A residual activity dataset is the resulting Activity dataset obtained when subtracting data for all available Unit process datasets within an activity class from the data in the activity dataset for the same activity class in a Supply-use table, for the same year and geographical area.
This concept addresses a fundamental challenge in Life Cycle Assessment: supply-use tables provide aggregate economic data for entire activity classes (such as "electricity generation" or "steel production"), whilst LCA practitioners typically work with specific unit process datasets representing individual facilities or technologies within those classes. The residual activity dataset bridges this gap by representing the portion of economic activity not captured by the specific unit process datasets available.
The calculation method is straightforward in principle. Starting with the total activity data from a supply-use table for a specific activity class, year, and geographical area, one subtracts the contributions from all available unit process datasets that fall within that same activity class, year, and geographical area. What remains is the residual dataset, which represents the aggregate behaviour of all the specific activities within the class that have not been individually documented as unit process datasets.
Residual activity datasets serve several important functions in LCA databases. They ensure complete coverage of economic activities, preventing gaps in system boundaries when linking processes. They provide a way to model average or unknown technologies within an activity class when specific process data is unavailable. They also help maintain consistency between bottom-up LCA data and top-down economic statistics.
The quality of a residual activity dataset depends heavily on the completeness and representativeness of the available unit process data. When many high-quality unit process datasets cover a large share of the activity class, the residual dataset represents a smaller, potentially less significant portion. Conversely, when few unit processes are documented, the residual dataset may represent the majority of activity and carry greater uncertainty.
