Pandora's Box: Can We Distinguish Groundwater
transport Hypotheses Given Observational
uncertainties?
Open Access
Author:
Urban, Rachel Lorah
Graduate Program:
Civil Engineering
Degree:
Master of Science
Document Type:
Master Thesis
Date of Defense:
March 13, 2013
Committee Members:
Patrick M Reed, Thesis Advisor/Co-Advisor Kamini Singha, Thesis Advisor/Co-Advisor
Keywords:
groundwater transport observational uncertainty
Abstract:
This study explores the diculties associated with using groundwater transport models to distinguish alternative process hypotheses given observational uncertainties. A well-characterized laboratory sand tank, packed with two sands, serves as the basis for our study. Using time-series breakthrough curves at multiport wells, we monitor the flow (steady state) and transport of a 1.7-h step-pulse injection of NaCl across nine replicate experiments. The replicate tracer tests show
signicant variability despite the simplicity and highly controlled nature of the experiments. Local and non-local transport models are shown to be indistinguishable given experimental uncertainty. Moreover, our results demonstrate how physically implausible parameterizations of the more complex non-local transport model lead to improved model performance measures with little basis in the true system transport dynamics. The replicate tracers experiments performed in this study represent a lower bound of problem complexity. The inability to distinguish transport hypotheses as observed in these results would only be exacerbated in higher complexity groundwater systems.