Data Quality and Uncertainty in a GHG Inventory

Every figure in a carbon footprint is an estimate. The TGO Requirements for Calculating and Reporting the Carbon Footprint of an Organization (2022) accept that, but they do not accept silence about it. Section 8 obliges an assessment, Appendix 7 explains where uncertainty enters and how to keep it small, and Appendix 11 gives a scoring example.

#What the requirements mean by uncertainty

The definitions describe uncertainty as a parameter associated with the result of a quantification that characterises the dispersion of the values and the qualitative causes of that dispersion (TGO CFO Requirements, 2022, p.12). Section 8.3 turns that into an obligation: the organization must assess the uncertainty associated with its quantification approach, the data and models it uses, and assess the uncertainty of the inventory. Where a quantitative estimate is not possible or not cost-effective, it must explain why and make a qualitative assessment instead, and it may draw on the method in Appendix 11 (TGO CFO Requirements, 2022, p.29).

Appendix 7 adds structure. The organization should set a specific uncertainty criterion for the final emission figure, assess uncertainty at the level of each emission category, and recognise three sources: parameter uncertainty in the factors and activity data, scenario uncertainty from conditions such as the use phase or end of life, and model uncertainty (TGO CFO Requirements, 2022, p.74).

#Where uncertainty enters: data types and tiers

Appendix 7 lists the kinds of data a quantification method uses: activity data such as mass, volume, energy or financial value; calorific values; emission factors; composition data such as carbon content; oxidation and conversion factors; emission rates per period; and monetary values. Data are primary or secondary depending on who collected them, and site-specific or not depending on whether they came from the actual source (TGO CFO Requirements, 2022, p.67).

The choice between those data types is what the document calls a tier, and its Box 1 makes the range concrete for fuel combustion. A simple tier takes annual fuel quantities from purchase invoices and multiplies them by an IPCC default factor, ignoring unburned carbon and other gases such as methane. A complex tier monitors natural gas flow continuously with temperature and pressure instruments, holds overall data uncertainty below 1.5 percent, determines the factor with a gas chromatograph that samples 4 to 8 times per hour under ISO 10715, and calibrates and validates the system with ISO 17025 certified gases and an annual ISO 10723 check (TGO CFO Requirements, 2022, p.68). The tier chosen decides the uncertainty, and the choice must be documented.

#Controls that keep uncertainty down

#Document the data flow

The organization should keep written procedures for data transfer that cover five things: identification of the primary data sources; each transfer step from primary data to annual emissions; the specific operations in each step, meaning the formulas and the data used; the processing and electronic storage systems involved, including manual entry by staff; and how the output of each step is recorded (TGO CFO Requirements, 2022, pp.69-70).

#Sampling, laboratories and calibration

Sampling and analysis should follow international or national standards, and results apply only to the batch they represent. The example is a cement plant that samples its limestone each month, analyses the calcium oxide content and uses the monthly average for that month's calcination emissions (TGO CFO Requirements, 2022, p.70). A written sampling plan should state responsibilities, location, frequency, quantity and sample handling (TGO CFO Requirements, 2022, p.71). Laboratories should be accredited or demonstrate the competence needed, and instruments should be calibrated at least at the minimum frequency set by the manufacturer (TGO CFO Requirements, 2022, p.71).

#Data gaps and non-site-specific data

Missing data should be filled with an estimation method that produces conservative substitute values for the missing period or parameter, and good practice is a written estimation procedure (TGO CFO Requirements, 2022, p.71). When site-specific data are not collected, or for processes of secondary importance, the organization may use independently verified global or regional averages, or secondary data such as default factors from the literature. In every such case it should record the numeric value and source of each factor used, including emission factors, oxidation factors and GWP values, together with the reason for choosing them (TGO CFO Requirements, 2022, p.72).

#Information management as a quality system

Section 8.1 requires a documented GHG information management procedure that keeps the inventory consistent with the guideline and its intended use, provides regular checks that the data are accurate and complete, identifies errors and omissions, and records the information management systematically (TGO CFO Requirements, 2022, p.27). The documented steps must cover roles and training, review of boundaries, sources and methods, consistency across business units, instrument calibration, accurate data storage, routine accuracy checks, periodic internal audits and technical reviews, and review of opportunities to improve (TGO CFO Requirements, 2022, p.28).

Appendix 7 lists the records that good practice keeps, from activity data, default values and analysis results to calibration records, the reasons for the chosen approach and the uncertainty assessment itself, and notes that where reporting is required by law the usual practice is to keep them for 10 years (TGO CFO Requirements, 2022, pp.71-72).

#Scoring data quality: the Appendix 11 example

Appendix 11 shows one quantitative approach. Company A has three sources: fuel oil recorded continuously with a supplier factor, electricity from invoices with a Thai database factor, and petrol for staff travel estimated with an IPCC factor. Data collection is scored 6 for continuous automatic collection, 3 for meters and invoices and 1 for estimates; the factor is scored 4 for a measured factor, 3 for a manufacturer or national factor, 2 for a regional factor and 1 for an international factor. The two scores are multiplied and the product graded: 1 to 6 is high uncertainty and poor quality, 7 to 12 slight uncertainty and moderate quality, 13 to 18 low uncertainty and good quality, and 19 to 24 low uncertainty and excellent quality (TGO CFO Requirements, 2022, p.79).

Fuel oil scores 6 times 3, or 18, and lands in level 3. Electricity scores 3 times 2, or 6, and petrol 1 times 1, or 1; both land in level 1. The organization can use these levels to plan uncertainty management for the next assessment (TGO CFO Requirements, 2022, p.80). The lesson: the electricity bill, usually the most trusted figure, sits at the lowest level in the example because of the factor paired with it. The quickest improvement is often to the factor, not the meter.

#What the report must contain

The required content of a GHG report includes an explanation of the effect of uncertainty on the accuracy of emissions and removals for each source category, and a description and the results of the data uncertainty assessment (TGO CFO Requirements, 2022, p.32). Appendix 15 gives this its own item 3.5, uncertainty management: the assessment and explanation of the uncertainty, and the approach to managing it in data collection, calculation method, assumptions and the factors chosen (TGO CFO Requirements, 2022, p.88).

#How CarbonBiz supports this

The two axes of the Appendix 11 score map onto fields the App already stores. Each activity is tagged as measured, calculated or estimated and carries a calculation-method string, the data collection axis. Each emission factor carries its source, standard version and GWP version, and custom Tier 2 and Tier 3 factors must state a reason and a source URL, the factor axis and the record Appendix 7 asks for. Special source modules store raw inputs so a result can be replayed, and edits are written to the audit log with before and after values. The uncertainty assessment itself, and the written explanation the report requires, remain the organization's work.