Abstract
The complexity of pollutant-mixing mechanism in open channels generates large uncertainty in estimation of longitudinal and lateral dispersion coefficients (Kx and Ky). Therefore, Kx and Ky estimation in rivers should be accompanied by an uncertainty analysis, a subject mainly ignored in previous studies. We introduce a method based on thorough analysis of different calibration datasets, resampled from a global database of tracer studies, to determine the uncertainty associated with five applicable intelligent models for estimation of Kx and Ky (model tree, evolutionary polynomial regression (EPR), gene-expression programming, multivariate adaptive regression splines (MARS), and support vector machine (SVM)). Our findings suggest that SVM gives least uncertainty in both Kx and Ky estimation, while EPR and MARS generate most uncertainty in Kx and Ky estimation, respectively. By considering significant uncertainty in the model estimations, we suggest that the methodology we introduce here for uncertainty determination of the models be incorporated in empirical studies on estimation of Kx and Ky in rivers.
Highlights
Driven pollutants discharged in rivers are a main concern that threaten freshwater supply for human and aquatic life (Bostanmaneshrad et al, 2018; Noori et al, 2019)
We aimed to determine the uncertainty associated with model tree (MTree), evolutionary polynomial regression (EPR), GEP, multivariate adaptive regression splines (MARS), and support vector machine (SVM) models for estimation of Kx and Ky in channels
Our study showed considerable uncertainty associated with these models, which are commonly used for estimation of Kx and Ky
Summary
Driven pollutants discharged in rivers are a main concern that threaten freshwater supply for human and aquatic life (Bostanmaneshrad et al, 2018; Noori et al, 2019). Vertical mixing in rivers usually occurs quickly near the contamination field from the pollutant discharge point, while lateral and longitudinal mixings take place in intermediate and distant fields from the point of pollutant release into rivers, respectively. Since Kx and Ky values estimated by these models are used in combination with two-dimensional (2D) numerical models in water quality studies, the uncertainty in MLM results induced by variation in the calibration dataset obtained from tracer studies needs to be better investigated. With respect to management practices, the more certain in the results of MLMs in the process of Kx and Ky estimation, the more confidence in the characterization of pollutant transport in rivers. Water quality management decisions based on poor Kx and Ky estimation could have detrimental effects on aquatic life in rivers and harm human health (Ramezani et al, 2019)
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call