GLEC v3.0 vs v3.2 — what changed in the freight emission factor tables

If your calculator is on GLEC v3.0 today and you are thinking about moving to v3.2, this post is about what your numbers actually do on the migration. Not what changed in the framework document — the v3.2 in detail post covers that section by section. This is the practical delta: same shipments, two factor tables, here is the percent move. I will walk through a Shanghai–Rotterdam ULCV voyage in both versions, the mode-by-mode portfolio impact, and the question I keep getting from CSRD-bound clients about whether to restate the prior year.

Quick context. GLEC v3.0 shipped in 2020. v3.1 was a minor revision in 2021. v3.2 shipped in 2023 and is the current edition as of May 2026; v4 is in draft with publication expected late 2026. Most calculators built between 2020 and 2023 are on v3.0; a smaller cohort moved to v3.1; the v3.2 migration is the live conversation right now. The numerical movement between v3.0 and v3.1 is small (sub-1% portfolio-wide). The v3.0-to-v3.2 jump is bigger. Our own per-revision dated record sits in the methodology changelog, which tracks exactly when we pinned each edition into the calculator.

Same shipment, two versions: Shanghai–Rotterdam by ULCV

Twenty tonnes of cargo in a 40-foot container, on a 14,000 TEU ultra-large container vessel, sailing Shanghai–Rotterdam via the Suez Canal. Routed distance: 19,600 km. Single deep-sea leg, no transshipment, no hub.

Under GLEC v3.0: the container-ship factor is undifferentiated — one “container ship, average fleet” WTW value of approximately 8.4 g CO₂e per tonne-km. The arithmetic is 19,600 × 20 × 8.4 = 3,292,800 g, or 3,293 kg CO₂e.

Under GLEC v3.2: the container-ship class is split into 3,000–8,000 TEU and 8,000+ TEU. A 14,000 TEU vessel falls into the 8,000+ TEU class with a WTW factor of approximately 7.6 g CO₂e per tonne-km. The arithmetic is 19,600 × 20 × 7.6 = 2,979,200 g, or 2,979 kg CO₂e.

Same shipment. v3.0 says 3,293 kg. v3.2 says 2,979 kg. That is a 9.5% drop on a single voyage. The reason is that GLEC v3.2 acknowledges what the fleet actually looks like — ultra-large vessels are meaningfully more efficient per tonne-km than the old undifferentiated average. The methodology did not change; the factor got more accurate.

Now swap the vessel. Same lane, same cargo, but on a 6,000 TEU vessel (typical for some round-Africa or feeder configurations after the 2024 Red Sea reroute pattern). v3.0 still gives 3,293 kg. v3.2 puts a 6,000 TEU vessel in the 3,000–8,000 TEU class with a WTW factor of approximately 9.1 g/tkm: 19,600 × 20 × 9.1 = 3,567,200 g, or 3,567 kg CO₂e. That is 8.3% higher than the v3.0 number on the same lane.

So the v3.2 split moves identical lanes by 8–10% in either direction depending on which vessel class you put your cargo on. Operators with portfolios that lean toward ULCV deployment will see their headline drop on migration. Operators with portfolios on smaller vessels will see their headline rise. Most large-shipper portfolios are mixed and the net effect is somewhere between −3% and +3% — the diversity of vessel deployment washes out at scale.

The full portfolio picture, by mode

Five changes between v3.0 and v3.2 move production numbers. I walked through them per-clause in the v3.2 in detail post; here is the migration-impact summary table you can hold in your head.

