An innovative procedure to assess multi-scale temporal trends in groundwater quality: Example of the nitrate in the Seine–Normandy basin, France

Abstract

• Large and heterogeneous dataset: how to improve the assessment of trends. • Punctual and regional assessment of groundwater quality. • Homogeneity and nonparametric tests to determine recent groundwater quality trends. • We modify geostatistical approach to create spatial unit for regional trend analysis. • Improving the management of groundwater in compliance with the WFD requirements. The European Water Framework Directive (WFD) asks Member States to identify trends in contaminant concentrations in groundwater and to take measures to reach a good chemical status by 2015. In this study, carried out in a large hydrological basin (95,300 km 2 ), an innovative procedure is described for the assessment of recent trends in groundwater nitrate concentrations both at sampling point and regional scales. Temporal variograms of piezometric and nitrate concentration time series are automatically calculated and fitted in order to classify groundwater according to their temporal pattern. These results are then coupled with aquifer lithology to map spatial units within which the modes of diffuse transport of contaminants towards groundwater are assumed to be the same at all points. These spatial units are suitable for evaluating regional trends. The stability over time of the time series is tested based on the cumulative sum principle, to determine the time period during which the trend should be sought. The Mann–Kendall and Regional-Kendall nonparametric tests for monotonic trends, coupled with the Sen-slope test, are applied to the periods following the point breaks thus determined at both the sampling point or regional scales. This novel procedure is robust and enables rapid processing of large databases of raw data. It would therefore be useful for managing groundwater quality in compliance with the aims of the WFD.