Comparison of ambient samples

Ward and Loftis (1986) recommend the calculation of measures of central tendency, such as arithmetic means, or geometric means if data are highly skewed, and the plotting of these on charts or graphs as the easiest method for communicating differences in environmental quality to the public and environmental managers. Confidence limits should be reported always with measures of central tendency, such as means and lethal concentration (LC)/effective concentration 

Because analyses of water quality usually involve the collection of data on several variables, multivariate statistical analyses often are relevant. For example, multivariate analysis of variance (MANOVA) and discriminant analysis were used by Alden (1997) to investigate water quality trends in Chesapeake Bay. 

Simple time-series plots of water or air quality can reveal a great deal of information (Ward and Loftis, 1986), such as  

---

If the aim is to detect time-series trends associated with environmentally impacting activities, then samples should be taken at regular intervals before and after the impact, with more samples taken after in case there is a lag in any effect. Of course, in many cases the impact may have occurred already if it was accidental, in which case the best design is that for comparison of ambient samples described above. It may be difficult to find truly independent sites that are sufficiently similar. For example, rivers are often monitored by comparing sites upstream of an impact with sites downstream of the impact. It is arguable whether such sites are truly independent. 

---