Groundwater Monitoring Network Design Using Additional Objectives in Dual Entropy Multi-Objective Optimization Method

Abstract

This study explores the applicability of including groundwater recharge and water table variation as additional objective functions in a multi-objective optimization approach to design optimal groundwater monitoring networks. The study was conducted using the Ontario Provincial Groundwater Monitoring Network wells in the Hamilton, Halton, and Credit Valley regions in southern Ontario. The Dual Entropy-Multiobjective Optimization (DEMO) model which has been demonstrated to be sufficiently robust for designing optimum hydrometric networks was used in these analyses. The importance of determining the applicability in using additional design objectives in DEMO, including groundwater recharge and groundwater table seasonal variation, is rooted in the limitations of groundwater data and the time required setting up the models. While recharge allows for the capturing of spatial variability of climate, geomorphology, and geology of the area, the groundwater table series reflect the temporal/seasonal variability. The two set of information are complementary and should provide additional information to the DEMO for optimal network design. Two sources of groundwater recharge data were examined and compared; the recharge provided by the local conservation authorities, calculated using both the Precipitation-Runoff Modeling System (PRMS) and Hydrological Simulation Program--Fortran (HSP--F), and the recharge calculated in situ using only PRMS. The entropy functions are used to identify optimal trade-offs between the maximum possible information content and the minimum shared information between each of the existing and potential monitoring wells. The additional objective functions are used here to quantify the hydrological characteristics of the vadose zone in the aquifer as well as the potential impacts of agricultural, municipal, and industrial uses of groundwater in the area, and thus provide more information for the optimization algorithm to use. Results show that including additional design objectives significantly increases the number of optimal network solutions and provides additional information for potential monitoring well locations. These results suggest that it is worthwhile to include recharge as a design objective if the data is available, and to include groundwater table variation for the design of monitoring wells for shallow groundwater system.