Abstract
In a three years study (1997-99) in the Polish-German National Park Lower Oder Valley, juvenile fish recruitment was assessed on a meso-scale. The work aimed to test the hypothesis that the 0+ fish assemblage structure of a large, modified lowland river is primarily controlled by extrinsic factors. Altogether 5474 samples containing 134447 0+ fish were sampled in different mesohabitat types. Logistic modelling of eurytopic 0+ fish occurrence in 1997 and 1999 retained solely emergent vegetation and water depth as significant predictor variables. Plant cover and shallow depth were considered to offer refuge from elevated current velocities and predation. The occurrence of rheophilic 0+ fish was linked to a higher number of predictor variables, of which water depth, sandy substrate and current velocity were most important. In 1998, which was characterised by comparably low water levels, model building revealed that the need to avoid the open water column was prevalent when plant cover was lacking. Abundance patterns and model results were combined to assess the role of changing water levels. The timing and duration of high or low-flow events and the associated availability of suitable habitats was found to account for both seasonal and annual variation in local abundance: 1) Elevated water levels increased the quantity of flooded riverside marshes, thus favouring typical flood plain species. 2) A 100-year summer flood in 1997 caused detrimental ‘wash-outs’ of (mainly eurytopic) 0+ fish, especially in poorly structured mesohabitats. The progeny of rheophilic species were able to colonise newly created sandbar mesohabitats and, therefore, benefited from the increased habitat heterogeneity after the flood. 3) A short-term low-flow event in June 1998 led to a habitat bottleneck in terms of ‘absence of cover’, thus affecting the recruitment of late spawning species. 4) Under low-flow conditions occurring in summer 1998, groyne-heads were identified as potential habitats for juvenile rheophilic fish.

In conclusion, fish assemblage structure was determined by how stochastically changing water levels influenced the existing biotic context. Furthermore, projections of 0+ fish densities for high water level scenarios occurring in spring and summer, respectively, showed that recruitment of most species might be substantially enhanced, if the nowadays poldered flood plain was available.

More information is available at: "http://www.shaker.de/Online-Gesamtkatalog/Details.idc?ISBN=3-8322-0191-2">