenient number would be measuring progress against something that never happened.