Processed Food Aggregation

The parts of the aggregation process that are specific to processedFood products.

Aggregation stages

Aggregations for the food processing stage can either be generated at the global level (to maximise data volume) or the country level. For example, if an aggregation for Wheat flour (whole) in Italy is triggered with the option to use global processing data, all Cycles describing flour milling, regardless of their region, are combined into one global milling aggregation. The background data in the aggregation (e.g., related to the crop that is processed, or related to electricity production) are replaced with country-specific values. This is permitted because food processing can be relatively similar in terms of its inputs and products in different countries, whereas the impacts of the agricultural inputs and energy can differ significantly. If sufficient country-level processing data are available, the aggregation can be triggered to use region-specific data instead.

The aggregation process is as follows:

1. When an aggregation is triggered to not use global processing data, all public Cycles that have the processedFood as a primaryProduct for the country specified are gathered. If it is triggered to use global processing data, all Cycles are gathered regardless of their country.
2. Cycles are then filtered to those that have a specific list of inputs, based on the aggregationInputTermIds lookup. For this list of inputs, a HESTIA-verified aggregation is used to fill in the background emissions.
3. The Cycles are rescaled to 1 kg of the primaryProduct (see Rescaling below) and then aggregated.
4. All background emissions are removed from the aggregated Cycle.
5. The background emissions are recalculated using country-specific data. For the crop/processedFood inputs, the verified aggregations for the same period are used for the background emissions. If no verified aggregation is available for the specified period, the next closest period is used.
6. The impact assessments are then recalculated.

Rescaling

All inputs, products, and emissions are rescaled to be relative to 1 kg of the primaryProduct. In practices, only operation data items are rescaled.

Each value a Cycle reports is divided by the amount of primaryProduct that Cycle produced, before the Cycles are combined - so a mill producing 10 t of flour and one producing 1 t contribute on the same basis:

$$R = \frac{v}{P}$$

Where:

• $R$ = the value relative to 1 kg of the primaryProduct, which is what is then aggregated
• $v$ = the value reported by the Cycle
• $P$ = the primaryProduct value of that Cycle, treated as 1 when it is missing or zero

Terms are rescaled only where the aggregationRescaleProductValue lookup is set for them, which is why a data item that is already an intensity (e.g. a percentage) keeps its value.