A Bayesian approach to model the trends and variability in urban stormwater quality associated with catchment and hydrologic parameters. (1st June 2021)

• Record Type: Journal Article
• Authors: Perera, Thamali; McGree, James; Egodawatta, Prasanna; Jinadasa, K.B.S.N.; Goonetilleke, Ashantha
• Is Part Of: Water research. Volume 197(2021)
• Publication Date: 2021-06-01
• Subjects: stormwater runoff -- Bayesian hierarchical modelling -- uncertainty analysis -- stormwater quality -- stormwater pollutant processes; Water -- Pollution -- Research -- Periodicals; 363.7394
• DOI: 10.1016/j.watres.2021.117076
• Languages: English
• ISSNs: 0043-1354

Abstract

Highlights: Bayesian hierarchical model for runoff pollutograph is proposed The model combines catchment and event specific characteristics Rain intensity, duration, antecedent dry period can explain pollutant variability Effective impervious area of the catchment create high variability in pollutographs Abstract: Stormwater runoff pollution has become a key environmental issue in urban areas. Reliable estimation of stormwater pollutant discharge is important for implementing robust water quality management strategies. Even though significant attempts have been undertaken to develop water quality models, deterministic approaches have proven inappropriate as they do not address the variability in stormwater quality. Due to the random nature of rainfall characteristics and the differences in catchment characteristics, it is difficult to generate the runoff pollutographs to a desired level of certainty. Bayesian hierarchical modelling is an effective tool for developing complex models with a large number of sources of variability. A Bayesian model does not look for a single value of the model parameters, but rather determines a distribution of the model parameters from which all inference is drawn. This study introduces a Bayesian hierarchical linear regression model to describe a catchment specific runoff pollutograph incorporating the associated uncertainties in the model parameters. The model incorporates catchment and rainfall characteristics including the effective impervious area, time of concentration, rain duration, average rainfall intensity and the antecedent dry period as the contributors to random effects. Graphical abstract: Image, graphical abstract