Formatted Title
Waste Not, Want Not: A Systematic Framework for Gauging Network Optimization
Background/Objectives
Over time, it is not uncommon for monitoring networks at large sites with impacted groundwater to become unwieldy with hundreds or even thousands of wells. Often, these networks evolved to meet multiple, wide-ranging objectives (e.g., plume delineation, source delineation, sentinel wells, remedy performance evaluation, etc.) and, therefore, lack cohesion, particularly with respect to water-level data collection and interpretation. In most cases, water levels are simply collected from the entire available network simply out of convention. However, at large sites this can be a sizable effort resulting in the collection of redundant and unnecessary data, which leads to overall project inefficiencies (i.e., increased time, costs, staff exposure, data management requirements, etc.). This can be especially true for sites that are required to demonstrate hydraulic control or containment (e.g., pump and treat). Water levels are gauged at an established interval to interpret groundwater conditions and evaluate hydraulic performance. In these cases, gauging network optimization, with the objective of defining a reduced network (and frequency), can result in increased project efficiencies by reducing data redundancies and/or the collection of non-essential data without loss of efficacy when assessing hydraulic conditions.
Approach/Activities
A systematic framework for gauging network optimization was developed, and two case studies are presented to demonstrate the potential benefits that can be realized through gauging optimization. The optimization steps include: (1) development of a time-series database for all wells within the current gauging network; (2) performing principal component analysis to define data correlation and redundancy within the gauging network to establish an initial optimized network; (3) executing iterative geostatistical and geospatial analyses to refine the initial optimized network; (4) validating the proposed network to verify that mapped water-level data using the proposed optimized network accurately reflects conditions inferred when using the comprehensive gauging network; (5) supplementing proposed network as necessary to meet regulatory and/or other project objectives; and (6) evaluating temporal changes in horizontal hydraulic gradients at key locations within the proposed gauging network to support modifying gauging frequency.
Results/Lessons Learned
Through applying this optimization process, a reduction in gauging locations (and frequency) of about 50% was realized in both case studies while providing the same level of fidelity as the full network when interpreting groundwater flow conditions (i.e., arriving at the same conclusions when evaluating plume containment). The results of the optimization exercise suggest that if this effort were to be applied at other sites with similar results (i.e., water-level data from only about one-half of the current gauging network and levels measured less frequently), the return on investment could be realized within a few years, allowing limited/valuable project resources to be better applied and/or directed toward advancing site cleanup or other project objectives. Additionally, this optimized network and gauging frequency significantly reduces the amount of time that workers would be exposed to site hazards.