Quelle surveillance de la qualité physico-chimique des cours d’eau ?

• A surveillance network comprises many measurement points located along waterways. Each point is a water sampling location. Each sample undergoes multiple physicochemical analyses by public or private laboratories. • Networks gradually developed from the 1970s onwards. They then evolved rapidly from 2006 onwards with the implementation of the Water Framework Directive (WFD) principles. Today, they are evenly distributed across the metropolitan area.

• Between 2000 and 2017, the number of measurement points per 1,000 km of waterway increased from 4 to 9 across France. Networks became denser in each river basin. The Artois-Picardie basin was an exception: its density remained relatively high throughout the period. From 2006 onwards, the upward trend intensified. This was the period when surveillance programmes were established following WFD requirements. In parallel, some water agencies maintained complementary networks to cover other parts of the territory and/or manage local situations not addressed in WFD network programmes. • Today, the highest densities are observed in the Seine-Normandy, Adour-Garonne, and Loire-Brittany basins.

The majority of measurement points are monitored for at most 3 years. Monitoring strategies are heterogeneous between basins: the Artois-Picardie basin monitors 80% of its points annually; Loire-Brittany, Rhône-Mediterranean-Corsica, and Seine-Normandy monitor around 60% over 1 to 3 years; Adour-Garonne and Rhine-Meuse have more balanced distributions of monitoring durations.

Across the metropolitan area, sampling frequency has remained around 9 samples/year throughout the period. The significant disparities between basins observed at the beginning of the period narrowed from 2006 onwards (from 5-18 samples/year in 2002 to 8-11 samples/year in 2017).

• Chemical measurements accounted for 92% of analyses during 2000-2004 (5.5 million analyses total). This increased to 98% during 2013-2017 (67.6 million analyses total). • The number of chemical parameters monitored between 2000 and 2017 increased across all pollutant groups (metals, organic compounds, pesticides, etc.). However, the rate of increase varied by basin. Currently, the "Pesticide" group has the widest range of monitored parameters. • These increases are linked to regulatory obligations introduced in European directives and to technical progress in metrology (e.g., multi-residue pesticide analysis enabling easier and less costly screening of multiple substances).

• In each basin, the majority of analysis results concern pesticides. The substances searched for depend on territorial characteristics and are not necessarily identical from one basin to another, except for a few dozen substances covered by common regulatory requirements. This is explained by differences in land use between basins (substances monitored in wine regions differ from those in cereal regions).

Analysis result quality must be controlled and qualified by its producer. The result is then distributed in the national Naïades database. This process aims to detect any anomalies in the result. A measurement is qualified as "correct" when its producer considers it to be neither erroneous nor doubtful. The proportion of "correct" measurements increased significantly from 2006 onwards and today stabilizes around 98%.

The performance of a chemical analysis method is assessed notably by its quantification limit. This limit is the smallest concentration of the analyte that can be quantified with sufficient reliability. Depending on the parameter considered, performance differences can be minimal (e.g., trichloromethane) or highly pronounced (e.g., metals). They are linked to the laboratories selected to carry out analytical services under public procurement contracts.