Deciphering Groundwater Quality Mechanisms in the Rhône-Mediterranean-Corsica Basin (RMC) Through Multi-Source Data Integration

In the Rhône-Mediterranean-Corsica (RMC) basin (130,000 km2, 14 million inhabitants), groundwater intended for human consumption has been monitored for decades. These data, stored in the SISE-EAUX database, were cross-referenced with information from the CORINE Land Cover (CLC) database, which describes human land use, in order to identify potential relationships between pollutant pressure and water quality at the basin scale, as well as the mechanisms specific to each geographical area. Data processing was carried out in three stages. The 27,741 water samples from 2825 abstraction points were assigned to the 224 groundwater bodies (GWBs), and average values for each physicochemical and bacteriological parameter were calculated for each GWB. At the same time, the percentage of surface area covered by each land use type was also extracted at the scale of each GWB. This information was subjected to statistical processing, first separately and then jointly, using principal component analysis (PCA) and hierarchical clustering of parameters. A redundancy in the information carried by the quality parameters, previously observed at the scale of administrative regions (four to five times smaller), is confirmed at this new analysis scale, paving the way for data consolidation and a more synthetic representation. Fecal contamination primarily concerns areas with crystalline lithology and, secondarily, a few karst sectors, with other livestock farming regions being less contaminated. Higher nitrate concentrations are observed in cereal-growing regions and areas of intensive row cropping, while metal concentrations are lower in the drier Mediterranean climate zone than under the more humid continental climate. Structuring factors, notably altitude and climate, emerge at the RMC basin analysis scale, which was not the case at the scale of administrative regions. These structuring factors influence land use, soil type, and hydrological regimes alike, which explains the correlations between the information contained in the CLC and SISE-EAUX databases.