depending on the potential audience. Elaborate graphics and sophisticated statistical analyses may be
included as an appendix or presented for the more technical reader in a different format.
There are four basic questions in sample design to be considered.
What should be sampled? What are the constituents of concern?
How many samples should be collected? What is the sample size?
When do I collect the samples? What is the sampling frequency?
Where do I collect the samples? What sites and depths should be used?
What should be sampled is a function of the perceived problem. In some case the problem is
easily identified and constituents to be monitored can be determined. For example, if hydrogen sulfide
odors are a problem, then monitoring for hydrogen sulfide is easily identified (this particular constituent
is not easily monitored). However, many processes may contribute to the cause of the problem and a
basic knowledge of water quality assists in determining which constituents should be monitored. In the
above example with hydrogen sulfide, the origin of the problem is the generation of dissolved sulfide
during anoxia in the upstream impoundment. Dissolved oxygen monitoring in the forebay of the
impoundment is a logical approach to identifying the development of the problem. Measuring the
oxidation-reduction potential may also prove useful in predicting the temporal extent of the problem.
More complex problems, such as a fish kill, may require a more detailed assessment. For example
several factors may contribute to the sudden death of fish in a reservoir tailwater (e.g., low dissolved
oxygen, thermal change, toxic compounds). Problems associated with increased concentrations of
reduced metals (e.g., staining, taste and odor), dissolved gasses, and other physico-chemical processes
may be repetitive and therefore can be predicted with a well-designed monitoring program. The
outcome of identifying the problem (and its source) should provide sufficient information for the design
of an enhancement technique.
2. Sample size
Determination of sample size should consider the amount of information needed to address the
objectives of the sampling effort with limitations of financial resources often the determining factor. If
insufficient funds are available, the option to not conduct the study (not desirable) may have to be used.
However, careful consideration of sample size versus data needs may allow an adequate sampling
effort. An equation for determining sample size for a simple random sample can be defined by: