Author(s): Wiktoria Cudna; Piotr Parasiewicz

Linked Author(s):

Keywords: Fish larval drift; Spatial modeling; Vistula River; Hydrodynamics; Logistic regression; Habitat suitability

Abstract: This study develops a predictive model for the spatial distribution of fish larval drift in the Vistula River, identifying the primary hydrological drivers of this critical ecological process. Research was conducted using nocturnal sampling during the peak drift season (April–July) near the ENEA Połaniec Power Plant, integrated with hydrodynamic modeling to interpolate depth and flow velocity. Larval density data were transformed into a binary presence/absence indicator based on a lower-quartile threshold of 0.044 larvae/m³. A logistic regression model was calibrated using AIC-based stepwise selection to determine the influence of environmental variables. The results identify water depth, flow velocity, and local morphological heterogeneity (depth variability) as significant predictors. The model achieved a success rate of 0.70 and an Area Under the ROC Curve (AUC) of 0.68, demonstrating its utility as a diagnostic tool for assessing hydrotechnical impacts on riverine larval niches.

DOI: https://doi.org/10.64697/iahr.proc.ise2026.abs.261

Year: 2026