Sawmill co-product allocation: mass versus economic keys
A sawmill processes 9,000 t of roundwood per year. Outputs are 4,500 t of kiln-dried sawn timber, 2,700 t of wood chips sold to a panel mill, and 1,400 t of sawdust and bark sold as biomass fuel; 400 t is lost as moisture and residue. The upstream forestry and transport burden is 100 kg CO2e per tonne of roundwood. Sawmill electricity is 450 MWh at 0.40 kg CO2e/kWh, and internal transport adds 40 t CO2e. Average gate prices are EUR 220/t for timber, EUR 45/t for chips and EUR 18/t for sawdust/bark. The LCA practitioner currently allocates the entire burden by output mass. The panel mill, which buys the chips, argues that this overstates the impact of its feedstock and threatens the mill's bio-based product claim. The sawmill wants a defensible allocation for an upcoming EPD.
1. Calculate the chips share of the total burden under mass allocation and under economic allocation, and explain why the two differ.
Read the worked answer
Total burden: upstream 9,000 t x 100 kg CO2e/t = 900 t CO2e; electricity 450 MWh x 0.40 kg CO2e/kWh = 180 t; internal transport 40 t; total 1,120 t CO2e.
Mass allocation uses the three saleable outputs: 4,500 + 2,700 + 1,400 = 8,600 t. Chips share = 2,700/8,600 = 31.4%, so 0.314 x 1,120 = 352 t CO2e.
Economic allocation: timber revenue = 4,500 x 220 = EUR 990,000; chips = 2,700 x 45 = EUR 121,500; sawdust/bark = 1,400 x 18 = EUR 25,200; total EUR 1,136,700. Chips share = 121,500/1,136,700 = 10.7%, so 120 t CO2e.
The difference arises because mass reflects the physical fibre flow, while price reflects the much lower market value of chips per tonne. Neither key is inherently true. The choice changes the chips burden by roughly a factor of three and must be justified by the causal relationship and the decision context, not by which result is preferred.
2. Apply the ISO 14044 allocation hierarchy to this sawmill. What would you try before selecting an allocation key, and which key would you defend for the EPD?
Read the worked answer
First try to avoid allocation by subdivision or system expansion. Subdivision is partly feasible: metered chipping and chip transport can be assigned directly to chips, and kiln drying can be assigned to timber. The forestry burden, however, is joint to the log input and cannot be physically traced to each output without a key. System expansion is weak here because chips displace either panel-board feedstock or fuel, and the displaced product is not unique.
If allocation remains, use a physical relationship that reflects causation. Wood fibre mass is defensible for the forestry burden because chips and timber both consume fibre. It may overstate chips if the economic driver of the sawmill is timber. For an EPD, I would use mass allocation as the primary method because it is transparent and physically based, disclose that timber dominates revenue, and run economic allocation as a sensitivity. A hybrid approach (mass for forestry, direct metering for processing) must be documented and applied consistently.
3. What sensitivity checks and disclosures would you require before the EPD is published?
Read the worked answer
Require a sensitivity table for mass, economic and hybrid allocation, showing both the chips burden and the timber burden, and the effect on the panel mill's downstream product. Check that the 400 t loss is included in the burden denominator and is not treated as a product. Disclose the allocation key, prices and reference year, because economic allocation changes with price volatility.
If the EPD makes a comparative assertion, ISO 14044 requires critical review; the reviewer should see the allocation calculation and the sensitivity, not just the final numbers. Also check that chips are not credited with avoiding wood waste, which would double count the avoided burden. State whether the result is intended for the sawmill's own EPD or for the panel mill's product, because the allocation choice affects them in opposite directions.