EWIF — Data Sources and Implementation Guidance
EWIF — Data Sources and Implementation Guidance
In short: to calculate indirect operational water you need a number for "how much water does generating electricity here actually cost" — that number is EWIF, and this page is where to find it. Use the most detailed data you can get; fall back to broader defaults only when you have to, and say so when you do.
The specification states the EWIF principle — reflect the water intensity of electricity generation at its point of generation, matched to the correct accounting basis, region, and period — but deliberately leaves dataset selection to implementers. This page collects the primary sources the SSWG has discussed as candidates.
Core principle: use the most granular data available
Before any specific source: the guiding rule agreed by the working group is to use the most granular data available for the electricity actually consumed, and to disclose whichever source was used. Regional or default factors exist as a fallback for when granular data isn't available — not as a first resort.
Primary data sources
NREL 2011 water intensity factors. The dominant dataset in this space today, and usable as a disclosed default when plant-level data is unavailable. Worth understanding its limitation directly: much of the secondary literature on electricity water intensity ultimately traces back to this same 2011 NREL dataset. That circularity means independently verified, ground-truth data outside the United States remains genuinely limited — using NREL 2011 as a fallback is reasonable, but it should not be mistaken for an independently corroborated global baseline.
USGS water-use surveys (2008–2020 reanalysis). Supplements NREL 2011 with updated plant-level water consumption figures for the United States.
EIA-923 Forms (United States). The primary source for plant-level water consumption and withdrawal data, reported monthly and annually. Electricity Maps already processes this same data for carbon-intensity purposes, and extending equivalent processing to water is understood not to be significantly harder — this is one of the more tractable near-term data pipelines available to implementers working in the US.
WRI Aqueduct. Primarily a stress-characterization dataset (see Stress Characterisation), but also relevant here: it provides monthly data at watershed-boundary resolution, which is what enables plant-level, seasonally-aware stress adjustment once EWIF values are available.
Regional availability
| Region | Status |
|---|---|
| United States | Available now, via EIA-923 (monthly and annual). |
| European Union | Not yet available; an equivalent is expected around 2028. |
| India | Legislation is advancing data availability faster than in most other regions, though a standardised plant-level source is not yet in place. |
Where granular data genuinely isn't available for a region — which today includes most of the world outside the US — that gap, and the fallback used to cover it, should be disclosed explicitly.
A data quality tier framework
The SSWG drew an explicit parallel to how the SCI framework handles grid carbon intensity data: the specification defines what is required, this guidance indicates where to get it and what quality tier it sits at, and implementers select a source and disclose it. The full tier table lives on the Data Quality and Implementation Tiers page; for EWIF specifically, the tiers map as:
| Tier | EWIF source | Disclosure requirement |
|---|---|---|
| Tier 1 | Plant-level measured data (e.g. EIA-923) | Preferred; disclose source and time period |
| Tier 2 | Regional grid-level data derived from plant-level aggregation | Acceptable; disclose methodology and source |
| Tier 3 | Default NREL 2011 regional factors | Permitted as fallback; must be explicitly disclosed |
GSF does not maintain, and does not currently plan to maintain, a single fixed set of EWIF numbers for implementers to plug in. This guidance points to methodology and sources; it is not a lookup table of ground truth.
Please submit any comments you have here.