The number does not exist yet. We make it exist.
Ask a facility how much water it consumed last month and you get a figure derived from an invoice. Ask what it was during the drought week and nobody knows. Here is what we do instead.
We instrument the lines
Sensors on make-up, blowdown, discharge, reuse, fills and sanitary. Clamp-on flow meters, so no pipe is cut and nothing shuts down. Where a building management system or a chemistry vendor already measures something, we read it rather than duplicate it.
We close the balance
Every fifteen minutes: what came in, what went back out, what was reused inside the fence, and what was actually consumed. Consumed is withdrawn minus returned, and it is the only figure the basin cares about.
We score the instruments
Each sensor carries a trust score based on calibration age, drift against reference, and cross-checks. A sensor that cannot be believed produces a HOLD, never a number.
We sign the window
Each window becomes a record: readings, limits, sensor states, verdict, and a hash of the record before it. Signed and chained. A month is 2,976 of them with no gaps.
One console. The operator, the council and the regulator read the same record.
Simulated feed until the first site is live.
| item | value | note |
|---|---|---|
| withdrawn | 3.8 m³ | municipal reclaimed |
| returned | 0.7 m³ | sanitary sewer · SW-22 |
| reused | 0.9 m³ | blowdown to make-up |
| consumed | 2.2 m³ | evaporation + drift |
| indirect | 0.9 m³ | grid electricity · shown separately |
All 7 sensors trusted. Last review 14:00Z by operator K.N. No excursions this window.
What a Water Statement says.
Note the fourth line. We publish our own instrument failing, because a month with no faults in it is a month nobody checked.
See the full document4,312 m³ withdrawn municipal reclaimed 1,788 m³ returned sanitary sewer, agreement SW-22 1,066 m³ reused blowdown recovered into make-up 2,524 m³ consumed 61.0% of the August envelope 1,035 m³ indirect embedded in grid electricity, shown separately windows 2,976 of 2,976, no gaps verdicts 2,961 PASS 15 HOLD 0 FAIL event turbidity sensor Q-02 drifted 14 Aug. 15 windows held and excluded from the consumed total. Recalibrated 15 Aug. Lab sample LB-2261 confirmed 0.12 NTU. ledger sha256 9f3a…c41e, signed
Four tests. All four have to be true.
Measured, not estimated
Taken on the pipe. Not derived from an invoice, and not a percentage assumption.
Continuous, not periodic
Every window, all month. When there is a gap, the gap is visible.
Trust-scored
We record whether the instrument could be believed at that moment, not just what it said.
Replayable
A third party can take the raw records, recompute the totals, and get the same answer. Or discover that they cannot.
The fourth is the one that matters. Everything else is monitoring. Replayability is what makes it verification.
The two fair objections.
Is this not what the municipal water meter does?
For total municipal withdrawal, a modern interval meter does that job well, and we reconcile to it rather than replace it. But a meter is a billing instrument. It measures what you bought.
It cannot see what went back out, and sewer discharge is usually not metered at all, it is estimated as a percentage of the water meter. So the returned figure is an assumption, and consumption is an assumption resting on an assumption. It cannot see reuse as a mechanism, only as a smaller bill. It cannot see water that never touches a city pipe, which in a closed Alberta basin is most of it. And it cannot tell you whether it was working.
Your meter tells you how much water entered the building. The hearing is asking for a record, not a receipt.
Do operators not already monitor their own water?
Many do, and we say so. A building management system usually meters the make-up line and reads conductivity on the tower loop. A chemistry vendor system does more. That is real instrumentation and we ingest it.
But all of it exists to run the plant, not to serve as evidence. Trend logs get recalibrated, backfilled and interpolated as normal practice. There is no signing, no chain, no record of whether a sensor could be trusted. It covers the cooling plant and not the sewer side, so the balance never closes. It is produced by the party being questioned. And in 2025 only 47% of operators measured water at all.
They can see their water. They cannot prove it.
Measurement, not magic.
We are not certifying drinking water. We are not guaranteeing treatment performance. We are not predicting anything.
We are saying: this is what this facility took, returned and consumed, this is how confident the instruments were when we measured it, and here is everything you need to check us.