Ex situ experimentation to determine if introduced artificial habitat can provide alternative refuge to hazardous anthropogenic structures

Highly degraded lowland river ecosystems are of global concern to restoration practitioners. Hazardous anthropogenic structures, such as those used for water level management (i.e. pumping stations), present a mortality risk to fish and associated channelization, dredging, and removal of in-channel and riparian vegetation during winter dramatically reduces habitat availability. Paradoxically, fish seeking habitat for predator refuge in these systems can lead to ecological traps, that is, the undesired occupation of pumping stations. Artificial habitats installed upstream could provide safe alternative refuge, but the effectiveness of this restoration technique is poorly understood. Here, we uniquely quantified habitat occupancy and preference of a ubiquitous European freshwater fish (Rutilus rutilus) between an artificial reed bed and pumping station habitat, with access to open water in a tank experiment. Generalized linear mixed models revealed that fish preferred the pumping station when the artificial habitat was absent (baseline) and when it was introduced (pre-exclusion). Habitat management (exclusion from pumping station) was performed, during which artificial habitat occupancy was highest. When the pumping station was reintroduced (post-exclusion), pumping station occupancy probability decreased from 87.5% (pre-exclusion) to 3.7%, while artificial habitat occupancy probability increased from 18.4 to 87.9%. Therefore, our results demonstrate a preferential change in habitat occupancy of R. rutilus and suggest introducing artificial habitat alone may lead to restoration failures and ecological traps, stressing the need for habitat management to accompany artificial habitat restoration plans which aim to provide a safe alternative refuge for fish which occupy hazardous anthropogenic structures.