The Value of Hydraulic Conductivity Information for the Optimal Restoration of an Over-exploited Aquifer

• Published in: Procedia environmental sciences 2015, Vol.25, p.227-234
• Main Authors: Sidiropoulos, P., Mylopoulos, N.
• Format: Article
• Language: English
• Subjects: aquifer restoration; geostatistical approach; hydrogeological uncertainty; Monte Carlo simulation and optimization
• ISSN: 1878-0296; 1878-0296
• DOI: 10.1016/j.proenv.2015.04.031

A stochastic management tool is developed and applied in order to evaluate the worth of hydraulic conductivity data on the optimal restoration and quantitative management of the over-exploited aquifer of Lake Karla watershed in Greece. This tool consists of six models (one geostatistical, four simulation models and one management model) and combines the methodologies of: stochastic simulation-optimization, Bayesian analysis and the value of information analysis. The four simulation models (surface hydrology, reservoir operation, lake-aquifer interaction and hydrogeology) are interlinked in order to satisfy the needs of integrated simulation at the watershed scale. The heterogeneity and the lack of sufficient data of hydraulic conductivity create uncertainty on the hydraulic heads estimation. Monte Carlo realizations of hydraulic conductivity are being performed with the use of geostatistical tools and imported to the groundwater model to give multiple stochastic realizations of the aquifer. A Monte Carlo based optimization problem is then applied for each aquifer realization in order to determine the optimal aquifer's restoration management strategy. Optimal strategy has been defined the one that combines the maximum possible volume of extracted groundwater and the optimal well's position with the least financial cost, under the environmental constraint of restoring the aquifer water table. The hydrogeological uncertainty is being transformed into financial uncertainty through the optimization problem, as certain risks for the decision maker are being introduced. To avoid hydraulic head underestimation, a Bayesian decision analysis for the hydraulic conductivity data collection is being applied on each optimal solution. The worth of the new hydraulic conductivity data can be evaluated by quantifying the reduction of both hydrogeological and financial uncertainties. The results prove that there is a certain number of new hydraulic conductivity measurements up to which the profit by reducing financial uncertainty exceeds the measurement cost.