Surface water quality assessment in a semiarid Mediterranean region (Medjerda, Northern Tunisia) using partial triadic analysis

A range of methods have been developed specifically to analyze several tables of data simultaneously (variable × space × time) in the field of ecological research, although they have been less widely used to examine water quality. In this study, we assessed the spatiotemporal variability of water quality in the Medjerda River basin (Northern Tunisia). Partial triadic analysis (PTA) provides an effective framework for the assessment of spatiotemporal variability of water quality in the Medjerda River basin (Northern Tunisia). Fourteen physicochemical variables were monitored from 12 sampling sites monthly during 2013. PTA allowed correlations among different physicochemical parameters to be identified and to assess overall water quality in the Medjerda River. Salinity (S), Cl−, SO42−, Ca2+, and Mg2+ ions were associated with intensive agricultural activities (agricultural pollution sources) leading to salinization. However, NH4+, PO43−, chemical oxygen demand (COD), and biochemical oxygen demand (BOD5) we more strongly associated with polluted urban sites. PTA helped illustrate that strong links exist between land uses and adjacent water quality. The advantages of this multi-table method approach for water quality monitoring include as follows: (1) identifying common multivariate spatial structures and problems associated with maintaining water quality, (2) allowing identification of consistent patterns in water chemistry, and (3) allowing analysis on the temporal variability of water chemistry.