Chemometric modeling for spatiotemporal characterization and self-depuration monitoring of surface water assessing the pollution sources impact of northern Argentina rivers

Abstract

In Argentina, both surface and ground water are used for a diverse priority purposes, such as drinking and basic hygiene, but they are also utilized as receivers of different types of industrial and urban and suburban effluents that affect their natural composition. This activity accompanied by the increase of the population and climate changes have activated the alarms of organism water management forced to implement strict quality controls previous to its use. In this work, a systematic evaluation of a set of physicochemical and biological parameters measured in 19 sampling sites during the period 2017–2019 is presented. Principal component analysis (PCA) and matrix augmentation-PCA (MA-PCA) were applied as exploratory analysis tools to visualize and interpret the information contained in the dataset. Both studies allowed to detect the relevant variables and to differentiate the samples based on pollution areas. These models led to similar conclusions; nonetheless, MA-PCA provided a more straightforward overview of the spatiotemporal variation of the samples in comparison to classical PCA. Finally, a significant and sensitive discriminant model (93% non-error rate) was developed to analyze and predict the self-depuration of the rivers. The excellent predictive ability achieved by this model makes its application suitable for the monitoring of the water quality.