Search

In order to facilitate the monitoring of groundwater quality in France, the groundwater bodies (GWB) in the Provence-Alpes-Côte d'Azur region have been grouped into 11 homogeneous clusters on the basis of their physico-chemical and bacteriological characteristics. This study aims to test the legitimacy of this grouping by predicting whether water samples belong to a given sampling point, GWB or group of GWBs. To this end, 8673 observations and 18 parameters were extracted from the Size-Eaux database, and this dataset was processed using discriminant analysis and various machine learning algorithms. The results indicate an accuracy of 67% using linear discriminant analysis and 69 to 83% using ML algorithms, while quadratic discriminant analysis underperforms in comparison, yielding a less accurate prediction of 59%. The importance of each parameter in the prediction was assessed using an approach combining recursive feature elimination (RFE) techniques and random forest feature importance (RFFI). Major ions show high spatial range and play the main role in discrimination, while trace elements and bacteriological parameters of high local and/or temporal variability only play a minor role. The disparity of the results according to the characteristics of the GWB groups (geography, altitude, lithology, etc.) is discussed. Validating the grouping of GWBs will enable monitoring and surveillance strategies to be redirected on the basis of fewer, homogeneous hydrogeological units, in order to optimize sustainable management of the resource by the health agencies. [ABSTRACT FROM AUTHOR]

When studying large multiparametric databases with very heterogeneous parameters (microbiological, chemical, and physicochemical), covering a wide and heterogeneous area, the probability of observing extreme values (Z-score > 2.5) is high. The information carried by these few samples monopolizes a large part of the information conveyed by the entire database. The study of the spatial structure of the data and the identification of the mechanisms responsible for the water quality are then strongly degraded. Data transformation can be proposed to overcome these problems. This study deals with a database of 8110 groundwater analyses (Occitanie region, France), on which the bacteriological load was measured in Escherichia coli and Enterococci, in addition to electrical conductivity, major ions, Mn, Fe, As and pH. Three modes of data conditioning were tested and compared to the treatment with raw data. The results show that log transformation is the best option, revealing a relationship between E. coli content and all the other parameters. By reducing the impact of extreme values without eliminating them, it allowed a concentration of information on the first factorial axes of the PCA, and consequently a better definition of the associated processes. The spatial structure of the principal components and their cartographic representation is improved. The conditioning of the data with the square root function led to an intermediate improvement between the logarithmic transformation and the absence of conditioning. The application of these results should allow a targeted, more efficient, and therefore, less expensive monitoring of water quality by Regional Health Agencies. [ABSTRACT FROM AUTHOR]