How much is too little to detect impacts? A case study of a nuclear power plant.

Several approaches have been proposed to assess impacts on natural assemblages. Ideally, the potentially impacted site and multiple reference sites are sampled through time, before and after the impact. Often, however, the lack of information regarding the potential overall impact, the lack of knowledge about the environment in many regions worldwide, budgets constraints and the increasing dimensions of human activities compromise the reliability of the impact assessment. We evaluated the impact, if any, and its extent of a nuclear power plant effluent on sessile epibiota assemblages using a suitable and feasible sampling design with no 'before' data and budget and logistic constraints. Assemblages were sampled at multiple times and at increasing distances from the point of the discharge of the effluent. There was a clear and localized effect of the power plant effluent (up to 100 m from the point of the discharge). However, depending on the time of the year, the impact reaches up to 600 m. We found a significantly lower richness of taxa in the Effluent site when compared to other sites. Furthermore, at all times, the variability of assemblages near the discharge was also smaller than in other sites. Although the sampling design used here (in particular the number of replicates) did not allow an unambiguously evaluation of the full extent of the impact in relation to its intensity and temporal variability, the multiple temporal and spatial scales used allowed the detection of some differences in the intensity of the impact, depending on the time of sampling. Our findings greatly contribute to increase the knowledge on the effects of multiple stressors caused by the effluent of a power plant and also have important implications for management strategies and conservation ecology, in general.