Multivariate methods and artificial neural networks in the assessment of the response of infaunal assemblages to sediment metal contamination and organic enrichment

Abstract

A 4-year annual sediment survey was conducted in an organically enriched tidal channel to compare the performance of univariate community descriptors, traditional multivariate techniques (TM) and artificial neural networks (AANs), in the assessment of infaunal responses to moderate levels of sediment metal contamination. Both TM approaches and the SOM ANN revealed spatiotemporal patterns of environmental and biological variables, suggesting a causal relationship between them and further highlighting subsets of taxa and sediment variables as potential main drivers of those patterns. Namely, high values of non-natural metals and organic content prompted high abundances of opportunists, while high values of natural metals yielded typical tolerant assemblages of organically enriched areas. The two approaches yielded identical final results but ANNs showed the following advantages over TM: ability to generalise results, powerful visualization tools and the ability to account simultaneously for sediment and faunal variables in the same analysis. Therefore, the SOM ANN, combined with the K-means clustering algorithm, is suggested as a promising tool for the assessment of the ecological quality of estuarine infaunal communities, although further work is needed to ensure the accuracy of the method.