Optimal characterization of pollutant sources in contaminated aquifers by integrating sequential-monitoring-network design and source identification: methodology and an application in Australia

Abstract

Often, when pollution is first detected in groundwater, very few spatiotemporal pollutant concentration measurements are available. The contaminant concentration measurement data initially available are generally sparse and insufficient for accurate source characterization. This requires development of a contaminant monitoring plan and its field implementation to collect more data. The location of scientifically chosen monitoring points and the number of measurements are important considerations in improving the source-characterization process, especially in a complex contamination scenario. In order to improve the efficiency of source characterization, a feedback-based methodology is implemented, integrating sequentialmonitoring-network design and a source identification method. The simulated annealing (SA) optimization algorithm is used to solve the models for optimal source identification and the monitoring-network-design optimization. This sequence is repeated a few times to improve the accuracy of source characterization. The methodology is based on the premise that concentration measurements from a sequence of implemented monitoring networks provide feedback information on the actual concentration in the site. This additional information, obtained as feedback from monitoring networks designed and implemented based on intermediate source characterization, can result in sequential improvement in the resulting source characterization. The performance of this methodology is evaluated by application to a contaminated aquifer site in New South Wales, Australia, where source location, sourceactivity initiation time and source-flux (mass per unit time) release history are considered as unknown variables. The performance evaluation results demonstrate potential applicability of the proposed sequential methodology.

Characterization of Groundwater Pollution Sources by Kriging Based Linked Simulation Optimization

“…Nonlinear regression model was applied to identify the release flux histories of groundwater pollution sources [1]. Linked Simulation Optimization (LSO) models are very popular in identifying the release histories of pollution sources [2][3][4][5]. Surrogate models are efficiently used to recover the release histories of groundwater pollution sources when locations of sources are well known [6][7].…”

Application of Simulated Annealing and Adaptive Simulated Annealing in Search for Efficient Optimal Solutions of a Groundwater Contamination related Problem

“…Designed monitoring networks [6][7][8][9] can reduce the nonuniqueness related to data availability. However, the nonuniqueness related to the search for a single global optimal solution to the inverse problem depends on the efficiency of the optimization algorithm.…”

“…The ill-posed nature of the inverse problem and the plausibility of nonunique solutions can be interrelated as well. More efficient monitoring networks can reduce the plausibility of nonunique solutions, and therefore, optimal monitoring network design is a related issue [9,22]. Also, only if the global optimum solution is found, it may represent the accurate solution to the source identification problem.…”

Adaptive Surrogate Model Based Optimization (ASMBO) for Unknown Groundwater Contaminant Source Characterizations Using Self-Organizing Maps

“…The linked simulation optimization procedures consist of two main components: 1) models for simulation of groundwater flow and contaminant transport processes, 2) optimization model with an optimization algorithm. Some of the optimization algorithms utilized are linear programming and multiple regressions technique [7]; a nonlinear optimization model with embedding technique [8] [9] [10]; Genetic Algorithm (GA) [11] and [12]; the Artificial Neural Network (ANN) [13] and [14]; a hybrid methodology based on GA [15] and [16]; the classical nonlinear optimization algorithm [17]; Simulated Annealing (SA) [18] [19] [20] [21] and Adaptive Simulated Annealing (ASA) [22], Genetic Programming (GP) [23] and [24]; ASA in conjunction with uncertainty modeling [25] and [26]. Application of these methodologies to real-world cases is generally computationally time intensive, and may need days or weeks of CPU time to obtain an optimal solution.…”