Predictive Accuracy of Fish Response to Two-Dimensional Modeling Using Field Data

Both water managers and researchers have the same goal when it comes to fish conservation that is to sustain, improve or restore the aquatic habitat. With this purpose two-dimensional hydrodynamic models have been widely used in aquatic habitat studies due to its high accuracy to simulate flow. The River 2D finite-element model is able to integrate the habitat suitability curves for fish with the simulated depth and velocity fields and the river-bed characteristics of substrate and cover, outputting the weighted usable area of fish (WUA), therefore predicting the potential distribution of fish species in the river reach. The fish WUA is currently used to plan engineered habitat restoration. However, little is known about the in situ variability associated to such prediction both for hydraulic and biological data, while ecological responses are known to be driven by many factors and thus noisy. To test the level of predictive accuracy of hydraulic and biological simulations, we modeled the habitat used by two species, the Iberian barbel Barbus bocagei and the Iberian straight-mouth nase Pseudochondrostoma polylepis, according to their preferences for depth, velocity, substrate and cover, and using new field results, we compared measured and simulated hydraulic and biological outcomes for different discharges using the River2D model. The results indicate that the 2D simulation is highly dependent on the density and location of points used for bed topography and therefore dependent on deviations related to data collection decisions. Substantial differences could be found for the biological response, only partly explained by the bed topography outliers. Likely, field habitat outcomes also result from other biotic and abiotic interactions occurring in the ecosystem, and restoration planning should be particularly aware of such variability.