Abstract :
[en] Discriminating the different sources responsible for nitrate contamination of groundwater is a complex task. Indeed, the excessive presence of this pollutant can be the result of a combination of several elements (poor assimilation of inorganic fertilisers, decomposition of manure, losses in drainage systems). An excess of nitrate represents a danger both for the quality of drinking water supplies and for the health of ecosystems. Therefore, identifying the causes of this contamination is an important step in establishing an appropriate mitigation policy. Largely exploited for public water production (50 million m³/year), the Mesozoic chalk aquifer of the Mons basin (Belgium), covers an area of over 400km². As a substantial water reserve for Belgium, the quality of this resource is continuously monitored. Over the last years, an increase in nitrate concentration has been observed in several monitoring wells and is progressively threatening the sustainability of some production sites. The land-uses in the area are various including fields, pastures, urban areas and industrial sites. It is, therefore, difficult to identify the origin of nitrate and mitigate the pollution. Commissioned by the SPGE (Société de Protection et de Gestion de l’Eau, Belgium) and conducted by the universities of Mons and Liège, the characterisation of the pollution and associated nitrate sources is carried out through multiple sampling campaigns covering the different land use zones and confined/unconfined areas. Classical hydrochemical analyses are performed to define the extent of the nitrate pollution, to locate potential denitrification zones and to highlight correlations with other major ions. In parallel, analyses of the stable isotopes of nitrate (δ15N and δ 18O) and boron (δ 11B) are carried out. These isotopes allow to differentiate different sources of nitrate, including mineral or organic fertilisers, household waste degradation in landfills and possible leakage from sewer systems in urban areas and also to highlight denitrification processes. The results of the sampling campaigns are interpreted using a combination of classical tools (maps of the ions spatial distribution, Piper diagrams, binary graphs) associated with multi-criteria analysis algorithms (PCA, T-SNE, SOM’S). These methods, combined with isotopic measurements, make it possible to delimit multiple sub-zones in the basin according to suspected nitrate sources. Finally, the presence of a denitrification front along the interface between confined and unconfined aquifer is also highlighted.
