Beyond BACI: the detection of environmental impacts on populations in the real, but variable, world

Abstract

BACI (Before/After and Control/Impact) sampling is widely used in investigations of environmental impacts on mean abundance of a population. The principle is that an anthroppogenic disturbance in the “impact” location will cause a different pattern of change from before to after it starts compared with natural change in the control location. This can be detectable efficiently as a statistical interaction in an analysis of variance of the data. Usually, samples are taken at replicated, random intervals of time before and after the putative impact starts; this ensures that chance temporal fluctuations in either location do not confound the detection of an impact. These designs are, however, insufficient because any location-specific temporal difference that occurs between the two locations will be interpreted as an impact even if it has nothing to do with the human disturbance. Alternatively, abundance in the single control location may change in the same direction, cancelling the effects of an impact. Here, asymmetrical designs are developed that compare the temporal change in a potentially impacted location with those in a randomly-selected set of control locations. An impact must cause a different temporal change in the disturbed location from what would be expected in similar locations. This can be detected for short-term (pulse) or long-term (press) impacts by different patterns of sigficance in the temporal interactions between of sampling and locations. Frolm these novel designs, tests are that demonstrate whether an usual pattern of temporal change in abundance of organisms is specific to the supposedly impacted location and correlated with the onset of the disturbance. Examples are presented of how to use these designs to detect impacts at different spatial scales. Other aspects of their use are discussed.