A Statistical Strategy for Determining Contaminant Impacts of an Industrial Monofill on Karst Springs

Collection of representative data in karst aquifers often requires monitoring at appropriately-located springs. Following dye tracing and geophysical investigations, one spring was selected as monitoring location for a landfill facility in southern Indiana. Nine rounds of sampling were conducted under different flow conditions. Sequential discrete water samples were collected across the hydrograph curve for each storm-pulse sampling event. The sampling frequency varied according to the flow conditions and the spring responses to the recharge events. Such sampling approaches characterized the natural variations of the water quality at the spring but complicated the development of an effective statistical evaluation plan for the landfill. A suite of combined Shewhart-CUSUM control charts were developed for parameters that are subject to statistical analyses to statistically determine the possible impact of the landfill on the karst system. The statistical power of the control charts was evaluated by Monte Carlo simulations.