Water & Compliance · Technical guide 9 min read Published 24 August 2026

District metering area design: how to zone a network so the losses have nowhere to hide

By Frank Guo · Technology & Product Leadership, addanode

TL;DR — A district metered area (DMA) is a hydraulically discrete zone of your network with every inflow metered — the unit at which water losses become measurable, rankable and fixable. Good design follows a few rules: size zones at roughly 500–3,000 connections (big enough to matter, small enough to localise), draw boundaries where the network already wants them (reservoir zones, pressure zones, trunk offtakes), meter the few inlets rather than the many customers, and close boundary valves you can prove stay closed. Then let minimum night flow rank zones and pressure management hold the gains. The classic failures are all design failures: leaky boundaries, zones too big to act on, meters installed but never telemetered, and programmes that measure for a year without fixing anything.

What a DMA is, and why it's the unit of loss reduction

A network-wide NRW percentage is unactionable — you can't send a crew to "43%". A DMA converts the network into zones where a simple balance holds: metered inflow − legitimate consumption = losses in this zone. With the inlet telemetered, that balance updates daily, and the night-flow signal (below) updates nightly. Loss reduction then becomes an ordered worklist: worst zone, fix, re-measure, next zone — the operating pattern behind our NRW solution.

Design rule 1: size for action, not elegance

International practice converges on roughly 500–3,000 connections per DMA. Below ~500, metering cost per connection climbs and night flow gets noisy (one farm filling a tank distorts everything). Above ~3,000, a "found" leak still hides among tens of kilometres of main, and the pinpointing work (see leak detection equipment) balloons. South African reality check: in dense townships, err smaller (illegal-connection dynamics localise better); across rural schemes, a "DMA" may simply be one village fed by one bulk line — perfect, meter it.

Design rule 2: let the network draw the boundaries

Reservoir and tower supply zones are natural DMAs — one source, gravity-defined edges, often needing zero new valves.
Pressure zones already have closed boundaries; add a meter and they're DMAs.
Trunk-main offtakes define zones for the suburbs they feed — meter at the offtake.
Avoid boundaries that need many closed valves. Every closed boundary valve is a future leak in your accounting: valves get opened during repairs and never closed, quietly cross-feeding zones. Fewer boundary valves = a DMA that stays discrete. Where valves are unavoidable, log them and re-verify with periodic zero-pressure tests.

Design rule 3: meter the inlets, telemeter the meters

The economics of DMA design are gentle: a zone of 1,500 connections usually needs one or two inlet meters — ultrasonic or mechanical-with-pulse, sized for the realistic flow range (oversized meters under-read the night flows you care about most; sizing to peak demand is the classic spec error). But a meter read monthly by a walker defeats the purpose: night flow is invisible, bursts show up weeks late. Telemetry — logging at 15-minute intervals or finer, reporting over 4G/LoRaWAN, buffered through load shedding — is what turns installed meters into a working programme. Pressure loggers at the inlet and the zone's critical point complete the instrumentation.

Minimum night flow: the number the whole design serves

Between roughly 2am and 4am, legitimate use collapses, so a zone's inflow approximates its losses. The design implications: log finely enough to see the trough; correct for genuine night users (hospitals, 24-hour plants — meter them separately or account explicitly); and trend MNF per zone over weeks, because the change is the alarm — a step up means a new burst or connection, a slow climb means background leakage growing. Dividing MNF by the zone's connection count normalises the league table so small zones compete fairly with big ones.

The interval in the parameters table below, measured against the window it has to resolve. Minimum night flow is a two-hour question — hourly sampling puts two readings inside it, and the minimum falls between them more often than not.

Pressure management: the follow-through

Once zoning exposes the worst areas, pressure management inside them multiplies the win: leakage scales strongly with pressure, so smoothing night-time excess (when demand is low and pressure spikes) cuts background losses and burst frequency without a single excavation — the mechanics are in our pressure management guide. A DMA with inlet metering, critical-point pressure logging and a controlled valve is the full toolkit: measure, reduce, hold.

DMA design parameters — typical starting points

Every network is different, but the ranges below are where working DMAs in South African municipalities tend to land; treat them as the first draft, and let the hydraulics and the crew capacity argue you off them.

Parameter	Typical starting point	Why it matters
Connections per DMA	500–3,000 (smaller in dense, high-loss areas)	Small enough that a night-flow step points a crew at a walkable area
Metered inlets	1 (ideally); 2 with both metered	Every unmetered inlet is a hole in the balance
Boundary valves	Closed, tagged, and tested by zero-pressure or step tests	One passing valve silently merges two zones
Inlet meter type	Electromagnetic or ultrasonic, sized for night flow, not peak	Mechanical bulk meters under-read exactly the low flows MNF depends on
Logging interval	15 min routine; 1–5 min during step tests	Night-flow minimum falls between hourly samples
MNF window	02:00–04:00, calculated nightly	The one number the whole design serves
Pressure logging	Inlet + critical (highest/farthest) point	Separates a leakage rise from a pressure rise
Legitimate night use allowance	Estimated per connection class and subtracted	Otherwise every zone looks like it leaks

The five classic DMA failures (all avoidable at design time)

Leaky boundaries — unverified valves cross-feeding zones, corrupting every balance. Verify at commissioning; re-verify annually.
Zones sized for the map, not the crews — beautiful GIS polygons too large to act on. Size to what a leak team can sweep.
Meters without telemetry — the programme dies in the reading backlog. Budget telemetry as part of the meter, not an option.
Data without action — a year of league tables and no repairs. Pair every measurement milestone with a repair commitment (the municipal NRW case study shows the rhythm).
No re-measurement — fixes unproven, recoveries unclaimed, budgets unrenewed. The after-measurement is the political capital of the whole programme.

Frequently asked questions

What is a district metered area (DMA)?

A discrete section of a distribution network — typically 500–3,000 connections — whose inflows are all metered, so a water balance can be computed for the zone alone. It's the standard international unit for locating and managing water losses, because it turns one network-wide percentage into comparable, actionable zones.

How many connections should a DMA have?

Convention says 500–3,000. Go smaller in dense or high-loss areas (faster localisation), larger where the network is sound and metering budgets are tight. The honest test: when this zone's night flow jumps, is the resulting search area small enough that your crew can actually sweep it?

What is minimum night flow and what's a "good" value?

The lowest stable inflow to a zone in the small hours (≈2–4am), when legitimate demand is minimal — so it approximates the zone's leakage plus unauthorised use. Absolute benchmarks vary with zone size and pressure; the operational answer is per-connection MNF compared across your own zones, and trended: the change is the alarm.

Can DMAs work with intermittent supply?

Yes, with adapted analysis: instead of classic night flow, use charge-window behaviour (how fast the zone draws when supply opens) and overnight tank decay. Telemetered inlets are even more important under rationing, because manual monthly totals can't compare zones with different supply hours fairly.

What does it cost to set up a DMA?

Typically one or two telemetered inlet meters, a pressure logger or two, and boundary-valve verification — modest against what the zone loses annually (see our metering cost guide for project-level numbers). The worst-first rollout pattern lets recovered water fund subsequent zones.

Primary sources

Zone the network. Rank the losses. Fix in order.

Send us your network map — we'll sketch the natural DMA boundaries, the metering points, and a worst-zone-first rollout priced for your budget.

Talk to our Water Team NRW solution →