Hydrionis benchmarks

Hydrionis Leakage Benchmark — France

1.83 m³/km/day

Median, 2019

0.40

Best held

9.30

Worst held

Subjects ranked

Volume lost from the distribution network per kilometre of main per day. Version 1.0.0.

Hydrionis Leakage Benchmark — France: median over time

The median of every eligible subject in each year Hydrionis can build the benchmark for

The set of reporting subjects changes between years, so a movement in the median can reflect a change in who reported as well as a change in performance.

View the data behind this chart

Underlying data
Period	Median across ranked subjects (m³/km/day)
2008	2.34
2009	3.00
2010	2.49
2011	2.66
2012	2.53
2013	2.49
2014	2.29
2015	2.46
2016	2.16
2017	2.25
2018	2.25
2019	1.83

Lowest leakage (2019)

Best-performing subjects in Hydrionis Leakage Benchmark — France
Rank	Subject	m³/km/day
1	Haute-Vienne	0.40
2	Lozère	0.60
3	Corrèze	0.65
4	Loire	0.65
5	Creuse	0.70
6	Tarn-Et-Garonne	0.70
7	Gers	0.80
8	Haute-Loire	0.80
9	Morbihan	0.80
10	Finistère	0.85

Highest leakage

Worst-performing subjects in Hydrionis Leakage Benchmark — France
Rank	Subject	m³/km/day
68	Bouches-Du-Rhône	9.30
67	Aveyron	8.90
66	Aude	5.80
65	Nord	5.70
64	Isère	5.30

Disclosure

Method, coverage and rights disclosure for Hydrionis Leakage Benchmark — France
Metric	Linear leakage index (m³/km/day)
Definition	Volume lost from the distribution network per kilometre of main per day.
Period	2019
Jurisdiction	France — departments, metropolitan and overseas
Why these are comparable	Every value is the French regulatory indicator P106.3, computed under one national method and defined in French regulation, aggregated by Hydrionis to department as the median across the water services that reported it. Because the definition, the method and the reporting year are identical across every subject ranked, the departments are comparable with one another.
Comparability warnings	A French regulatory indicator (P106.3), normalised by network length. The English and Welsh leakage-per-property measure is normalised by connections instead, so the two rank networks differently and cannot be placed on one scale.
Missing data	A department with no reported value for the year is absent from the ranking. No value is imputed, carried forward from an adjacent year, or estimated from neighbouring departments, and the ranking never presents itself as covering all of France.
Coverage	68 subjects ranked, 1 values excluded as outside the jurisdiction or duplicated. Benchmark available for 12 years.
Method	Hydrionis Leakage Benchmark — France v1.0.0. Subjects ranked on Linear leakage index for 2019-01-01; lower is better. No weighting, no normalisation and no imputation are applied.
Sources	Hub'Eau / SISPEA — Indicateurs des services — service performance (Licence Ouverte / Open Licence v2.0 (Etalab))

What this benchmark does not establish

Valid within France only. The English and Welsh leakage measures are computed under different rules from a night-flow water balance and are normalised by connected properties rather than by network length; they are held as separate metrics and cannot be placed on this scale.
The subject is a department, not a water utility. A department containing one poorly performing large service and many good small ones is ranked on the median of its services, so the rank describes the typical network in that department and is never a judgement about any single operator.
The linear leakage index rewards dense networks: the same volume lost over fewer kilometres of main gives a worse score. It is the fairest of the available normalisations, not a neutral one.
Each department value is a median across its reporting services, unweighted, because the observatory publishes no served-population figure through this API. A department of many small rural services is therefore described mostly by small rural services.
Underlying returns are voluntary in practice, and coverage varies by department and year. The count of services reporting is published as its own series and should be read alongside any rank.
A leakage figure describes the network, not the water resource. A department with high leakage and abundant water may face less pressure than one with low leakage and none.

Hydrionis Leakage Benchmark — England

37.83 l/person/day

Median, 2024

26.14

Best held

53.59

Worst held

Subjects ranked

Leakage divided by the resident population the company supplies, expressed per person per day. Version 1.0.0.

Lowest leakage (2024)

Best-performing subjects in Hydrionis Leakage Benchmark — England
Rank	Subject	l/person/day
1	entity-utility-essex-and-suffolk-water	26.14
2	entity-utility-ses-water	26.52
3	entity-utility-cambridge-water	27.82
4	entity-utility-bristol-water	28.83
5	entity-utility-southern-water	36.07
6	entity-utility-anglian-water	36.36
7	entity-utility-severn-trent	36.57
8	entity-utility-affinity-water	36.64
9	entity-utility-portsmouth-water	37.83
10	entity-utility-south-staffs-water	41.54

Highest leakage

Worst-performing subjects in Hydrionis Leakage Benchmark — England
Rank	Subject	l/person/day
17	entity-utility-thames-water	53.59
16	entity-utility-united-utilities	50.62
15	entity-utility-wessex-water	49.15
14	entity-utility-yorkshire-water	47.05
13	entity-utility-south-west-water	46.78

