|
Optimal Impact Study Design and Analysis Pages: 26 Published: Jan 1978
Download this paper for $25
PDF (368K)
View License Agreement An optimal impact study design is possible if (1) the impact has not yet happened, (2) the type of impact and when and where it will happen are known, (3) measurements on both biological and environmental variables can be obtained from the samples, and (4) an area which will not receive the impact, but is otherwise similar to the impact area, is available to serve as a control. The example presented here is based on simulated data consisting of sample estimates of species abundances and environmental variables. A total of 36 samples are distributed among locations in the impact and control areas and between “before” and “after” impact times. A multivariate analysis of variance approach permits a test of the null hypothesis that any change in the biotic community of the impact area does not differ from that in the control area. A procedure which permits impact-related biotic change to be related to environmental change is illustrated, and the results are displayed visually. Use for subsequent biological monitoring to detect a future impact of the same type is illustrated. A cluster analysis approach, which both effectively displays the impact effects and corroborates the analysis of variance results, is illustrated. Design and analysis options are discussed extensively. | ||