• Container ship segmentation (sea, v3.2 §5.3). ULCV (8,000+ TEU) factor: ~7.6 g/tkm. Mid-class (3,000–8,000 TEU): ~9.1 g/tkm. v3.0 uniform value: ~8.4 g/tkm. Portfolio swing: typically −3% to +3%. ULCV-heavy portfolios benefit; feeder-heavy portfolios go up.
• LNG marine engine-type breakout (sea, v3.2 §5.4). WTW for modern high-pressure dual-fuel: ~5% lower than v3.0. WTW for older low-pressure dual-fuel: ~5% higher than v3.0. If your LNG bunker exposure is dominantly modern dual-fuel (which it should be in 2026), you see a downward move; legacy fleet exposure moves up. Average across the operating LNG fleet: about −3%.
• Articulated truck >32 t lane utilisation (road, v3.2 §6.2). v3.0: ~60 g CO₂e/tkm with embedded 25% empty-mileage assumption. v3.2: ~62 g/tkm with embedded 28% empty mileage. Portfolio impact: about +3% on road freight using Tier 1 defaults. Operators with Tier 2 carrier-reported empty data can claim back the difference; the change only affects Tier 1.
• Belly cargo allocation (air, v3.2 §7.4). v3.0 used a hybrid volume-and-mass allocation; v3.2 uses ISO 14083:2023’s mass-based default. Numerical impact: under 2% on the typical long-haul belly factor. The bigger win is consistency — different calculators on v3.0 could produce different numbers using nominally the same factor; v3.2 closes that.
• Hub and transshipment line item (all modes, v3.2 §9.1). v3.0 bundled hub emissions into the upstream or downstream leg; v3.2 breaks them out. Per-handling factors are 2–5 kg CO₂e per TEU transshipment and 0.6–1.2 kg per truck-rail transfer. Portfolio impact: +1% to +3% on multi-leg shipments depending on transshipment exposure.

Stack these up across a typical 3PL portfolio doing mixed sea, road, and air freight, and the net migration impact from v3.0 to v3.2 is usually somewhere in the −2% to +5% range on the WTW total. The direction depends on mode mix and on whether you are LNG-heavy and ULCV-heavy or not. The one direction the v3.2 migration almost never moves is sharply downward — if your aggregate emissions drop by more than about 3% on a v3.2 swap, double-check that you have applied the hub breakout, because dropping it is the most common migration error and it makes the headline look better than it should.

When to migrate

Three triggers I tell clients to watch for.

CSRD next cycle. CSRD ESRS E1 assurance providers through 2026 are increasingly asking for the current GLEC edition. v3.0 is two revisions behind. The audit-finding language I have seen in 2025 was “the company uses an emission factor framework that is two minor revisions out of date; methodology equivalence has not been documented.” That is not a fatal finding but it is the kind of thing that turns into a remediation action plan you have to deliver against by the next cycle. If you are CSRD-bound and currently on v3.0, the next cycle is the time to migrate.

CDP submission. CDP’s C6.5a question on methodology scoring favours the current GLEC edition. There is no explicit penalty for v3.0 but the methodology score band is harder to land in at the A- level on an older factor table. If you are pushing for a CDP A-list result and are currently on v3.0, the migration is worth doing before submission rather than after.

Year-on-year comparability concerns from your assurance provider or your shipper. If a shipper customer or an internal sustainability team is starting to ask why year-on-year deltas are noisy, the factor-table version is often part of the answer. A migration that lands at the start of a reporting year is clean; one that lands mid-year is messy. If you can plan the migration to coincide with your reporting year boundary, the year-on-year story stays cleaner.

Restating the prior year

The question I get most: do I have to restate last year on v3.2 if I migrate this year?

Not strictly. Neither CSRD nor CDP requires retrospective restatement when a methodology framework revises. Both accept a forward-looking change disclosed as a methodology footnote. But assurance providers I have worked with prefer to see a restated prior-year baseline because it makes the year-on-year delta interpretable — if your headline moves 8% between 2025 and 2026 and 5% of that is from the factor-table swap, you want to be able to say so cleanly. The clean way is to restate the prior year on v3.2 and footnote both versions.

The slightly less clean way is to disclose the prior year on v3.0 (as originally reported) and the current year on v3.2, with a footnote stating what would have happened if v3.2 were applied to the prior year. That works for CDP. It is less popular with CSRD assurance because the assurance opinion has to cover both years on a comparable basis and the methodology equivalence footnote becomes load-bearing.

