The IPCC framework offers two routes to the same number, and the choice shapes what you measure, how often, and where your uncertainty comes from.
The Two Approaches
Stock-change measures the carbon stock at two points in time and takes the difference. It asks: how much is here now, how much was here before?
Gain-loss tracks the flows, including carbon added through growth as well as carbon removed through harvest, mortality, and decay, and sums them. It asks what came in and what went out.
| Stock-change | Gain-loss | |
|---|---|---|
| Measures | Standing stock, twice | Annual flows |
| Field burden | Intensive inventory, less often | Lighter, more frequent, plus models |
| Main uncertainty | Sampling error on two large numbers | Model error on growth and decay rates |
| Best when | Change is large relative to stock | Change is small, or flows are well characterised |
The Small-Difference Problem
Stock-change has a mathematical weakness that catches projects out. You are subtracting two large, uncertain numbers to get a small one and the uncertainty on the difference can exceed the difference itself.
A forest holding 200 tonnes per hectare measured with 10% uncertainty, gaining 4 tonnes per hectare per year, produces an annual change well inside the noise on either measurement.
The remedies are permanent plots, re-measured, so the change is observed directly rather than differenced, and longer intervals between measurements so the signal grows relative to the noise.
Why Gain-Loss Is Not the Easy Alternative
Gain-loss avoids the small-difference problem and imports a different one: it depends on growth models, mortality rates and decay constants, each carrying its own error and each requiring local validation.
A gain-loss approach built on default growth rates from a different region is not more accurate than a stock-change approach with adequate plots. It is less transparent about where its uncertainty lives.
The Practical Guidance
Most credible forest projects use stock-change on permanent plots for the dominant pool, and gain-loss for pools where direct measurement is impractical, such as deadwood decay, or soil where sampling is destructive.
Mixing them is normal. What matters is that each pool's approach is stated, its uncertainty quantified, and the two are not double-counted where they meet.
Turn climate goals into a verified portfolio
TREEO connects the full carbon cycle, encompassing eligibility, simulation, real time monitoring, and registry ready reporting, thereby combining expert consulting with dMRV technology so the evidence exists before anyone asks for it.
