Search

A large dataset from 78 Italian lakes was collated to analyze the relationships between species of benthic macroinvertebrates and environmental factors and to test if different lake types erected according to the Water Framework Directive (WFD) system B classification matched with distinct communities. The lakes have different morphometry (altitude, volume, surface, depth), origin (glacial, volcanic), chemistry (conductivity, pH) and anthropogenic pressures (eutrophication, acidification, hydromorphological alteration, climatic change). More than two hundred species (276) were found of which 189 belonged to Chironomids and 30 to Oligochaetes. Samples for which both quantitative benthic samples and environmental data (morphometric and chemical) were available were selected (2404 samples). Taxa present in the database with > 50 specimens (83 taxa) were included in data analysis. A self organizing map (SOM) from artificial neural network (ANN) followed by a k-mean clustering was used to order and classify sites. Environmental variables were included in the analysis as supplementary variables. SOM analysis gave a map of 240 cells, a k-mean cluster analysis grouped cells into 12 clusters: cluster 4 was characterized by species preferring high altitude and low water temperature, cluster 5 by species tolerating high TP and low dissolved oxygen; cluster 2 grouped species preferring high conductivity, cluster 1 those preferring high pH, cluster 10 those preferring high water temperature, cluster 3 included species with optimal response in sublittoral stations of large lakes, whereas cluster 8 species preferring profundal stations of large lakes; clusters 7 and 11 were defined by a characteristic species composition but no environmental variables could be associated. Natural variables well separated lakes into different types, but clusters of macroinvertebrate species did not match with the types erected according to the WFD system B classification. Altitude, depth, volume, conductivity used in the system B influenced the community structure, but others were the key factors (i.e. water temperature, dissolved oxygen and total phosphorous) determining species composition in different lakes. To sum up there was no agreement between the WFD system B classification and the one suggested by macroinvertebrates. Using weighted means of total phosphorus, Secchi transparency and minimum hypolimnetic oxygen % saturation an optimum value for each species was calculated and a benthic quality index (BQIM) was built. With the data available, it was not possible to build distinct BQIM for each lake type and the present results do not support the need of separate indexes.