Apply online to any of UB's Civil, Structural and Environmental Engineering Graduate programs
Development of a framework to account for uncertainty sources in modeling toxic concentrations in the Niagara River
The Niagara River is an important natural resource for drinking water and inexpensive hydroelectric power production. However, the quality of its water is less than pristine. Past lax management of toxic waste has left behind contaminants that are slowly seeping into the river. These loadings along with incoming concentrations from Lake Erie are just two of the contaminant sources to the Niagara River. The over all scope of this study it to model the fate and transport of selected contaminants under uncertainty. The work herein presents a framework to quantify the overall variability of the model estimations of toxic concentrations based on the uncertainty of several model parameters and the natural variability embedded in some of the model input variables. The results of the uncertainty analysis are used to understand the importance of stochastic model components and their effect on the overall reliability of the system, to evaluate multiple sources of uncertainty that might need to be further studied, to quantify in a probabilistic sense the ability of the natural system to meet pre-established standards, and to propose improvements to current water quality practices. The uncertainty analysis is performed using two Point Estimate Method (PEM) - the Rosenblueth and the Modified Rosenblueth methods. The water quality along the Niagara River is simulated by coupling two numerical models the Environmental Fluid Dynamic Code (EFDC)- for the hydrodynamic portion of the study and the Water Quality Analysis and Simulation program (WASP)- for the fate and transport of contaminants. Three priority toxic contaminants are analyzed: Total PCBs, fluoranthene and arsenic. This work is also one of the first attempts to introduce the Hilbert-Huang Transform (HHT) method to the environmental engineering field. The HHT method is used for the analysis and prediction of nonstationary and nonlinear time series of the contaminant concentrations. This research has brought significant contributions to modeling the fate and transport of toxic contaminants in the Niagara River. Primarily, it provides a framework to account for the uncertainty of both input variables and model parameters as a first step to optimize and improve the modeling effort. It also offers a systematic approach to developing better management practices based on the results of the uncertainty analysis.