In coastal aquifers, the seawater intrusion can mask the effects of high salinity regional flows, connate waters mobilization or contaminant process. Therefore, to discriminate between all the processes that have taken place in the coastal aquifer, is a complex task. Normally, traditional hydrogeochemical methods (e.g., Piper and Durov) together with statistical multivariate techniques (e.g., cluster and factorial analysis) and other methods (e.g., ionic deltas and isotopic studies) have been used to understand the hydrogeochemistry of aquifers and to confirm previous hypothesis.
This paper presents a characterization of the salinization process in coastal aquifers, by means a fuzzy logic and data mining based methodology, which has not been used before for this purpose in this environmental area. The proposed fuzzy methodology is based on the use of the data mining computer tool Predictive Fuzzy Rules Generator (PreFuRGe).
The results have been obtained by processing groundwater samples analyses with PreFuRGe. The parameters used for the experimentation have been: temperature, electric conductivity, redox potential, total dissolved solids, silicon dioxide, oxidability, major ions (chloride, sulfate, bicarbonate, nitrate, calcium, magnesium, sodium and potassium), and minor ions (arsenic, bromide, lithium, boron, strontium, chromium and fluoride).
The application of this method has made it possible to differentiate several overlapping hydrogeochemical processes, such as seawater intrusion, the entry of regional groundwater flows with high concentrations of strontium, magnesium, lithium and sulfates, and the effect of contamination from agricultural activities, with the presence of nitrates.
The qualitative obtained results in this paper have been compared to previous researches carried out in the same environmental area, and it is proved that the used fuzzy methodology is a powerful tool for discriminating between overlapping geogenic and anthropogenic processes in coastal aquifers.