Water quality and macrophytes in the Danube River: Artificial neural network modelling

Ecological assessment of large rivers such as the Danube is a challenging task. Eutrophication was reported one main drivers that structure aquatic communities in basin. Due to their sedentary nature, relatively slow growth/ long life spans, and engineering role ecosystems, macrophytes are widely used detection nutrient enrichment. In this study, macrophyte presence-absence data within 3 km reaches obtained from Joint Survey (JDS3) were predict water quality river its tributaries. For each variable (dissolved oxygen, nitrate-nitrogen, orthophosphates), multi-layer feed-forward artificial neural network model (ANN) constructed using explanatory variables. Despite limited number samples (123) along wide trophic gradient Danube, showed good predictive performances for channel. The highest discrepancy between observed predicted collected tributaries or downstream tributaries' mouth, where better conditions compared measured ones. From 64 analysed species, 28 selected by sensitivity analysis key indicators (KIS) at least environmental variable. KIS mainly belonged eutrophic tolerant submerged emerged species with broad ecological amplitude, which reflects significance developed use on subjected pollution. However, restricted sections having velocity suitable growth. ANN architecture represents modelling approach could be applied other lotic systems biological elements.