Does your product really need location specific and hourly consumption based electricity data?
An LCA embeds one electricity factor — usually a national average, from whichever release was current when the study was done. Three things separate that from the grid the plant actually runs on.
Vintage
The release has moved. A study using the 2021 eLCI embedded a US consumption mix near 551 kg CO₂e/MWh; the current release publishes 405.3 — a 26% shift with no change of location or method.
Location
A national average against the region the plant is in. Regional figures across the fleet span roughly 90 to 530 kg CO₂e/MWh.
Time
An annual average against what was actually consumed, hour by hour, including imports traced to whoever generated them.
Read the share before the correction
A grid correction moves a footprint only in proportion to how much of it is electricity. The arithmetic is one line — change in footprint = electricity share × leverage — where leverage means how far the per-MWh emission factor itself can move. Both leverages below are measured from the hourly record, not assumed.
- Where it’s made moves the emission factor by ~41%
- Relocating a plant from a typically clean region (10th percentile) to a typically dirty one (90th) changes the electricity emission factor by about 41% of the US-average factor. Between the single cleanest and dirtiest regions the swing is 69%.
- When it’s made moves the emission factor by ~67%
- Inside a single region, the gap between the dirtiest and cleanest quarter of hours equals about 67% of that region’s own annual-average factor, for the median region. In the most variable regions the gap reaches 198%.
When you produce moves the number about as much as where. That is the part an annual regional average cannot show you, however specific the region.
The threshold
Inverting that line gives the electricity share a product needs before better data changes the answer by even one percent of its footprint:
- Below ~1.5% electricity — neither a specific location nor a specific hour moves the total by 1%. A national annual average is defensible, and anything finer is precision the rest of the assessment does not have.
- Around 1.5–2.4% — real-time data starts to earn its place before location does, because the swing between hours is wider than the swing between regions.
- Above ~2.4% — both levers clear 1%, and by roughly 10% they clear 5%. Location-specific, hourly data is the right choice and a national average is the largest assumption in the study.
If you would rather not compute a share: the same thresholds are about 0.036 kWh per kg CO₂e of declared footprint for timing and 0.058 kWh per kg CO₂e for location — two numbers already in your report, divided, with no factor assumption in the way.
Computed from the hourly record against a national reference of 420.6 kg CO₂e/MWh, so these move with the grid rather than ageing into a rule of thumb. Thresholds scale linearly: for 5% materiality instead of 1%, multiply the shares by five.
Does location matter for your product at all?
The question before “what is my footprint here”. Run one product against every region and read the spread: if the cleanest and dirtiest grids move the total by a fraction of a percent, a national average is defensible and a location-specific factor is false precision. If they move it by twenty, the grid is the largest assumption in the assessment.
Takes electricity’s share rather than its absolute impact, because the share is what a contribution chart already shows — no need to reopen the model.
Read it off your contribution chart. These three recover the electricity factor your study embedded, without needing the model.
Enter a footprint to see how far it moves across regions.
Correct one footprint for one place
Once you know location matters, this puts a number on it: a ZIP code, the electricity your study assumed, and what it would be on the grid that actually served the plant. Start from one of four published footprints or enter your own.
No subregion on file for this ZIP.
Fill in the fields to see the correction.
The worked examples
Four published footprints, chosen because their electricity shares differ by seventy times. Product A is the only one whose source publishes both the electricity use and its impact, so it is the only one where the embedded factor can be recovered by division: 1.532 ÷ 2.78 × 1000 = 551.1 kg CO₂e/MWh, about 36% above the current eLCI US figure.
Product C is the interesting one. No EPD for it publishes an electricity quantity at all. But two facility declarations for the same product — one on a hydro grid, one on a fossil-heavy grid — differ by 45%, and dividing that gap by the measured difference between those two grids recovers roughly 477 kWh per declared unit. That lands inside the published range for its process, which is what makes the derivation worth reporting. It is an upper bound: the facilities differ in other ways too.
Electricity is 0.43% of Product B and 30.5% of Product D. The correction is irrelevant for the first at any grid and decisive for the second — and Product C is 2% or 32% depending only on where it was made.
Boundary. Subregion-level, because that is what a ZIP code resolves to and the level EPA recommends. Our hourly balancing-authority factors are aggregated up using eGRID’s own plant-level generation shares. Methodology · Built from EIA-930 · eLCI · eGRID
Disclaimer. Academic project. US Government data used responsibly; outputs used at your own risk, with no responsibility or liability accepted. Opinions are my own and reflect no position of my employer.