Potential risks of nitrate pollution in aquifers from agricultural practices in the Nurra region, northwestern Sardinia, Italy

The paper describes the methodological and innovative approach, which aims to evaluate the potential risk of nitrate pollution in aquifers from agricultural practices by combining intrinsic aquifer vulnerability to contamination, according to the SINTACS R5 method, with agricultural nitrates hazard assessment, according to the IPNOA index. The proposed parametric model adopts a geographically based integrated evaluation system, comprising qualitative and semi-quantitative indicators. In some cases, the authors have modified this model, revising and adjusting scores and weights of the parameter to account for the different environmental conditions, and calibrating accordingly. The method has been successfully implemented and validated in the pilot area of the Alghero coastal plain (northwestern Sardinia, Italy) where aquifers with high productivity are present. The classes with a major score (high potential risk) are in the central part of the plain, in correspondence with the most productive aquifers, where most actual or potential pollution sources are concentrated. These are mainly represented by intensive agricultural activities, by industrial agglomerate and diffused urbanisation. For calibrating the model and optimizing and/or weighting the examined factors, the modelling results were validated by comparison with groundwater quality data, in particular nitrate content, and with the potential pollution sources census data. The parametric method is a popular approach to groundwater vulnerability assessment, in contrast to groundwater flow model and statistical method ones: it is, indeed, relatively inexpensive and straightforward, and use data commonly available or that can be estimated. The zoning of nitrate vulnerable areas provides regional authorities with a useful decision support tool for planning land-use properly managing groundwater and combating and/or mitigating desertification processes. However, a careful validation of the results is indispensable for reliable application.