Current Status of Emerging Contaminant Models and Their Applications Concerning the Aquatic Environment: A Review

Abstract

Increasing numbers of emerging contaminants (ECs) detected in water environments require a detailed understanding of these chemicals’ fate, distribution, transport, and risk in aquatic ecosystems. Modeling is a useful approach for determining ECs’ characteristics and their behaviors in aquatic environments. This article proposes a systematic taxonomy of EC models and addresses gaps in the comprehensive analysis of EC models and their applications. The reviewed models include conventional water quality models, multimedia fugacity models, and machine learning (ML) models. Conventional water quality models have higher prediction accuracy and spatial resolution; nevertheless, they are limited in functionality and can only be used to predict contaminant concentrations in aquatic environments. Fugacity models are excellent at depicting how contaminants travel between different environmental media, but they cannot be used directly to analyze contaminant variations in different parts of the same environmental media because the fugacity model assumes that contaminant concentrations are constant within the same environmental compartment. Compared to other models, ML models can be applied to more scenarios, such as contaminant identification and risk assessments, rather than being confined to the prediction of contaminant concentrations. In recent years, with the rapid development of artificial intelligence, ML models have surpassed fugacity models and conventional water quality models, becoming one of the newest hotspots in the study of ECs. The primary challenge faced by ML models is that the model outcomes are difficult to interpret and understand, and this influences the practical value of an ML model to some extent.