HR: 1340h
AN: NG23D-0118    [Abstracts]
TI: Linear Stochastic and Non-linear Deterministic Paradigms for Improved Spatial Interpolation of Groundwater Contamination: Implications for Management of Arsenic Contamination in Bangladesh
AU: * Hossain, F
EM: fhossain@tntech.edu
AF: Tennessee Technological University, Department of Civil and Environmental Engineering, Box 5015, 1020 Stadium Drive, Cookeville, TN 38505-0001 United States
AU: Hill, J
EM: ajhill21@tntech.edu
AF: Tennessee Technological University, Department of Civil and Environmental Engineering, Box 5015, 1020 Stadium Drive, Cookeville, TN 38505-0001 United States
AU: Sivakumar, B
EM: sbellie@ucdavis.edu
AF: University of California Davis, Department of Land Air and Water Resources, Davis, CA 95616 United States
AU: Bagtzoglou, A C
EM: acb@engr.uconn.edu
AF: University of Connecticut, Department of Civil and Environmental Engineering, 261 Glenbrook Road, U 2037, Storrs, CT 06269 United States
AB: Since the discovery of large-scale arsenic contamination of groundwater in Bangladesh more than a decade ago, studies related to its spatial characterization have relied on linear stochastic (geo-statistical) approaches. In this work, we explore an alternative nonlinear paradigm alongside the linear stochastic approach, with a motivation to possibly improve spatially interpolated contamination fields. First, we report spatial variability analyses on the basis of a purely geostatistical framework (variogram analyses), where regional anisotropy in the spatial dependence of arsenic for Northwest region of Bangladesh was found to be stronger than that in the Southwest. We then investigate the effectiveness of geostatistical approaches, specifically ordinary kriging, for making rapid decisions on zonal (~25 km2) management of arsenic contaminated shallow groundwater. Next, we identify the possible presence of nonlinear deterministic and chaotic pattern via the Grassberger-Procaccia correlation dimension algorithm. Our analyses reveal correlation dimension values ranging anywhere from 8 to 11 depending on the region, suggesting that the arsenic contamination in space, from a chaotic dynamic perspective, is a medium- to high-dimensional problem. We then apply ordinary logistic regression to demonstrate the possibilities of such multi-dimensional modeling to predict arsenic concentrations in space. Finally, we propose a low-cost and non-structural simulation approach for improved spatial interpolation that aims at robust identification of safe/unsafe wells. Our proposal is founded on the premise that the near impossibility of testing every single shallow well in a rural setting (such as in Bangladesh) requires a simulation methodology that can accurately characterize a well as being safe/unsafe without the need for extensive and expensive in-situ sampling tests. Such a method can then act as a fast-running and inexpensive proxy to the time-consuming field campaigns and save considerable testing resources by judiciously directing them to those wells pre-determined by the simulation approach to have a high likelihood of being unsafe. Our proposed approach is based on the recognition that linear stochastic tools alone may only be partially adequate to understand the large-scale variability of arsenic contamination in the context of identifying drilling locations of safe wells in a rural setting. The proposed approach therefore includes a pattern recognition approach based on non-linear deterministic and chaotic dynamic theory that could be exercised to achieve a more ?complete? estimate of the arsenic field.
UR: http://iweb.tntech.edu/fhossain
DE: 1831 Groundwater quality
DE: 3265 Stochastic processes (3235, 4468, 4475, 7857)
DE: 3275 Uncertainty quantification (1873)
DE: 4420 Chaos (7805)
SC: Nonlinear Geophysics [NG]
MN: Fall Meeting 2005