Main content area

A survey on river water quality modelling using artificial intelligence models: 2000–2020

Tiyasha,, Tiyasha,, Tiyasha,, Tiyasha,, Tiyasha,, Tiyasha,, Tiyasha,, Tung, Tran Minh, Yaseen, Zaher Mundher
Journal of hydrology 2020 v.585 pp. 124670
anthropogenic activities, artificial intelligence, automation, climate change, cost effectiveness, decision making, early warning systems, funding, hydrologic models, issues and policy, monitoring, pollutants, prediction, problem solving, risk assessment, river water, rivers, surveys, time series analysis, water quality, watersheds
There has been an unsettling rise in the river contamination due to the climate change and anthropogenic activities. Last decades’ research has immensely focussed on river basin water quality (WQ) prediction, risk assessment and pollutant classification techniques to design more potent management policies and advanced early warning system. The next challenge is dealing with water-related data as they are problematic to handle owing to their nonlinearity, nonstationary feature and vague properties due to the unpredictable natural changes, interdependent relationship, human interference and complexity. Artificial intelligence (AI) models have shown remarkable success and superiority to handle such data owing to their higher accuracy to deal with non-linear data, robustness, reliability, cost-effectiveness, problem-solving capability, decision-making capability, efficiency and effectiveness. AI models are the perfect tools for river WQ monitoring, management, sustainability and policymaking. This research reports the state of the art of various AI models implemented for river WQ simulation over the past two decades (2000–2020). Correspondingly, over 200 research articles are reviewed from the Web of Science journals. The survey covers the model structure, input variability, performance metrics, regional generalisation investigation and comprehensive assessments of AI models progress in river water quality research. The increasing contaminants, the lack of funding and the deficiency in data, numerous variables and unique data time series pattern based on the geological area have increased the need for river WQ monitoring and control even more. Hence, this is highly emphasising the involvement of AI models development which can deal with missing data, able to integrate the features of a black-box model and white-box models, benchmarked model and automated early warning system are few of many points need more research. Despite extensive research on WQ simulation using AI models, shortcomings remain according to the current survey, and several possible future research directions are proposed. Overall, this survey provides a new milestone in water resource engineering on the AI model implementation, innovation and transformation in surface WQ modelling with many formidable problems in different blossoming area and objectives to be achieved in the future.