Every carbon forecast is a single number produced by a dozen assumptions. Sensitivity analysis asks which of those assumptions the number actually depends on, and the answer is almost never the one the team spent most time on.
The Method
Vary one assumption across a plausible range, hold everything else fixed, and record how much the output moves. Repeat for each assumption. Rank by effect.
| Assumption varied | Typical effect on a restoration forecast |
|---|---|
| Survival rate | Large — scales the whole result |
| Growth trajectory | Large, and compounds over time |
| Allometric equation choice | Moderate to large, species-dependent |
| Wood density source | Moderate, linear |
| Planting density | Moderate, but interacts with survival |
| Carbon fraction | Small — well constrained |
The pattern recurs: the parameters with the widest plausible range and the most direct scaling effect dominate, and they are frequently the ones taken from a default without much examination.
Why Single-Point Forecasts Mislead
A forecast stated as one number implies a precision the model does not have. Worse, it invites the reader to treat it as a commitment.
Presenting a range, with the assumptions that produce each end explicitly stated, is both more honest and commercially safer. A buyer who understands that the downside case assumes 60% survival can assess whether that is plausible for the site. A buyer given one number can only accept or reject it.
What a Credible Downside Case Contains
Not a uniform percentage haircut. A downside case built by applying a discount to the central estimate communicates nothing about why the result might be lower.
A useful one names the mechanism: survival below plan because of drought in establishment years, or growth below plan because site quality proves lower than the reference stands. Each of those is checkable, and each points to a mitigation.
The Link to Sampling
Sensitivity analysis on the model tells you which ex-ante assumption matters most. The same logic applies to ex-post measurement: identify which error term dominates total uncertainty before deciding where to spend on plots.
Teams that do both find they are usually pointing at the same thing, which is the parameter that is least well constrained by local evidence.
Turn uncertainty into a plan
Sampling design and growth assumptions decide how much measured carbon survives conservative deduction. TREEO Carbon Simulator models sequestration scenarios and biomass growth before the field season.
