Geoscience Reference
In-Depth Information
SENSITIVITY ANALYSIS FOR OPTIMIZATION MODEL
FOR COASTAL GROUNDWATER
SUNG-HUN HONG
∗
and NAMSIK PARK
Institute of Construction Technology and Planning, Dong-A University
P.O. Box 604-714, Hadan-Dong, Saha-Gu, 840, Busan, Korea
∗
wghsh72@smail.donga.ac.kr
A simulation-optimization model can be a useful tool in decision-making
regarding groundwater developments. A flexible model is developed to deter-
mine optimal solutions of not only the distribution of the pumping rates and the
locations, but also control measures to prevent or reduce saltwater intrusion.
This paper discusses the result of a sensitivity analysis for the model. Influence
of hydraulic conductivity, recharge rate and aquifer thickness on optimal pump-
ing rates and optimal well locations is studied. It is shown that the hydraulic
conductivity has more pronounced impact than other parameters on optimal
pumping rates and optimal well locations.
1. Introduction
Sensitivity analysis is important in understanding not only the characteris-
tics of the model, but also the behavior and influence of various parameters
in the model. The general purpose of a sensitivity analysis is to quan-
tify the uncertainty in the calibrated model caused by uncertainty in the
estimates of aquifer parameters, stresses, and boundary conditions.
1
Tran
2
studied the response in saltwater intrusion lengths with respect to stresses
and transmissivities using a SHARP simulation model. The objective of
this study is to observe how the optimal solutions change depending on the
parameters. Sensitivity analysis of major parameters on optimal solutions
is investigated to aid potential users of the model.
2. Sensitivity Analysis for Optimization Model
A hypothetical cross-sectional unconfined aquifer (Fig. 1) is used to conduct
the sensitivity analysis. The hydraulic conductivity (
K
), the total recharge
rate (
Q
rech
), and the aquifer thickness (
B
) are selected as control param-
eters. The optimal pumping rate (
Q
opt
) and the optimal location of the
well (
L
opt
) are state variables that are affected by the control parameters.
167
