Dissecting the distribution of brittle stars along a sewage pollution gradient indicated by organic markers.

We have assessed variation in brittle star distribution patterns along a contamination gradient identified by fecal steroids and aliphatic hydrocarbons in Paranaguá Bay, southern Brazil. A hierarchical design using multiple spatial scales (centimeters-kilometers) was applied. Generalized linear mixed models (GLMMs) were used to investigate the spatial and temporal variability of brittle stars. Main principal components from the contamination and environmental matrices were used to investigate the best explanatory dataset. The abundance of brittle stars was significantly lower in sites with high concentrations of fecal steroids and aliphatic hydrocarbons. The best model fitting always included components from the contamination gradients, which precludes a purely environmental driving of brittle star abundance. Variability in spatial scales lower than kilometers was probably driven by sediment characteristics. We highlighted the importance of a robust multi-scale sampling design for a better biological indication of coastal contamination.