Hierarchical analysis of habitat use by 0+ juvenile fish in Hungarian/Slovak flood plain of the Danube River

Synopsis To address the lack of information on the distribution and habitat use of 0+ juvenile fishes in the Hungarian/Slovak flood plain of the middle Danube River, we undertook the first cross-border ichthyological investigation, examining three levels of ecological perception (hydrosystem, macrohabitat, microhabitat) during August 1992 using ‘Point Abundance Sampling’ by electrofishing. Being that the Gabćófcovo hydroscheme was about to begin diverting most of the river's discharge away from the flood plain during the winter of 1992, the present investigation represented the last chance to record the distribution and microhabitat use of 0+ fishes within the flood plain. At each sampling point, numerous environmental variables were measured quantitatively, or as percentages. At the hydrosystem level, 25 species of 0+ fishes were captured in the 1170 point samples collected from 52 sites (27 in Hungary, 25 in Slovakia), ranging from 10 to over 200 mm standard length (i.e. pikeEsox lucius). No significant differences were found between the Hungarian and Slovak specimens with respect to standard length (ANOVA, p> 0.31), nor in the relative densities (ind.m⁻²) of 0+ fish (Student's t-test: df 24, t = 0.601, p = 0.553). A typology of macrohabitats using principal components analysis of the sites X species data matrix in absence/presence revealed three groupings of sites: (1) lotic channels, weirs and wing-dams; (2) partially-abandoned channels; (3) abandoned channels; the results corroborated our assumption that weirs of the anabranch systems represent a quasi-lotic refuge for rheophilous 0+ fishes of the flood plain during late summer. At the microhabitat level, an empirical model of microhabitat use was generated using canonical correspondence analysis and association analysis (based on chi-square probabilities). Water velocity was the most influential variable, with the O+ juveniles ordinated along the first canonical axis according to their increasing rheophily. The second most influential microhabitat variable was water transparency, followed by the percentage abundance of macrophytes and substrate composition.