Published: Jan 1988
| ||Format||Pages||Price|| |
|PDF (224K)||15||$25||  ADD TO CART|
|Complete Source PDF (8.5M)||15||$76||  ADD TO CART|
Systematic and random sampling (stratified or not) are widely applied to monitor toxic chemicals in streams, municipal wastes, sediments, dredge spoils, and so on. One advantage of systematic sampling is the simplicity of identifying exact sampling points; however, it can lead to biased results. The advantage of simple random and stratified random sampling is the expectation that results will be unbiased; however, it is difficult to identify exact sampling points. The primary attraction of systematic sampling is that, if no cyclic phenomenon is present, it can produce unbiased results with better precision than possible with simple random or stratified random sampling. The test described in this paper provides an answer to the difficult choice of selecting the most appropriate sampling methodology for a particular case. The distribution of toxic chemicals in a number of sites is an autocorrelated function of space or time most of time. The study of this function provides an effective means of deciding how many increments a sample should contain, how often an increment should be taken, and how the increment should be collected to keep its representativeness when cyclic phenomena occur. The test provides essential information on the sampling selection error by separating this error into three major components. The first component is a short-range quality fluctuation error, which includes errors introduced by small local composition and distribution heterogeneities. The second component is a long-range quality fluctuation error, which represents the continuous trend introduced by the waste producer. The third component is a periodic fluctuation error, which may render systematic sampling a difficult and risky operation. Systematically implementing a variographic experiment on sites where a long-term commitment for monitoring is essential is recommended. Not only does it effectively optimize the sampling methodology, but it also gives valuable information on the variability of the parameter of interest, which in turn gives valuable clues to the waste producer in solving his problem.
sampling methodology, monitoring of streams, variography, systematic sampling, stratified random sampling, random sampling, trends and sampling, cycles and sampling, heterogeneities and sampling
President of Pierre Gy and Francis Pitard Sampling Consultants, Broomfield, CO