What uncertainty should I put on biomass fuel tonnes in my greenhouse gas inventory, and why is it so much wider than for coal or diesel?

The Intergovernmental Panel on Climate Change (IPCC) recommends plus or minus 50 percent for biomass fuel data when nothing better is known, against plus or minus 5 percent for fossil fuel. That default is written for national statistics. If you weigh every load you can justify less, but the heat value range is usually the bigger problem.

What IPCC says

• Section 1.5.2: if no further data are available, the recommended default uncertainty range for fossil fuel combustion data should be assumed to be plus or minus 5 percent, meaning 0.95 to 1.05 times the point estimate.
• Since data on biomass as fuel are not as well developed as for fossil fuels, the uncertainty range for biomass fuels will be significantly higher. A value of plus or minus 50 percent is recommended.
• The chapter explains why: a large fraction of biomass used for energy may be part of the informal economy, and the trade in fuel wood, agricultural residues and dung cakes is frequently not registered in national statistics.
• That is a statement about national energy statistics. A plant that weighs each load on a weighbridge has better evidence than that. Using a smaller figure then is my reading, and the evidence should be on file.

Three inputs, three ranges

Uncertainty here is half the 95% range as a share of the value, as the inventory uncertainty answer on this site defines it. For other primary solid biomass:

• Quantity: 50% by IPCC's recommended default, which is a recommended value and not a fitted interval, or your own figure if the loads are weighed.
• Net calorific value (NCV) if you use the default 11.6 megajoules (MJ) per kilogram (kg): IPCC's Table 1.2 range is 5.90 to 23.0, which IPCC's footnote marks as expert judgement, so half the range is 8.55 and 8.55 / 11.6 = 73.7%. A tested value replaces this.
• Carbon dioxide (CO2) factor: 84,700 to 117,000 around 100,000 kg per terajoule (TJ), so 16,150 / 100,000 = 16.2%. IPCC's footnote also marks this carbon-based range as expert judgement.

Combining them

• The three multiply, so combine them by root sum of squares, as for any product of independent inputs.
• IPCC's Approach 1 holds well while the standard deviation stays under about 30% of the mean, roughly 60% on a 95% range, as the inventory uncertainty answer notes. The ranges for heat value and CO2 factor here are expert judgement, not fitted, so a result near 90% is a rough size built on them. It is also outside that limit, so use a simulation if the number matters.
• The quickest way to narrow it is to test the heat value and moisture, not to buy a better scale.

1,000 tonnes of rice husk, defaults against measured

• Biogenic CO2: 1,000 t x 11.6 gigajoules per tonne = 11.6 TJ; 11.6 x 100,000 = 1,160,000 kg = 1,160 t
• IPCC defaults, a rough size: square root of (50 squared + 73.7 squared + 16.2 squared) = square root of (2,500 + 5,432 + 262) = 90.5%, so about plus or minus 1,050 t
• Weighed loads at 2% (an assumption), tested NCV at 5% (an assumption), default factor 16.2%: square root of (4 + 25 + 262) = 17.1%, so about plus or minus 198 t
• Same fuel, same tonnes: the range falls from about 1,050 t to about 198 t because two inputs are measured

The 2% and 5% are example figures for a weighbridge and a lab test, not published values. Use your instrument's stated accuracy.

What Senseible does here

Senseible reads fuel invoices and electricity bills into monthly records. Weighbridge slips and any heat value test are yours to keep beside them, because they are what lets you state a narrower range.

• How to estimate inventory uncertainty
• Biomass or waste fuel in a boiler: what is Scope 1
• Which biomass line fits husk, sawdust or dung?
• Evidence for a biomass fuel supplier

Sources: IPCC 2006 Guidelines, Volume 2, Chapter 1 (sections 1.4.1.3 and 1.5, Tables 1.2 and 1.4)

Last reviewed 2026-11-05. Check with your buyer or verifier whether they accept a plant-level range narrower than IPCC's national default.