All CDR projects are required to account for uncertainty in the quantification of net CO₂e removal by conducting a sensitivity analysis at their first verification. Learn how to automate your sensitivity analysis in Certify, then account for datapoint uncertainty with the tutorial below.

​
Purpose and approach

The purpose of a sensitivity analysis is to assess the impact of datapoint uncertainty on the calculation of net CO₂e removed. The sensitivity analysis is performed by varying each datapoint by a conservative standard of ±20%. All datapoints are included in the analysis except list inputs related to CO₂ storage, such as measurements of biochar carbon content. These inputs are assumed to be accurate primary data and can be omitted.
In future users will be able to provide custom uncertainty factors to vary each datapoint in the sensitivity analysis instead of ±20%, but this is not yet available.Additionally, whilst in the a contribution analysis is recommended in the Isometric Standard (to reduce the number of datapoints undergoing sensitivity analysis) in Certify this step is skipped, which is acceptable.

​
Steps

1

Input parameter information

Prepare your first GHG Statement ensuring that:
  1. All inputs are filled - Certify will not let you run the sensitivity analysis with missing inputs
  2. No datapoints are duplicated where they correspond to the same real-world data
Prevent datapoint duplication by sharing datapoints across components. The same datapoint must be used to reference the same real-world data across your project’s carbon accounting to ensure correct sensitivity analysis results.
2

Run sensitivity analysis

Run the sensitivity analysis on your draft GHG statement by opening the Sensitivity analysis tab and clicking ‘Calculate sensitivity’. Datapoints that cause a >1% impact to net CO₂ removal are identified as sensitive and flagged by Certify with a yellow dot.
3

Add uncertainty justification

For each input that is flagged as sensitive, a justification should be added in all future GHG Statements. These can be three types:
  1. Adding a standard deviation to the input datapoint
  2. Confirming that the input provided was a conservative value
  3. Justifying that the datapoint is accurate primary data or a scientific constant
See examples in the section below for more information.
Your sensitivity analysis must be recalculated upon any changes to the GHG Statement calculations. If a datapoint’s attribute has changed and the analysis needs recalculating this is flagged on the Sensitivity analysis tab. A sensitivity analysis must be run for the project’s first verification, thereafter it is optional.

​
Justification types

​
Standard deviation

Standard deviations added to inputs are used by Certify in a variance propagation to generate an uncertainty discount term.
Example justification: The carbon content of biochar has a standard deviation of 0.02 (evidenced by the source provided). This standard deviation has been provided in Certify.

​
Conservative input value

If an uncertainty discount has already been applied to the input off platform, then this justification can be used. Alternatively, if the most conservative value was already selected from among multiple possible inputs.
Example justification: The emission factor has an uncertainty factor of 5% and an original value of 3.2 kgCO₂e/kg (evidenced by the source provided). The uncertainty factor has been applied and the datapoint value updated to this in Certify: 3.36 kgCO₂e/kg.
Example justification: When selecting between two similar steel emission factors (1 kgCO2e/kg and 1.2 kgCO2e/kg), the 1.2 kgCO2e/kg was selected.

​
Accurate primary data or scientific constant

If the datapoint is accurate primary data or a scientific constant, it does not require an uncertainty discount.
Example justification: The mass of biochar is measured with a weigh scale as evidenced by the source provided, so the mass is accurate primary data and does not require an uncertainty discount.

​
How are sensitive inputs identified?

If varying a datapoint by ±20% results in a >1% change to net CO₂e removal then the input is defined as sensitive. This is calculated in Certify automatically, in line with the Isometric Standard as follows:
  1. The GHG Statement’s net CO₂e removal (CO2eRemovalCO_2e_{Removal}) is quantified without accounting for any input’s uncertainty.
  2. For each datapoint:
    • The datapoint is varied by ±20% as a conservative standard and net CO₂e removal is recalculated. Whichever calculation (+20% or -20% variation) results in the greatest reduction is taken as the net removal value under the Sensitivity Analysis scenario.
    • The percentage difference (UpU_p) in net removal under the Sensitivity Analysis scenario is calculated. If this is greater than 1% then the datapoint is considered sensitive.