Write a carbon inventory methods note someone else can use

A methods note passes its most useful test when the next preparer can rebuild the calculation without asking its original author what a cell means. A readable workbook is helpful, but it may conceal unit conversions, rounding rules or assumptions that disappear during a software change. Write the note around a repeatable result.

This fictional case follows one generator calculation through a workbook migration. All organisations, records, quantities and emission factors are invented. It is a narrow training extract, not a complete corporate inventory or an actual reporting factor recommendation. Its practical output is a note that explains both the approved result and why a replacement workbook initially fails.

GHG Protocol’s Corporate Standard connects transparency to documented assumptions, an audit trail and enough information for another person to reproduce results from the same source data. Its consistency guidance also addresses explanations of changes to methods and other relevant factors. The worked note below is an original teaching design.

Give the note a specific calculation to govern

Methods note MN-07, version 1.0, governs generator DG-01 at the fictional Cedar workshop for 1 April–30 June 2026. Boundary decision B-04 already includes this generator throughout the quarter. The note covers its fuel-combustion calculation only. Other fuel sources, refrigerants and purchased energy belong to separate inventory workstreams; the result below cannot be labelled the workshop’s total footprint.

The frozen input pack contains three fuel-use summaries: FU-APR records 0.8258 kilolitres; FU-MAY records 1,161.8 litres; FU-JUN records 933.8 litres. Each identifies DG-01 and its complete calendar month. For this exercise, these are reconciled quantities consumed, with no stock movement, return, missing period or estimate to resolve. They are not purchase invoices awaiting interpretation.

Calculation reference F-DEMO-1 supplies an invented combined factor of 2.500 kg CO2e per litre. The note preserves that identifier, value and unit in both workbook versions. A real note would point to the retained applicable factor record and its approval; this exercise does not select a factor or change its coverage.

Describe the transformations outside the spreadsheet

Write the executable sequence in ordinary language: select the three named records for DG-01 and the quarter; preserve each original quantity and unit; convert kilolitres to litres by multiplying by 1,000; retain litre entries unchanged; multiply litres by 2.500; divide kilograms CO2e by 1,000 to obtain tonnes CO2e. Keep unrounded calculated values until the final total.

The input extract needs record ID, generator ID, period, quantity and unit as separate fields. Reject an unfamiliar unit or a blank quantity for investigation. Do not infer litres from the column position. Keep the selection rule in the note even if the workbook normally opens with the right filter already applied.

MN-07 specifies decimal arithmetic and a final two-decimal display, with an exact halfway value rounded upward for these positive quantities. The rounding rule belongs to the calculation instructions. Merely formatting a cell to two decimals is different from replacing its stored value with a rounded number. This distinction becomes decisive during the migration.

Rebuild the original result in three lines

April: 0.8258 × 1,000 = 825.8 litres; 825.8 × 2.500 = 2,064.5 kg CO2e; 2,064.5 ÷ 1,000 = 2.0645 tonnes CO2e.

May: 1,161.8 litres × 2.500 = 2,904.5 kg CO2e, or 2.9045 tonnes CO2e. June: 933.8 litres × 2.500 = 2,334.5 kg CO2e, or 2.3345 tonnes CO2e.

The activity control total is 825.8 + 1,161.8 + 933.8 = 2,921.4 litres. The unrounded emissions total is 2.0645 + 2.9045 + 2.3345 = 7.3035 tonnes CO2e. As a cross-check, 2,921.4 × 2.500 ÷ 1,000 also gives 7.3035. The final displayed result is 7.30 tonnes CO2e.

Record where those results appear: workbook version 1.0, calculation rows APR, MAY and JUN, and output Q2-DG01. Use these stable labels alongside any cell addresses. If the next workbook rearranges columns, the labels still identify what must be recreated. Retain the original workbook as read-only evidence rather than using it as the only explanation.

Original fictional method test. Keeping calculation precision gives 7.3035 tonnes CO2e, displayed as 7.30 after final rounding. The rejected early-rounding workflow displays 7.29. The example uses an invented factor and supplies no factor for real reporting.

Make the replacement workbook prove equivalence

Candidate workbook 2.0 imports the same three records and F-DEMO-1. Its activity total is still 2,921.4 litres. However, its developer has applied rounding to every monthly tonnes value before adding the rows. April becomes 2.06, May 2.90 and June 2.33; the candidate therefore reports 7.29 tonnes CO2e.

The 0.01-tonne difference between displayed totals is a processing defect in this case. Nothing changed in the consumed quantities, factor, generator or period. Matching the activity total did not test the aggregation rule. Record the failed migration test against candidate 2.0; do not silently overwrite 7.30 or describe the lower number as an emissions reduction.

In candidate 2.1, retain 2.0645, 2.9045 and 2.3345 for summation and round only 7.3035 for the final display. Monthly cells may show fewer decimals if their underlying values remain intact. The corrected candidate now reproduces both the unrounded control and the 7.30 output. Keep the failed candidate and the repair description with the test evidence.

Hand the pack to someone who did not build it

Give a second preparer MN-07, B-04, the three summaries, F-DEMO-1 and a blank calculation file. Ask them to show the litres total, unrounded tonnes and displayed output using fresh formulas built from the inputs. Compare their working with version 1.0 and candidate 2.1. A copied total is not evidence that the note was sufficient.

If they obtain a different answer, identify the first divergent step. A litres mismatch points towards selection or conversion; matching litres with different emissions points towards the factor or subsequent arithmetic. A match at 7.3035 but a different published number points towards presentation or export. Repair the note wherever the preparer had to guess.

After the successful test, release MN-07 version 1.1 with the clarified rounding instruction, approved workbook 2.1, reviewer, approval date and unchanged calculation result. Preserve version 1.0 and its outputs. The handover decision is now explicit: the software implementation changed, the approved calculation method and underlying activity did not.

Reproduction establishes that the instructions can be followed. It does not establish that the boundary, fuel records or invented factor would be suitable for real reporting, nor does it provide assurance. Those decisions require their own evidence and review. If an actual method change affects issued figures, the reporting owner must assess the applicable correction and comparison requirements.

For help structuring a calculation handover, contact Nitium Advisory with a non-confidential description of the records, method and unresolved questions.

Source and scope

GHG Protocol Corporate Standard, revised edition, chapter 1, printed pages 8–9: consistency and transparency. Official source checked 14 October 2026.