grgmiller
commented
1 year ago In general, I agree, although we want to be sure to tailor the regions to the use case and the dynamics of power systems.
For carbon accounting data, we stayed away from aggregations like state and NERC regions because these are administrative/geopolitical boundaries that do not necessarily align with grid topology. In addition, my understanding is that the use of NERC regions and eGRID subregions was EPA's way of accounting for consumption-based emissions (instead of explicitly accounting for power flows they just aggregated together the BAs that tend to trade with each other). However, since we are explicitly accounting for these power flows, these broader regions become less necessary.
For the power sector data however, I think it does make sense to aggregate using other boundaries like states, especially if the data will be used for state policymaking. It could also make sense to split up the plant-level data into separate files by state to make it easier to use.
Ultimately, our regional aggregations should be driven by user needs, but also be informed by grid physics.
EPA eGRID offers a range of physical aggregations: plant, state, BA, region, US. We currently offer only one physical aggregation per data type: plant-level for plant data, BA-level for power_system and carbon_accounting.
We should offer regional aggregation for power_system and carbon_accounting for regions with many small BAs (eg, Florida/the southeast). This would make it easier for users to get started even if they're not sure of their BA.
If we want consistency with eGRID, we could offer US averages, though I don't think that's a good use of the data. I think using high temporal granularity (hourly) with low spatial granularity (all US) would give users an unrealistic sense of precision.