Transportation and distribution emissions — GHG Protocol Scope 3 Categories 4 (upstream) and 9 (downstream) — account for 20–35% of total Scope 3 for most manufacturers. They're also among the more tractable categories to calculate, because the source data usually exists in structured or semi-structured form: freight manifests, carrier invoices, transportation management system (TMS) exports, and increasingly, carrier emissions APIs.
The challenge isn't conceptual — the activity-based methodology is well-defined. The challenge is data normalization: freight data arrives in inconsistent formats, missing key fields (especially weight), and without the mode classification you need to apply the right emission factor. This post walks through the methodology step by step and flags the specific data quality problems that cause calculations to go wrong.
Category 4 vs. Category 9: Getting the Classification Right
GHG Protocol defines these categories from your company's perspective, not the carrier's:
Category 4 — Upstream transportation and distribution: Transportation of purchased goods from your suppliers to your facilities. Also includes transportation between your own facilities (warehousing, inter-facility transfers). The emissions belong to you even if the freight is arranged by your supplier, as long as the cost is reflected in the purchase price or you control the transportation decision.
Category 9 — Downstream transportation and distribution: Transportation of goods you sell, from your facilities to your customers or to intermediate warehouses/distribution centers. If you sell FOB origin (customer arranges pickup), those emissions are technically Category 4 for your customer, not Category 9 for you — though some companies include them voluntarily for completeness. If you sell FOB destination or DDP (you arrange the freight), it's Category 9.
The most common misclassification: companies with mixed Incoterms across their customer base calculate Category 9 inconsistently. An annual review of your standard Incoterms by customer segment is worth doing before finalizing your boundary statement.
The Activity-Based Calculation: Core Methodology
The GHG Protocol guidance for freight transportation uses tonne-kilometers (tkm) as the activity metric:
Step 1: For each shipment, determine the weight in tonnes and the distance in km. Weight typically comes from the bill of lading or freight invoice. Distance is calculated from origin and destination zip codes or locations — using either straight-line (great-circle) distance with a detour factor, or a routing engine for more precision.
Step 2: Calculate tkm per shipment: weight (tonnes) × distance (km) = tonne-kilometers.
Step 3: Apply the emission factor for the relevant mode and vehicle type. The emission factor is expressed in kg CO2e per tonne-km.
Step 4: Sum across all shipments by category (upstream vs. downstream) and mode.
The calculation at the shipment level is straightforward. The difficulty is in having clean weight and mode data across thousands of shipments from multiple carriers in a single year.
Emission Factors by Mode: What to Use and Where to Get Them
For North American freight, the primary emission factor sources are EPA's MOVES model outputs, the SmartWay program emission factors, and for international freight, the GLEC Framework (Global Logistics Emissions Council) published by Smart Freight Centre.
Indicative factors (these change with annual EPA and GLEC updates — always use current-year values):
- Road — full truckload (FTL): 0.095–0.130 kg CO2e/tkm, depending on truck type and load efficiency. SmartWay carrier-specific values, where available, are more precise than fleet averages.
- Road — less-than-truckload (LTL): 0.140–0.200 kg CO2e/tkm. Higher than FTL because of partial load efficiency and more stops. LTL emission factors should reflect the consolidated load's full vehicle emissions divided by the actual payload, which depends on your shipment's weight relative to the full truck capacity.
- Rail — domestic intermodal: 0.020–0.035 kg CO2e/tkm. Significantly lower than road.
- Ocean — container ship: 0.010–0.020 kg CO2e/tkm depending on vessel size and route. Smaller vessels and older fleets run higher.
- Air freight: 0.500–0.800 kg CO2e/tkm. Typically 30–50× the road factor. Even a small air freight volume can dominate freight emissions if weight is significant.
A practical note on LTL: the correct emission factor for LTL depends on whether you're using a vehicle-level approach (allocate total truck emissions by weight fraction) or the SmartWay carrier-level approach (carrier reports fleet-wide tkm and total fuel use). For ESRS E1 and GHG Protocol, both are acceptable — document which you used.
