Could we estimate industrial wastewater flows composition using the UN-ISIC classification system?

Although we have extensive datasets on the location and typology of industries, we do not know much on their generated and discharged wastewater. This lack of information compromises the achievement of the sustainable development goals focused on water (Sustainable Development Goal 6) in Europe and globally. Thus, our goal was to assess to which degree the chemical composition of industrial wastewater could be estimated based on the industry's typology according to its International Standard Industrial Classification of All Economic Activities (ISIC) class. We collected wastewater effluent water samples from 60 industrial wastewater effluents (before any wastewater treatment process), accounting for 5 samples each of 12 ISIC classes, analyzed the composition of key contaminants (i.e. European Commission rated priority compounds and watchlist), and statistically assessed the similarities and differences amongst ISIC classes using ordination and random forest analyses. The results showed statistically significant linkages between most ISIC classes and the composition of produced wastewater. Among the analytical parameters measured, the random forest methodology allowed identifying a sub-set particularly relevant for classification or eventual contamination prediction based on ISIC class. This is an important applied research topic with strong management implications to (i) determine pollution emission caps for each individual ISIC class, (ii) define monitoring schemes to sample and analyze industrial wastewater, and (iii) enable predicting pollutant loads discharged in river basins with scarce information. These encouraging results urge us to expand our work into other ISIC classes and water quality parameters to draw a full picture of the relationship between ISIC classes and produced wastewater.