Water Accounting Basis
Water Accounting Basis
In short: do the whole SWI add-up twice, not once — once counting only water that left the local watershed for good (consumption), and again counting all water pulled from the ground or a river regardless of whether it came back (withdrawal). These are two different questions about two different kinds of pressure on a water system, not two ways of asking the same thing — so they're never averaged or summed together. This classification is the water accounting basis, and it is not optional or a matter of convenience: the specification requires both to be quantified and reported.
Precise notation:
W_c = W_O,c + W_M,c SWI_c = W_c / R
W_w = W_O,w + W_M,w SWI_w = W_w / R
The sections below define each term precisely and explain why they can never be conflated.
Consumption
Water consumption is water removed from the local water system and not returned to the same watershed in a readily available form — water that is evaporated, transpired, incorporated into products or waste, or otherwise lost to that watershed. The clearest example: water that evaporates during cooling-tower operation at a data centre or power plant. Once it has evaporated, it is no longer available to that local watershed, even though it hasn't been "destroyed" in any absolute sense.
Withdrawal
Water withdrawal is water abstracted from a surface-water or groundwater source, regardless of whether it is subsequently returned. The clearest example: water drawn from a river for once-through cooling and then discharged back into that same river. The water was withdrawn — removed from the source, however briefly — even though most of it goes back.
Every withdrawal event has a consumption event somewhere inside it (some fraction of withdrawn water is typically lost to evaporation), but the two numbers are not simply related by a fixed ratio — it depends entirely on the cooling technology and process in question.
Why they must never be summed or substituted
Consumption and withdrawal are different physical phenomena that stress water systems differently, and the specification is explicit that they shall not be summed into a single water volume or SWI value. A facility that withdraws a large volume of water but returns nearly all of it (once-through cooling) has a very different local impact from one that consumes a smaller absolute volume outright (evaporative cooling) — adding the two together would erase exactly the distinction that makes each number meaningful.
Both are required reporting outputs
Per the specification's Reporting clause, the consumption-basis results
(SWI_c, and its stress-adjusted counterpart SWI_c^adj) are
mandatory; the withdrawal-basis results (SWI_w, SWI_w^adj) are
recommended — reflecting, among other things, the withdrawal data gap
described on the Direct Operational Water page. Neither
is an optional alternate to the other where both can be produced — SWI_c
and SWI_w answer different questions and both should be reported when
the underlying data supports it.
Please submit any comments you have here.