The one thing I would absolutely not do is silently swap the factor table mid-cycle. Auditors find this and the finding is fatal in a way that an out-of-date framework is not.

What the migration costs an engineering team

For practical purposes, the v3.0-to-v3.2 migration is a factor-table swap with two schema additions:

• Container ship class field (3,000–8,000 TEU vs 8,000+ TEU), populated from vessel capacity which most carriers will give you on the bill of lading or the SCAC + voyage number lookup.
• LNG engine class field, populated from vessel registry data (the IMO number maps to engine class via the IHS Fairplay register or equivalent).
• Hub line item per multi-leg shipment, populated by your routing data — if your shipment goes via Singapore or Algeciras or Rotterdam, the hub-line emissions get added.

The calculation logic does not change. The data quality tier logic does not change. The WTW / TTW / WTT split does not change. If you have versioned your factor table (you should — v4 is coming), pinning v3.0 and v3.2 side-by-side is a few hours of work and lets you offer customers a switch-over date that does not coincide with their year-end disclosure crunch.

The EcoFreight calculator runs v3.2 as the default and supports v3.0 as a named historical methodology for restating prior-year baselines. The factor source is stamped on every per-leg response, so a calculation made today and the same calculation made on a v3.0 baseline a year ago are distinguishable in the audit log. The factor table is documented on the methodology page, and one published case study walks through what a mid-market forwarder ran into when they restated a full year of trade-lane data on v3.2.

One honest caveat

I do not have a clean number for the v3.0-to-v3.2 delta on inland waterway barge traffic. The factor moved — v3.2 tightened the barge vessel-class definitions — but the published delta is small (sub-2%) and I have not run enough barge calculations through both editions side-by-side to give a confident range. If you ship significant volume by barge and have a v3.0 baseline you are thinking about restating, email and we can look at the specific lane together. The other modes I am confident on; this one I would rather not guess.

The other genuine gap: the hub line item in v3.2 has variable data quality. Some terminals publish their electrification grade and per-handling energy data; most do not. If your calculator falls back to the GLEC default for hubs (which it should, in the absence of better data), you are landing at the upper end of the v3.2 hub-emission range. A Tier 2 override using terminal-specific data can move the hub line by 30–60%, which on a heavy-transshipment portfolio is more impactful than the rest of the v3.0-to-v3.2 migration combined.

Closing

The v3.0-to-v3.2 jump is a real factor-table revision, not a cosmetic one. The portfolio-level impact is usually a few percent in either direction depending on mode mix. The single biggest single-shipment move is on container shipping where the ULCV split moves identical lanes by 8–10% one way or the other. The migration itself is straightforward engineering — factor table swap, two new schema fields, a hub line item. The disclosure-side work (restating the prior year, footnoting the methodology change, getting the assurance provider to sign off) is where most of the actual effort lives.

For the section-by-section changelog of what specifically moved in v3.2, see the v3.2 in detail post. For the broader question of how GLEC factors sit underneath ISO 14083 methodology, see the GLEC vs ISO 14083 post. For the UK-specific factor-table comparison against DEFRA, see the GLEC vs DEFRA post.

Sources

GLEC Framework v3.0, Smart Freight Centre, 2020 — smartfreightcentre.org. GLEC Framework v3.2, Smart Freight Centre, 2023 — same site. ISO 14083:2023 Quantification and reporting of greenhouse gas emissions arising from transport chain operations — iso.org/standard/78864.html. IMO Fourth GHG Study (2020) for marine factor lineage and engine-class breakdown reference. ICCT methane slip studies (2020, 2023) for the LNG WTW revision. European Commission Eurostat road-freight statistics for the post-2022 empty-mileage data underlying the articulated-truck adjustment. Section references throughout (§5.3, §5.4, §6.2, §7.4, §9.1) refer to GLEC v3.2.

The EcoFreight calculator runs GLEC v3.2 as the default and supports v3.0 and v3.1 as named historical methodologies for restating prior-year baselines. For the per-leg data quality tier in the API response, see the freight emissions API documentation.