Assessment of hot groundwater in an arid area in Tunisia using geochemical and fuzzy logic approaches

In this study, thermal groundwater from arid area in southeastern Tunisia was assessed for irrigation use. For this purpose, thirty-one water samples were collected and physiochemical parameters (EC, pH, TDS, major ions) were measured and analyzed. A fuzzy logic model was developed in which six parameters were integrated: electrical conductivity, sodium adsorption ratio, sodium percentage, Kelly ratio, permeability index and temperature. The membership functions for a fuzzy logic model were developed using linguistic terms and trapezoidal shapes. The fuzzy logic model developed was validated with a dataset of chemical analyses from groundwater sampled in the study area. The assessment indicated that 26% of the samples were in the “good” class, 10% in the “good to permissible” class, 55% are in the “permissible” class, 6% are in the “permissible to harmful” class and 3% were considered to be harmful and therefore unsuitable for use in irrigation. The effectiveness, simplicity and robustness of the fuzzy model assessment make this approach a more consistent and reliable way of assessing water quality than conventional methods of assessing water quality data.