The short answer
The fastest way to expose hidden leaks on a large site is to compare meter readings during a quiet night window when no legitimate use occurs. Record the master meter and each submeter before the window and again in the morning; any difference is loss, and dividing it by the hours gives the leak rate in litres per hour. Once a leak is confirmed, isolate it by shutting branch valves one at a time, then rank repairs by size, safety risk and cost rather than order of discovery.
Key takeaways
- Hidden leaks rarely show on the surface, so the reliable signal is unaccounted flow: the gap between what the master meter measures and what the submeters and legitimate uses account for
- A minimum-night-flow check compares readings across a window with zero legitimate consumption; the difference divided by time gives the leak rate in L/h or m³/h
- One night is rarely enough: confirm over several quiet windows and average, because a lone night can catch a shift change, cleaning or an automatic regeneration cycle
- Large sites are isolated top-down: master and sub-meters point to a building or zone, then branch valves are shut one at a time until the suspect leg stops drawing flow
- Repair priority follows risk and volume, not discovery order: fix large, safety- and supply-critical leaks first, then moderate flows, and schedule slow seepage before it becomes a bigger failure
Why hidden leaks are a large-property problem
On a business campus, distribution centre or industrial park, much of the piping is buried, under slab or inside crowded plant rooms. A slow leak in such a route rarely reaches the surface, so it can run for months, quietly inflating water and sewer charges. The U.S. EPA notes that leaks frequently go undetected when a facility is not routinely monitoring use and that leaks can account for more than 6 percent of a facility's total water use.
The scale changes the economics: a running tap here and there seems small, but EPA's own loss table shows how a 10 gpm unattended hose wastes thousands of litres a day and an unattended broken line compounds quickly. Because damage is invisible, the practical early-warning signal on a large territory is not a wet patch but a number — water that was metered at the boundary yet never consumed. Regular comparison of master and sub-meter readings turns a vague suspicion into a measurable, actionable gap.
- Underground or slab-buried piping hides small failures from visual inspection
- A leak that never surfaces is easy to blame on process use, seasons or occupancy
- Every undetected week adds thousands of litres to the site's bill
- Metering and regular balance checks convert the problem from guesswork into a figure
Map your meters before you measure
You cannot manage what you do not measure. For a large property this means two levels of metering: a master or source meter at the point where water enters the site, and sub-meters on each significant end use — individual buildings, cooling tower make-up, irrigation, process or boiler feed. EPA's guidance recommends metering all water conveyed to the facility, tracking each source separately, and using sub-meters to spot leaks and malfunctioning equipment.
Correct meter selection matters as much as installation. EPA warns that an undersized meter causes pressure loss and noise, while an oversized meter is uneconomic and does not accurately measure minimal flow rates. For leak hunting, the low-flow behaviour of the meter is what lets you see a slow night-time drip. Where possible, meters and sub-meters should feed a building management system so usage can be reported hourly and anomalies or continuous flow flagged automatically.
- Master meter records all water entering the site; sub-meters split it by building and end use
- A water balance — master reading minus the sum of sub-meters — exposes the unexplained remainder
- Choose meters for the actual working flow, not the pipe diameter, so small night flows register
- Integrating meters into a central system enables hourly logs and automatic leak alarms
Run the night-flow test correctly
The cleanest reading of hidden loss comes from a window when no water is legitimately used. For most sites that is late night to early morning — often around 01:00–05:00. The method is simple: record the meter indexes at the start of the window, ensure all taps, equipment and toilet cisterns are shut, then re-read before anyone uses water. The leak rate is (morning index − evening index) ÷ hours, giving a figure in litres or cubic metres per hour.
A direct visual check adds confidence: if the meter's low-flow indicator keeps turning while nothing is open, water is passing through it. Where full-night closure is impractical, two off-peak readings hours apart can still reveal a persistent flow. Repeat the test over several nights rather than trusting a single one — a lone reading can coincide with a cleaning shift, an automatic regeneration cycle or equipment that runs briefly. Consistency across nights separates a real leak from a schedule artefact.
- Choose a window with zero legitimate use, typically 01:00–05:00 or full site shutdown
- Before recording, close taps, stop equipment and confirm cisterns and make-up valves are not running
- Note full index readings with date and time at both ends of the window
- Compute loss rate = (morning − night reading) ÷ hours, in L/h or m³/h
- Watch the meter's low-flow indicator directly for a minute or two; a steady turn signals a small leak
- Repeat across 2–3 nights and average before drawing conclusions
Interpret results and avoid false alarms
Not every night flow is a leak. Cooling tower make-up, boiler feed, automatic softener regeneration, overnight irrigation and a night shift all legitimately draw water. Before concluding there is a fault, list every night-time consumer and either exclude them from the clean window or meter them separately. The other common trap is the meter itself: an impulse-based meter may not register an extremely slow trickle, and instrument error adds uncertainty.
Scepticism is cheap compared with an unnecessary excavation. If two or three nights give mixed results, cross-check against the work calendar and equipment schedule, repeat on different days of the week, and if possible install a temporary data logger or test meter on the suspect branch. The point of confirmation is not pedantry — it protects the maintenance team's trust, because an alert that turns out to be a false positive discourages people from acting on the next, real one.
- List legitimate night users — shifts, make-up, regeneration, irrigation — and exclude or meter them separately
- Check the meter's low-flow resolution; very slow drips may not move the register
- Cross-check readings against maintenance and occupancy calendars before concluding
- Re-test over several nights and days; confirm before scheduling any excavation
Isolate the leak from zone to branch
Once a night flow is confirmed, narrow it down from the top. Compare sub-meter readings to find the building or circuit with the anomaly, then work inside that circuit by closing branch isolation valves one at a time while watching the flow. The branch that stops the night draw when shut is the source. This step-testing approach is the practical property-scale equivalent of the district isolation utilities use on municipal networks.
