The impact of acidification on macroinvertebrate assemblages in welsh streams: towards an empirical model.

The acidification of surface waters has profound ecological consequences. There is a need to predict the effects of possible future patterns of acid deposition on the biological components of fresh waters. This paper describes a model of the relationships between water chemistry and macroinvertebrate assemblages in eighteen streams in the upper Tywi whose catchments are subject to different land uses. Using established statistical techniques on data sets derived from riffle and margin samples taken in spring and summer, the macroinvertebrate assemblages were classified into three groups, which corresponded with streams draining conifer afforested catchments, acidic moorland streams and circumneutral moorland streams. Following principal components analysis to select key environmental variables, the application of multiple discriminant analysis generated two discriminant functions which were related most strongly to mean filterable aluminium concentration and mean total hardness, respectively. The discriminant functions were used to assign site-group membership with 100% success in the case of the spring data set with combined habitats. In addition, multiple regression of the primary ordination axis of each data set on mean aluminium concentration and mean hardness or pH, produced equations which explained 62.0%-87.2% of the variance. We conclude that the methods used here provide an effective analytical and potentially predictive tool for use in the understanding and management of the impact of acidification on freshwater ecosystems.