Disclosure

Method, coverage and rights disclosure for Hydrionis Leakage Benchmark — England
Metric	Leakage per person (l/person/day)
Definition	Leakage divided by the resident population the company supplies, expressed per person per day.
Period	2024
Jurisdiction	England — appointed water companies
Why these are comparable	Every value is reported by an appointed water company to the Environment Agency for the same April-to-March regulatory year, from the company-reported water balance, under one regulator’s definition. Definition, method, unit and period are identical across every company ranked.
Comparability warnings	Normalising by population rather than by property shifts what the figure reflects: a company serving many small households will look worse per property and better per person than one serving few large ones. It is a different ranking of the same losses, not a better one.
Missing data	A company that did not report for the year is absent. Nothing is imputed, carried forward or estimated from a neighbouring company, and the ranking never presents itself as covering companies it does not hold.
Coverage	17 subjects ranked, 1 values excluded as outside the jurisdiction or duplicated. Benchmark available for 2 years.
Method	Hydrionis Leakage Benchmark — England v1.0.0. Subjects ranked on Leakage per person for 2024-04-01; lower is better. No weighting, no normalisation and no imputation are applied.
Sources	EA water resources — Analysis of the water industry’s annual water resources performance (Open Government Licence v3.0)

What this benchmark does not establish

Valid within England only. The French linear leakage index normalises by network length rather than by person, and international non-revenue water counts apparent losses and unbilled authorised use as well as physical leakage. Neither can be placed on this scale.
Normalising by population rewards dense supply areas: the same losses spread over more people give a better score. It is one defensible normalisation of several, and the same companies rank differently per property.
Leakage reflects pipe age, material, pressure, ground conditions and network density as much as management. This ranks circumstances alongside performance and should not be read as a judgement of either.
Two regulatory years are held, so a company’s position here is a position and not a trend. A single year-on-year movement can be weather, a revised water balance or a change in what the regulator asks for.
Ofwat holds a leakage series running from 1992 which would put any single year in context. It is absent because ofwat.gov.uk refuses automated requests, not because its licence forbids reuse.

Hydrionis Drinking Water Nitrate Benchmark — France

0.00 %

Median, 2025

0.00

Best held

13.22

Worst held

101

Subjects ranked

Percentage of drinking-water samples analysed for nitrate, expressed as the nitrate ion whose result exceeded the regulatory limit of 50 mg/l as NO₃. Version 1.0.0.

Hydrionis Drinking Water Nitrate Benchmark — France: median over time

The median of every eligible subject in each year Hydrionis can build the benchmark for

The set of reporting subjects changes between years, so a movement in the median can reflect a change in who reported as well as a change in performance.

View the data behind this chart

Underlying data
Period	Median across ranked subjects (%)
2022	0
2023	0
2024	0
2025	0

Lowest leakage (2025)

Best-performing subjects in Hydrionis Drinking Water Nitrate Benchmark — France
Rank	Subject	%
1	Ain	0.00
2	Hautes-Alpes	0.00
3	Alpes-Maritimes	0.00
4	Ardèche	0.00
5	Ariège	0.00
6	Aude	0.00
7	Bouches-Du-Rhône	0.00
8	Calvados	0.00
9	Cantal	0.00
10	Charente	0.00

Highest leakage

Worst-performing subjects in Hydrionis Drinking Water Nitrate Benchmark — France
Rank	Subject	%
101	Oise	13.22
100	Somme	11.03
99	Aube	10.35
98	Moselle	8.51
97	Marne	7.10

Disclosure

Method, coverage and rights disclosure for Hydrionis Drinking Water Nitrate Benchmark — France
Metric	Nitrate limit exceedance rate (%)
Definition	Percentage of drinking-water samples analysed for nitrate, expressed as the nitrate ion whose result exceeded the regulatory limit of 50 mg/l as NO₃.
Period	2025
Jurisdiction	France — departments, metropolitan and overseas
Why these are comparable	Every value is computed by Hydrionis from per-sample results published by the regional health agencies under one statutory control programme, against the single national limit of 50 mg/l, for the same calendar year. The determinand, the limit, the method and the period are identical across every department ranked.
Comparability warnings	Computed against the French regulatory limit of 50 mg/l as NO₃. Limits for the same substance differ between jurisdictions, so exceedance rates are not comparable internationally without checking both standards.
Missing data	A department with fewer than thirty results in the year is absent rather than ranked on a handful of samples. Nothing is imputed or carried forward.
Coverage	101 subjects ranked. Benchmark available for 4 years.
Method	Hydrionis Drinking Water Nitrate Benchmark — France v1.0.0. Subjects ranked on Nitrate limit exceedance rate for 2025-01-01; lower is better. No weighting, no normalisation and no imputation are applied.
Sources	Hub'Eau / SISE-Eaux — Qualité de l'eau potable — drinking water (Licence Ouverte / Open Licence v2.0 (Etalab))

What this benchmark does not establish

This is a rate of samples exceeding the limit, not a rate of people exposed. Sampling is risk-directed: a department that samples its worst supplies more often will look worse than one that does not, and an exceedance in a small rural supply counts the same as one serving a city.
Results below the laboratory detection limit are censored and excluded from every mean rather than substituted with zero, so a department’s figure is an upper bound on what was measurable rather than an estimate of what was present.
Valid within France only. Exceedance rates elsewhere are computed against different limits, different sampling regimes and different aggregation units.
A department is not a supply zone. Water quality varies within a department by source and treatment works, and a department-level rate describes the typical sampled supply rather than any particular tap.