In comparison to lentic systems, the species composition and community structure of phytoplankton in lotic habitats are still poorly understood. We investigated the spatial and temporal dynamics of the phytoplankton community in a German lowland river, the Kielstau catchment, and the relationships with environmental variables. Among the 125 taxa observed, Desmodesmus communis, Pediastrum duplex and Discostella steligera were dominant species at lentic sites while Tabellaria flocculosa, Euglena sp., Planothidium lanceolatum, Cocconeis placentula and Fragilaria biceps dominated at lotic sites. Remarkable spatial and temporal variations of the phytoplankton community were revealed by non-metric multidimensional scaling. Canonical correspondence analysis indicated that physical factors (e.g. hydrological variables) and major nutrients [e.g. total phosphorus, dissolved inorganic nitrogen (DIN)] were of equal importance controlling the variation in structure of riverine phytoplankton assemblages. Weighted averaging regression and cross-calibration produced strong models for predicting DIN, water temperature (WT) and total suspended solid (TSS), which enabled the selection of algal taxa as potentially sensitive indicators: for DIN, Ulnaria ulna var. acus, U. ulna, D. communis and Euglena sp.; for WT: D. steligera, Scenedesmus dimorphus, D. communis and Euglena sp.; for TSS, Nitzschia sigmoidea, D. communis and Oscillatoria sp. The results from this relatively small survey indicate the need for further monitoring to gain a better understanding of riverine phytoplankton and to capitalize on the environmental indicator capacity of the phytoplankton community.

The classical ‘decoupled’ approach for fluvial flooding makes use of hydrographs as input boundary conditions. The catchment hydrology is determined by empirical semi‐distributed rainfall–runoff models, the flood processes by the use of hydrodynamic models. However, for urban floods, the distributed rainfall is set directly as input (‘direct rainfall modelling’ – DRM) to the elements of the 2D model. This ‘integrated approach’ aims to include hydrological and hydraulic processes in one single model. In this study, both modelling approaches are applied and evaluated for their suitability to determine flood hazards in small, rural catchments. The resulting flood maps and flow hydrographs are compared for selected rainfall–runoff events in a catchment located in Central Germany. In the first approach, the hydrological model (HEC‐HMS) from the Hydrologic Engineering Center (HEC) is used to generate the inflow boundary hydrographs for the 2D model (HEC‐RAS), which is then used to simulate the flow variables for the river network and its floodplains. For the second approach, the DRM is applied over the whole catchment by the use of HEC‐RAS. Special focus is given for the integrated approach to the difficulties occurring during the model optimisation and calibration. The comparison of the results and modelling processes of both approaches give insights into the advantages, disadvantages and difficulties or limitations of each presented approach.