Assessing Environmental Impacts: Effects of Spatial and Temporal Variability at Likely Impact Scales

Ecological and environmental variability in natural ecosystems precludes the widespread use of simplistic design and analysis tools to detect the effects of human activities on natural ecosystems. BACI (Before–After and Control–Impact) sampling and analysis is often recommended to overcome the problem of inferring that changes are due to an impact rather than natural variability. Using five different BACI design and analysis techniques (from nested ANOVA to time-series intervention) on data with spatial, temporal, and sampling variability at scales relevant to impact assessments, we investigated how BACI designs cope with interactions between spatial and temporal variability and whether natural variability is best treated as random. A number of simulated impacts of different sizes and types were imposed on long-term observations of six macrofaunal species and three diversity indices. A major problem in detecting the impacts was the differences in variability between sites, although the statistical detection of impacts was aided by treating temporal variability as an underlying process with structure, rather than random noise. However, the sensitivity of any method depended on the spatial and temporal variability exhibited by the ecological measure. Our results emphasize that scale is one of the most important concepts in impact assessment. As spatial or temporal scale increases, both the number of processes and their importance in influencing local populations and communities will change, increasing the variability encompassed by the study. We suggest monitoring a number of variables (populations) and developing sampling strategies that enable a number of methods of analysis to be employed.