Challenge
A metropolitan council in New Zealand had identified two water loss zones within its distribution network as having persistently high Infrastructure Leakage Index (ILI) values — a key international benchmark used to assess the physical condition of pipe networks and prioritise capital investment.
One zone served a dense urban commercial core anchored by a major hospital, civic recreation facilities, and an active hospitality sector. The other was a heavily industrialised logistics and manufacturing hub where many facilities operated around the clock.
Despite their high ILI scores suggesting significant physical water loss, the council had not been able to reconcile these figures with observable network conditions. The concern was that inflated ILI values could be misdirecting millions of dollars in planned infrastructure renewal towards zones where the pipes were actually in better condition than the data suggested.
Aqua Analytics was engaged to conduct a targeted diagnostic water loss investigation to determine whether these zones were genuinely losing large volumes of water through pipe deterioration, or whether the reported leakage was a mathematical artefact of flawed assumptions in the underlying Minimum Night Flow (MNF) methodology.
The Diagnostic Approach to Water Loss
The project was structured across three phases, each building on the findings of the last
-
Phase 1 — Diagnostic Review:
The initial review examined the data inputs and assumptions underpinning the existing MNF calculations. Three critical deficiencies were identified. First, historical assessments had applied a generic residential nighttime usage allowance — typically 2 litres per property per hour — uniformly across all properties, including heavy manufacturing plants and 24/7 medical facilities. This drastically understated how much water was being legitimately consumed at night, causing the remainder to be incorrectly classified as leakage. Second, the zone inlet flow data and customer meter readings used in the calculations had been captured at different times — in some cases months apart — creating a fundamental temporal mismatch that invalidated the water balance equation. Third, pressure-leakage adjustments had relied on inlet pressure readings alone, without accounting for the actual hydraulic conditions across each zone's distribution network. -
Phase 2 — Field Data Acquisition:
With the diagnostic gaps mapped, a field team was deployed to collect the empirical data needed to replace assumptions with measurements. This involved two parallel workstreams. An active leak detection survey was completed across both zones, identifying 28 physical leaks — 24 in the urban zone and 4 in the industrial zone — with a combined estimated flow of approximately 185 litres per minute. Notable findings included a customer-side leak masked by a seized meter mechanism, a major main leak, and a significant valve leak at a key intersection. A "straight-pipe" connection was also discovered where a meter was missing entirely, meaning consumption at that property was completely unrecorded. Simultaneously, a team manually logged 24-hour consumption profiles for high-volume commercial customers. Any meter with average daily consumption exceeding 5 kilolitres was targeted. Across both zones, over 50 commercial meters were read at four precise intervals over a 24-hour cycle: early morning on Day 1 to establish a baseline, then at approximately 1:30 AM and 3:00 AM on Day 2 to capture the MNF window, and again early morning on Day 2 to close the volume balance. This approach generated interval-specific hourly consumption rates rather than relying on broad daily averages. -
Phase 3 — Analysis, Framework Development, and Reporting:
The field data revealed the true scale of legitimate commercial night consumption and its impact on the leakage calculations.
Deriving Empirical Night-Use Factors
By comparing MNF-window consumption against 24-hour averages, the analysis established verified Night-Use Factors for each zone. In the urban zone, the gross night-use ratio across all read meters was initially 83.7%, but this was heavily skewed by a single exceptional outlier — the major hospital, whose meters alone accounted for over 30 kL/hr of night consumption. Once the hospital was mathematically isolated and treated as a discrete deduction (as its 24/7 operational profile would grossly overestimate consumption if applied to standard commercial premises), the normalised Night-Use Factor for the zone's general commercial sector was calculated at 67.6%. The industrial zone returned a factor of 63.8%, with high-volume accounts at logistics and food processing facilities maintaining continuous water draws throughout the night window.
These figures were dramatically higher than the generic residential allowances previously applied, confirming that vast quantities of legitimate customer consumption had been misclassified as network leakage.
The "Negative Leakage" Anomaly
When the newly verified commercial night consumption rate from the urban zone was subtracted from the last available historical MNF inlet snapshot (captured approximately nine months earlier), the result was a mathematically impossible negative value — implying the network was somehow generating water. This "negative leakage" result proved conclusively that historical inlet data could not be retroactively corrected with consumption data from a different time period. Commercial and industrial demand fluctuates dynamically with production schedules and seasonal requirements, meaning a valid water balance can only be achieved when inlet logging and customer consumption logging occur simultaneously.
Pressure Corrections
A desktop GIS review identified that both zones sit on exceptionally flat terrain, allowing a single mid-point node in each zone to serve as a representative Average Zone Pressure (AZP) location. Specific intersection areas were identified for permanent or temporary logger deployment in future tests. Theoretical Leakage Exponents (N1) were assigned based on pipe material composition — 1.0 for the urban zone's mixed metallic/plastic network, and 1.2 for the industrial zone's predominantly plastic pipework — enabling the calculation of a Night-to-Day Factor to convert 2-hour night leakage rates into 24-hour volumes.
Metering Anomalies
The intensive field sweep uncovered several operational issues that would compromise future data capture if left unaddressed: a critical hospital meter that could not be physically located and was absent from GIS records; a meter fitted with a pulse reader that prevented manual reading; a meter that had been replaced and relocated under a new serial number without the asset database being updated; a new smart meter installation at a major recreation centre that did not match existing records; and a meter lacking any physical serial number, creating identification risks for field crews.
Investigation Outcome
Because the historical data lacked synchronised consumption logging, valid average zone pressure measurements, and proven hydraulic isolation, Aqua Analytics concluded that a reliable historical ILI value could not be calculated for either zone.
The existing ILI figures were confirmed to be overinflated, primarily due to systematic underestimation of legitimate commercial night-time consumption, compounded by temporal mismatches and the absence of pressure corrections.
Rather than attempting to retrofit unreliable historical data, the project delivered a refined MNF Calculation Framework, a step-by-step Standard Operating Procedure designed to govern all future testing in these zones and applicable to similar zones across the wider network.
The framework requires synchronised data capture (inlet logging and customer meter reading on the same night), verified hydraulic isolation through zero-pressure testing, empirically derived commercial Night-Use Factors applied to unread accounts, topographically corrected average zone pressures, and ideally, site-specific leakage exponents validated through physical pressure step-tests.
Key Outcomes at a Glance
Strategic Value for Water Loss
This project demonstrated that what appeared to be a significant infrastructure problem was, in fact, a data and methodology problem.
By replacing generic assumptions with empirical field measurements, the council gained a clear understanding of where mathematical reporting errors end and genuine physical water loss begins.
The refined framework positions the council to produce high-confidence ILI values that can reliably inform network investment decisions — ensuring capital is directed toward zones with genuine infrastructure need rather than zones with overstated leakage metrics.
Aqua Analytics delivered this project through our New Zealand team, which specialises in water loss diagnostics, minimum night flow analysis, and network performance improvement for councils and water utilities. Contact us here.
