Every carbon project begins life in a spreadsheet. Species, areas, growth assumptions, mortality, a discount rate, a tonnage. This is entirely appropriate at the design stage — a workbook is the fastest way to think through a model. The problem begins when that workbook starts producing numbers that go into documents.
Why Workbooks Win Early
| Strength | Why it matters at design stage |
|---|---|
| Immediate | No build step, no environment |
| Visible | Every intermediate value is on screen |
| Shareable | Everyone can open it |
| Editable by non-programmers | The agronomist can change an assumption |
That last row is genuinely valuable and worth preserving in whatever replaces it. Models that only a developer can modify get modified less often and drift out of date.
Why They Fail Later
Logic is invisible. A formula lives in a cell. Understanding the model means clicking through cells, and nobody does this systematically — which is why spreadsheet errors are famously common and famously undetected.
No version history that means anything. `model_v4_final_REVISED.xlsx` is not version control. When a number in a report needs tracing, identifying which file produced it is guesswork.
Copy-paste propagation. A workbook adapted for a second project inherits assumptions specific to the first, silently. This is the single most common way a wrong parameter travels between projects.
No tests. There is no way to assert that a change did not break something elsewhere in the sheet.
The Specific Risk for Carbon Work
A monitoring report figure must be traceable years later, as discussed elsewhere on this site. A workbook edited over three years, by several people, with no history, cannot support that.
The failure is not that the arithmetic is wrong — usually it is right. The failure is that it cannot be shown to be right, and an unverifiable number is treated as an unreliable one regardless of its accuracy.
What Migration Actually Requires
Not necessarily a rewrite into a full application. The properties that matter are narrower than that:
Version control on the model logic, so any reported figure maps to a known state.
Separation of parameters from logic. Parameters in a configuration file that a non-programmer can read and edit; logic in code that does not change per project.
Re-runnability. Same inputs, same outputs, on someone else's machine.
Tests on known cases. A handful of worked examples with expected results, checked automatically.
A scripted model with a parameter file satisfies all four and remains editable by the agronomist, which preserves the workbook's real advantage.
The Pragmatic Position
Keep the workbook for exploration. Scenario testing, sensitivity analysis and conversations with clients are things spreadsheets do well and should continue doing.
Move to versioned, tested code the moment outputs start appearing in documents that will be audited. The migration is not expensive at that point, and it becomes progressively harder as more history accumulates in the sheet.
Projects that make this transition early rarely think about it again. Projects that defer it usually confront it during a verification, which is the most expensive possible moment to discover that the model cannot be reproduced.
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.



