EWIF — Hydropower Classification and Treatment
EWIF — Hydropower Classification and Treatment
Of all the generation types an EWIF calculation might encounter, hydropower is the one the SSWG identified as the most methodologically unsettled. Getting a plant's classification wrong here has an outsized effect on the result, because the two kinds of hydro sit at opposite extremes of water intensity.
Reservoir hydro
A reservoir hydro plant impounds water behind a dam, creating a reservoir with a large surface area exposed to evaporation. The NREL factor for reservoir hydro accounts for that evaporation — but the water evaporating from a reservoir is not attributable to electricity generation alone. The same reservoir typically also serves agriculture, flood control, and other industries, so the evaporation is apportioned across those uses rather than charged entirely to electricity.
That apportionment is the genuinely unsettled part: it requires a dataset describing how a given reservoir's water is shared across its uses, and applying that split before treating any of the evaporation as attributable to electricity generation. This apportionment methodology is tracked as an open question pending further resolution within the SSWG — implementers should disclose which apportionment dataset and method they used, since it is not yet a solved, standardised step.
Run-of-river hydro
A run-of-river plant has no meaningful reservoir: water flows through generation and continues downstream essentially unimpeded. There is no large standing reservoir surface to evaporate from, so run-of-river hydro's freshwater consumption is negligible to zero. Treating a run-of-river plant with a reservoir-hydro factor would substantially overstate its water intensity.
Classifying a plant correctly
Getting this distinction right requires plant-level classification data — in the United States, EIA-923 forms identify which type a given hydro facility is. This is one more reason plant-level data is preferable to regional aggregates for indirect operational water: a regional average that blends reservoir and run-of-river hydro together can misrepresent both.
Plant level before aggregation
This generalises beyond hydro: stress adjustment (and, implicitly, correct EWIF treatment) must happen at the level of the individual power plant, before any aggregation up to a grid-region figure. Aggregating first and adjusting for stress afterward loses exactly the information — like hydro's reservoir/run-of-river split — that plant-level data was meant to preserve. See Stress Characterisation — Practical Application for how this ordering applies more generally.
Please submit any comments you have here.