Published: Jan 1985
| ||Format||Pages||Price|| |
|PDF (272K)||19||$25||  ADD TO CART|
|Complete Source PDF (5.6M)||426||$55||  ADD TO CART|
A design for studying sources of environmental contamination must start with a basis for distinguishing the contamination of interest from the background. As part of this basis, the design should provide a method for assessing the sampling and measurement error. Because of problems with reports of none detected, the design should also include a plan for analyzing intermediate laboratory results in addition to the reported values. This paper discusses these aspects of design in the context of monitoring the groundwater around a waste management facility. A design appropriate for spatially and temporally varying backgrounds is proposed and illustrated with monitoring results from Alabama and Florida. To assess the sampling error, the proposed design specifies resampling each well after a period of a few days. Experiments to check this procedure are suggested. The proposed design incorporates and supplements the Environmental Protection Agency laboratory method for total organic halide. In addition, this paper illustrates some difficult design problems that involve nonnormality and nonlinear measurement methods.
analysis of variance, environmental contamination, experimental design, groundwater monitoring, low-level measurements, measurement, nonlinear measurement methods, nonnormal errors, pollution sources, sampling, statistical design, statistics
Center for Applied Mathematics, National Bureau of Standards, Gaithersburg, MD