HR: 0800h
AN: H11D-1307    [Abstracts]
TI: Quantifying Uncertainty in Complex Groundwater Flow Models
AU: * Ni, C
EM: nichuenf@egr.msu.edu
AF: College of Engineering, Michigan State University, A121 Research Complex-Engineering, Department of Civil and Environmental Engineering, Michigan State University, East Lansing, MI 48824 United States
AU: Li, S
EM: lishug@egr.msu.edu
AF: College of Engineering, Michigan State University, A121 Research Complex-Engineering, Department of Civil and Environmental Engineering, Michigan State University, East Lansing, MI 48824 United States
AB: Despite the intensive research over the past decades in the field of stochastic subsurface hydrology, our ability to analyze and model heterogeneous groundwater systems remains limited. Most existing theories are either too restrictive to handle practical complexity or too expensive to be applied to realistic problem sizes. In this study we present a hybrid spectral method that allows quantifying uncertainties (caused by unmodeled conductivity heterogeneity) in complex groundwater flow models. This method, based on solving stochastic perturbation equations, involves two major computational steps after the deterministic mean flow equation is solved. We first apply a set of closed-form formulas, valid for mildly nonstationary systems (Ni and Li, WRR, in press), to predict the variances for the entire modeling area. We then employ locally first-order numerical spectral method (Li and McLaughlin, WRR, 1991; Li et al., WRR, 2002) to correct the 'regional' solution in areas where the variance distribution is highly nonstationary (e.g., around discontinuities and singular sources/sinks). The boundary conditions for the local numerical solutions are based on the closed-form formulas and are implemented in a real-time visual steering software environment called Interactive Ground Water (IGW) [Li and Liu, Ground Water, in revision]. Since the "regional" closed form solution can be evaluated instantaneously and the more expensive first-order numerical analysis is only applied locally, the overall hybrid approach can be very efficient, making it possible to model large-scale, complex flow systems in the presence of general trends and sources and sinks. We illustrate the accuracy and effectiveness of hybrid spectral method with a number of examples.
DE: 1829 Groundwater hydrology
DE: 1847 Modeling
DE: 1849 Numerical approximations and analysis
DE: 1869 Stochastic hydrology
DE: 3265 Stochastic processes (3235, 4468, 4475, 7857)
SC: Hydrology [H]
MN: Fall Meeting 2005