When the branch is known, pinpointing follows: acoustic listening, correlators or specialist equipment can locate a buried failure, and night work helps because daytime plant and traffic noise drowns the sound of a leak. If in-house effort stalls, a specialist leak-detection service can usually cover even a very large site in under a day. In most cases the cost of one such visit is far below the running cost of an undetected leak.
- Sub-meters narrow the search to a building or process circuit
- Shut branch valves one at a time and watch the night flow to identify the guilty leg
- Inside the suspect leg, use acoustic and correlator methods, ideally at night
- If in-house location fails, bring in a specialist service rather than digging at random
Set repair priorities by risk, not discovery order
Leaks differ enormously in consequence, so schedule repairs by risk and volume. Highest priority goes to large or burst leaks that threaten safety, damage property or interrupt supply — these should be repaired immediately. Next come moderate leaks flowing at roughly the rate of a running tap, with no threat to people or property; these can be addressed within days. Slow, hidden seepage without visible impact is planned maintenance rather than an emergency.
Yet a slow leak should not drift indefinitely: EPA advises repairing all detected leaks as quickly as possible because small failures accumulate into large volumes, and an old trickle can become tomorrow's burst. The discipline that makes this work is the feedback loop — after each repair, repeat the night-flow reading and confirm consumption returns to its baseline. That closes the loop from measurement to location to repair to verification.
- Tier 1 — immediate: large or burst leaks, risks to people or property, supply interruption
- Tier 2 — within days: moderate running-tap flows with no safety impact
- Tier 3 — scheduled: slow hidden seepage, but tracked so it is not forgotten
- After any repair, re-run the night reading and compare with the baseline
- Keep a log: discovery date, zone, flow rate, fix date and post-repair verification
Put it into practice
Night-flow leak survey and repair-priority checklist
For a facility manager who suspects hidden losses on a large site, this sequence converts a vague concern into a documented plan. Complete it the evening before a quiet window and again the following morning before anyone uses water.
- Confirm a clean window (usually 01:00–05:00) with no shifts, cleaning or irrigation; note exceptions
- Walk the site: close all taps, stop equipment, confirm cisterns and make-up valves are not running
- Record full master-meter and every sub-meter reading with date and exact time
- Re-read the same meters after the window, before any water is drawn
- Compute each zone: (morning − night) ÷ hours in L/h or m³/h; flag any meter that moved
- Watch each suspect meter's low-flow indicator for 1–5 minutes; steady movement means a leak too small to see
- Repeat across 2–3 nights if readings are ambiguous and average the result
- In a suspect circuit, shut branch valves one at a time to find which leg keeps drawing flow
- Schedule repairs by tier: large or safety-critical first, moderate within days, slow seepage as planned work
- Re-run the same log 2–4 weeks after repairs to confirm readings return to the baseline
Questions people ask
How do I run a night-flow test with my water meter?
Record the full index on the master meter and each sub-meter at the start of a quiet window, typically 01:00–05:00, after closing all taps and stopping equipment. Re-read the same meters in the morning before anyone uses water. Subtract the night reading from the morning reading and divide by the number of hours to get a leak rate in litres per hour. Any meter that moved while nothing was open points to a leak.
What is a normal overnight reading for a large building or site?
In an ideal, fully shut-down facility the overnight reading is near zero, limited only by legitimate night-time uses such as cooling tower make-up, boiler feed or a night shift. Because a meter's register may not move for very slow drips, a stable, non-zero reading repeated across several nights is the practical trigger for investigation rather than a fixed universal threshold.
What if my site has equipment running at night?
List every legitimate night-time consumer — night shift, cooling tower make-up, softener regeneration, overnight irrigation — and either exclude them from the test window or meter them separately. Choose a window when these are guaranteed to be off, or measure the suspect branch in isolation. Only a circuit with no legitimate night draw gives a trustworthy leakage figure.
How do I set a leak alert threshold in a monitoring system?
Avoid hard-coding one absolute alarm value, because legitimate flows vary with season and occupancy. Better to learn the site's normal night-time profile first, then alert on continuous flow or a clear increase above the observed baseline for the same time of day. Confirm any alert over a couple of readings before dispatching a crew, since false alarms quickly erode trust in the system.
When should I call in a specialist leak-detection contractor?
Call when you have confirmed a night flow and narrowed it to a zone or branch but cannot find the exact point yourself — for example a leak running into ground under slab or along an inaccessible buried line. Specialists use loggers, acoustic listening and correlators and can usually cover even a very large site within a day. Their cost is typically far below months of an undetected leak.
Which leak should I fix first?
Prioritise by risk and volume rather than order of discovery. Repair large or burst leaks that threaten safety, damage property or interrupt supply immediately. Address moderate running-tap flows within days. Schedule slow, hidden seepage as planned work, but track it — small leaks accumulate into significant volumes and an old trickle can become a burst.
Sources and further reading
Sources were checked when this page was generated. Confirm changing dates, rules and prices with the original publisher.
- EPA WaterSense: Managing Water in Commercial and Institutional FacilitiesU.S. Environmental Protection Agency (EPA)
- EPA WaterSense at Work: Metering and SubmeteringU.S. Environmental Protection Agency (EPA)
- The 4 best tips for detecting costly water leaksShayp
- How we prioritise and repair faultsWatercare Services (Auckland)
- Приказ Минпромэнерго РФ от 20.12.2004 N 172: Методика определения неучтённых расходов и потерь водыLegalacts.ru (нормативная база РФ)
- МУП «Находка-Водоканал»: Как оценить исправность водомера в домашних условияхМУП «Находка-Водоканал»