Characterizing And Incorporating Uncertainty In Drinking Water Quality And Treatment
Towler, Erin L 1 ; Rajagopalan, Balaji 2 ; Summers, Scott 3 ; Seidel, Chad 4
1 University of babyÖ±²¥app at Boulder
2 University of babyÖ±²¥app at Boulder
3 University of babyÖ±²¥app at Boulder
4 Damon S. Williams
Drinking water utilities face complex decisions when balancing new and changing regulatory requirements with competing finished water quality objectives. Tools are needed to help utilities better understand treatment plant performance in light of natural influent water quality variability and regulation compliance. To this end, the natural variability of select water quality variables in the United States were characterized in terms of their spatial and temporal variability. Based on these relationships, a K-nearest neighbor (K-NN) bootstrap technique was developed to generate ensembles of influent water quality. Next, a statistical model was developed to simulate conventional water treatment using local polynomial (nonparametric) regression methods. Finally, input uncertainty was incorporated into the model to see output scenarios and the probability of exceeding a given regulation limit.