Analyzing the effect of numerical solution method and empirical coefficients on the accuracy predictions of the ADE for BOD modeling in riverine basins

摘要

Advection-Dispersion Equation (ADE) in one-, two- or three-dimensional form is commonly used for dynamically modeling of water quality and sediment transport processes in riverine basins, estuaries and coastal waters. The numerical method used for solving this partial differential equation and the empirical coefficients which are as a main part of it are considerably effective on the model stability and accuracy. This paper shows how effective is the numerical method used for solving the ADE on the accuracy of model predictions of Biochemical Oxygen Demand (BOD) and compares the importance of these methods against the empirical coefficients involved in the ADE. The sensitivity level of model due to numerical solution and empirical coefficients wereanalyzed. It was found that using an appropriate experimental equation for the dispersion coefficient in the ADE, can significantly increase the accuracy of estimation of BOD concentrations, rather than using different numerical solutions. For example, comparison of the BOD predictions by the model with different numerical solutions with the corresponding measured values showed a maximum difference error of 6% between the best and the worst numerical solutions, whereas this value was about 40% when an inappropriate dispersion coefficient was used.