A predictive multimetric index based on macroinvetebrates for Atlantic Forest wadeable streams assessment

Abstract

Abstract: Multimetric Indices (MMIs) have been widely applied for ecological assessment in freshwater ecosystems. Most MMIs face difficulties when scaling up from small spatial scales because larger scales usually encompass great environmental variability. Covariance of anthropogenic pressures with natural environmental gradients can be a confounding factor in assessing biologic responses to anthropogenic pressures. This study presents the development and validation of a predictive multimetric index to assess the ecological condition of Atlantic Forest wadeable streams using benthic macroinvertebrates. To do so, we sampled 158 sites for the index development. We adjusted each biological metric to natural variation through multiple regression analyses (stepwise-forward) and considered that the residual distribution describes the metric variation in the absence of natural environmental influence. For metric selection we considered normal distribution, variation explained by the models, redundancy between metrics and sensitivity to differentiate reference from impaired sites. We selected five metrics to the final index: total richness, %MOLD, %Coleoptera, EPT richness and Chironomidae abundance. The residuals were transformed into probabilities and the final index was obtained through the mean of these probabilities. This index performed well in discriminating the impairment gradient and it showed a high correlation (r = 0.85, p <0.001) with a specific index developed for a particular basin indicating a similar sensitivity. This index can be used to assess wadeable streams ecological condition in Atlantic Forest biome, so we believe that this type of approach represents an important step towards the application of biomonitoring tools in Brazil.