Files in this item

FilesDescriptionFormat

application/pdf

application/pdfJi_Xiang.pdf (3MB)
(no description provided)PDF

Description

Title:Impact of uncertain input on parameter estimation in groundwater model
Author(s):Ji, Xiang
Advisor(s):Valocchi, Albert J.
Department / Program:Civil & Environmental Eng
Discipline:Civil Engineering
Degree Granting Institution:University of Illinois at Urbana-Champaign
Degree:M.S.
Genre:Thesis
Subject(s):Thiem
Theis
groundwater
parameter estimation
Parameter ESTimation (PEST)
uncertainty
Abstract:Description of the aquifer characteristics accurately and efficiently is the most commonly encountered and probably the most challenging aspect of groundwater modeling. In the context of groundwater modeling, although many studies have focused on parameter estimation problems, these issues are far from being solved. When important hydrogeological parameters like transmissivity and storativity are estimated using regression-based inverse methods, it is assumed that all other parameters and quantities are known. In particular, it is assumed that pumping rates are known. This will not be a valid assumption for groundwater basins subject to intensive irrigation pumping since farmers are normally not required to report their pumping amounts to any government regulatory office. In this thesis, we study the impact of uncertainty in pumping upon estimation of hydrogeological parameters. We use three typical simplified groundwater models to test the impact of uncertain pumping on the parameter estimation and we use statistical methods to assess the results. The uncertainty analysis using the Matlab Regression Toolbox of the Thiem and Theis model shows that the impact of uncertain drawdown is less than the impact of uncertain pumping. The uncertainty analysis using PEST for a more complex model with a partially penetrating stream shows that the stream depletion cannot be used to estimate the transmissivity and the drawdown cannot be used to estimate the riverbed conductivity. The biases of estimated parameters commonly exist and they increase with the increasing uncertainty of model input. The impact of uncertain pumping rate is also more significant than the impact of uncertain observations. Finally, we estimate the pumping uncertainty in a real case by studying the data from the Republican River Compact Administration (RRCA) model. In this unusual case, we have actual metered pumping data, as well as an assumed pumping rate that was used in the RRCA model. For the Upper Natural Resources District of Nebraska, the error (uncertainty) in pumping rates approximately follows a Gaussian distribution. But the pumping rate used in the model is underestimating the actual pumping data.
Issue Date:2012-06-27
URI:http://hdl.handle.net/2142/31967
Rights Information:Copyright 2012 Xiang Ji
Date Available in IDEALS:2012-06-27
2014-06-28
Date Deposited:2012-05


This item appears in the following Collection(s)

Item Statistics