A Worked Scenario: Mid-Market Auto Supplier, Inbound Freight
A mid-market auto parts manufacturer based in the Detroit area receives components from 22 tier-1 suppliers. Their inbound freight for a typical year looks like:
- FTL road, domestic: 3.2 million tkm
- LTL road, domestic: 680,000 tkm
- Rail intermodal: 420,000 tkm
- Air freight, expedited: 8,400 tkm
Applying midrange emission factors:
- FTL: 3,200,000 × 0.110 = 352,000 kg CO2e = 352 tCO2e
- LTL: 680,000 × 0.165 = 112,200 kg CO2e = 112 tCO2e
- Rail: 420,000 × 0.028 = 11,760 kg CO2e = 12 tCO2e
- Air: 8,400 × 0.640 = 5,376 kg CO2e = 5 tCO2e
Total Category 4: approximately 481 tCO2e. The air freight volume is tiny in tkm terms but contributes disproportionately because of the much higher emission intensity. This is a useful internal finding: if the air freight is driven by a few urgent components from specific suppliers, there's a targeted reduction opportunity worth analyzing.
The Weight Problem: When Weight Is Missing or Estimated
The most common data quality gap in freight calculations is missing shipment weight. Carriers are inconsistent about including weight on invoices — some bills of lading have gross weight, some have dimensional weight, and some have nothing useful.
Options when weight is missing:
Map to product type and apply a density factor. If you know a shipment contains 48 cases of a specific component and you know the product weight from your ERP item master, you can reconstruct shipment weight. This works well for manufacturers with structured order-to-shipment data in their ERP.
Use billed weight for LTL. LTL carriers bill on the greater of actual or dimensional weight. The billed weight on an LTL invoice is a reasonable proxy for actual weight, particularly for dense industrial goods.
Apply carrier-level spend-based estimates as fallback. If you have carrier spend data but no weight data for a subset of shipments, you can use the EPA SmartWay carrier factor expressed per dollar of freight spend as a fallback. Document this clearly — it's a spend-based estimate within an otherwise activity-based calculation, and your disclosure methodology statement should note it.
We're not saying you should accept significant weight data gaps and fill them all with spend-based estimates. If more than 20% of your freight tonnage is estimated rather than measured, that's a data quality issue worth flagging in your disclosure and addressing in your carrier data agreements. Many carriers — particularly the larger LTL networks — can provide structured shipment-level data including weight via API or EDI if you request it.
Carrier Emissions APIs and Third-Party Verification
Several major freight carriers now offer emissions data through APIs or portals. The SmartWay program publishes annual carrier-level emission intensity data for enrolled US carriers — roughly 3,500 carriers covering a significant fraction of domestic freight. If your carriers are SmartWay-enrolled, using their program-specific emission intensity rather than mode-average factors is a methodological upgrade that your auditor will appreciate.
For international freight, the GLEC Hub allows shippers to query carrier-level emission factors for ocean and air carriers that have submitted data to the program. Coverage is growing but not universal; for carriers without GLEC data, the GLEC default factors by mode and route are the fallback.
An important nuance: carrier-reported emission factors are not audited the same way your own Scope 1 and 2 are. You're taking the carrier's word that their reported intensity reflects their actual operations. For limited assurance purposes, using SmartWay or GLEC data is generally considered acceptable — these programs have their own verification requirements. But it's a weaker provenance than your own measured data.
Building the Calculation Model
For a manufacturer running 5,000–20,000 freight shipments per year, the Category 4 and 9 calculation model should be structured as:
A shipment-level table with columns: shipment ID, date, origin zip, destination zip, mode, weight (kg), calculated distance (km), tkm, emission factor applied (source + value), and calculated CO2e. Summed by month and by mode for the annual total. This structure makes spot-checks trivial — if your auditor asks about a specific quarter or a specific carrier, you can filter the table rather than reconstructing a calculation.
The distance calculation deserves care. Great-circle distance with a standard detour factor (typically 1.25–1.30 for domestic road, 1.05 for ocean, 1.10 for rail) is widely used and acceptable. If you use a routing API that produces actual road distances, that's more precise — document which approach you used and apply it consistently.
— Natasha Rivera, CEO & Co-Founder, Circulyft