The IWA water balance explained.

Utilities have not always used the same definitions for water losses. For example, firefighting and meter errors could be included in different categories, making comparisons difficult.

The International Water Association (IWA) and American Water Works Association (AWWA) established a common framework. It separates authorised use from apparent and real losses, so utilities can account for water consistently.

The balance also provides the annual real-loss volume used in the Infrastructure Leakage Index (ILI). ILI compares this with estimated unavoidable annual real losses, which requires network and pressure data. The EU Drinking Water Directive allows ILI or another appropriate method for assessing leakage.

The key point: a balance that adds up is not proof that its inputs are correct. In a top-down water balance, real losses are calculated as the remainder.

The table

The two main calculations are:

Water losses = System input volume − Authorised consumption

Real losses = Water losses − Apparent losses

• Revenue water is billed authorised consumption, whether or not the bill has been paid.
• Non-revenue water (NRW) is unbilled authorised consumption + apparent losses + real losses.

NRW includes leakage, but also authorised unbilled use and losses caused by metering or data errors. It should not be reported as leakage alone.

System input volume

System input volume is all water entering the defined system during the reporting period, including water from the utility’s own production and imports. For a distribution-system balance, use water entering the network after treatment, not raw water abstracted upstream.

Define the boundary first and account for every inlet. In the table above, billed metered exports are included in authorised consumption. Do not also subtract them from input. If you use a balance that excludes exports, adjust the boundary and calculations consistently.

Check bulk-meter accuracy and use the same period for all volumes. With other inputs unchanged, a 2% error in system input changes calculated real losses by 2% of the input volume.

Authorised consumption

Authorised consumption is water used with the utility’s permission. It has four components:

• Billed metered consumption: recorded customer use and billed metered exports.
• Billed unmetered consumption: estimated volumes billed without meter readings, for example under flat-rate arrangements.
• Unbilled metered consumption: measured use that is not billed, such as supplies to the utility’s own facilities.
• Unbilled unmetered consumption: estimated use for firefighting, network flushing, street cleaning or other permitted activities within the system boundary.

Document these estimates. Leaving unbilled use at zero without evidence can overstate real losses.

Apparent losses

Apparent losses are unauthorised use and consumption missed or incorrectly recorded by customer meters and data systems. The table groups them into two components:

• Unauthorised consumption: illegal connections, meter tampering and unauthorised hydrant use.
• Customer metering inaccuracies and data-handling errors: under-registration, incorrect readings, missing consumption records and unit-conversion errors.

Apparent losses can reduce billing revenue. They are generally valued using the applicable customer tariff, while real losses are generally valued using the cost of producing or purchasing water. Recoverable revenue depends on tariffs and whether the consumption can be billed.

Mechanical meters can under-register as they wear, particularly at low flow. Age alone does not establish the error: test a representative sample of the meter fleet.

Real losses

Real losses are water physically lost from the system:

• Transmission and distribution mains: bursts and background leakage from joints and fittings.
• Storage tanks: leakage and overflows.
• Service connections up to the customer meter: leakage before the metering point, including pipes on private land if the meter is inside the building.

This is an accounting boundary. It does not, by itself, determine who owns the pipe or must repair it.

How data errors affect real losses

In a top-down balance, real losses are calculated by subtraction:

Real losses = System input volume − Authorised consumption − Apparent losses

With all other inputs unchanged, the effect on the calculated result is:

• Under-count system input by 10,000 m³ → real losses fall by 10,000 m³.
• Over-count billed consumption by 10,000 m³ → real losses fall by 10,000 m³.
• Leave unbilled unmetered consumption at zero when it is really 15,000 m³ → real losses rise by 15,000 m³.
• Set meter inaccuracy to zero when the fleet under-reads by 30,000 m³ → real losses rise by 30,000 m³.

Errors in input volume, authorised consumption or apparent losses affect the calculated remainder. They can increase or decrease the estimate, and some may cancel out. A total that adds up does not detect these errors; negative losses or implausible trends can still flag a problem.

Check the data quality and compare the result with an independent leakage estimate before using it to plan work.

An independent check

Minimum night flow is one useful check in a metered zone. Demand is often lowest around 02:00–04:00, but the timing and level depend on local users.

Start with net zone inflow, accounting for outlets and changes in storage. Subtract customer night use, including households, industry and tank filling. Closed-loop circulation inside a building does not itself consume supply water. The remainder estimates leakage at night, not annual leakage volume.

To compare it with the water balance, convert it to a volume for the same area and period, allowing for pressure changes and changes in leakage over time. A large difference calls for checking both methods: meters, boundaries, night-use assumptions and pressure adjustments.

Data quality, component by component

Record the source and uncertainty of each volume. Measured values also need checks:

Weak estimates can leave substantial uncertainty in real losses. Report that uncertainty and prioritise improvements by their likely effect on the result, not just by the size of a component.

Where to start

Start with a clear system boundary and a common reporting period. Verify bulk meters, record unbilled use and test customer meters. Where practical, divide the network into metered zones so you can investigate differences locally.

Divako water loss management brings input readings, consumption data and network zones together. It helps validate readings, align periods and update calculations as data arrives. Estimates still need evidence and review.

Sources

• AWWA: IWA/AWWA water audit framework
• US EPA: Water-loss definitions and ILI, chapter 2.
• Directive (EU) 2020/2184: Article 4(3): ILI or another appropriate method.