Avaliação da integridade biótica em riachos: um protocolo para a identificação de locais de referência e escolha de métricas

Abstract

The identification of minimally impacted sites and the selection of metrics for assessing the effects of human activities on aquatic environments has been a key issue in the management of these ecosystems. This paper presents a practical and easy method to identify reference sites and selection of the most significant metrics that reflect the general attributes of the assemblies, altered by human impacts. For this we evaluated the biotic integrity in some neotropical streams, located in Paraná state, Brazil, through the benthic macroinvertebrates and fish. The identification of reference sites was based on the physical and chemical quality of the habitat, and the physical and chemical variables summarized by principal component analysis, and submitted to cluster analysis to identify the reference sites. Sixteen metrics were calculated, which were submitted to a discriminant analysis to select the most significant metrics. The principal component analysis and cluster analysis showed that the reference sites were selected according to composition ande physical quality of the sediment. These sites were positively correlated with larger particles such as pebbles and negatively associated with smaller particles. The discriminant analysis showed that few metrics are sufficient to identify changes in the assemblies by human activities and that assemblies respond differently to environmental changes. In the pooled analysis of the assemblies the metrics selected for macroinvertebrate were different than metrics selected in the analysis performed solely for that assembly, in the pooled analysis were selected metrics abundance, percentage of tolerant and percentage of shredders, whereas in exclusive analysis were selected metrics richness and percentage of tolerant. Greater richness of macroinvertebrates were observed at pristine sites, while the percentage of tolerance increased with the gradient of impact. On the other hand, for the fish assemblage were selected metrics of richness, diversity and percentage of omnivores. Greatest richness were observed in moderately impacted sites, while higher values of diversity and percentage of omnivores were observed in reference sites. The analysis of the metrics of the two assemblages provided a more robust evaluation the sites categorized according to a gradient of impact, than the separate evaluation of the assemblages. However, the benthic macroinvertebrates assemblage was equally effective to confirm the reference sites. The procedure presented in this paper can easily assist managers in selecting effective methods for protecting and restoring stream ecosystems