In the framework of a research project focusing on mitigation measures for hydropeaking, a lateral embayment at the channel bank is studied as a fish refuge. Systematic experiments with different refuge configurations were carried out. The basic configuration is rectangular with a length of 2 m and a width of 1.2 m installed at the right bank of a 12 m long and 1.2 m wide flume supplied with freshwater from a natural river. In order to trigger water exchange between the flume and the rectangular refuge a wall acting like a groyne was installed inside the refuge protruding slightly in the main channel. Position, inclination and protrusion rate of this groyne were varied systematically in order to obtain an optimal water exchange and the best attractiveness of the shelter for fishes during hydropeaking. Each configuration was tested three times with juvenile wild brown trout (Salmo trutta fario) (0+ and 1+), with 2 different groups of 10 and the combined group of 20 brown trout. They where exposed during 3 hours each time to a hydropeaking flow of 220 l/s in the main channel. During every test, the movements of the fish were recorded continuously by video camera and their positions were observed every 20 minutes. 6 series of 20 fishes were used for 36 sequences corresponding to the 12 configurations tested. For each configuration the analysis of the fish positions gave a global frequentation rate as well as the favorite staying places in the shelter. Some in- and outgoing fish trajectories were obtained by the treatment of video pictures. A particular focus was given to the interface section between the refuge and the main channel in order to relate the spatial distribution and the frequency of fish passage from up- and downstream into the shelter. In order to link the swimming trajectories of the trout with the flow conditions, systematic measurement of the velocity field was performed using UVP technique. The flow velocities were analysed in several horizontal and vertical transects across the refuge and flume. Comparing the velocity patterns with the fish trajectories, the attractiveness of different configurations of fish refuges could be analyzed. The tests reveal that a very basic refuge configuration, with low water exchange between shelter and channel, is not interesting for fish. When forcing a water exchange by introducing a deviation groyne into the shelter, its frequentation can be increased significantly. The fish can easily detect the refuge by the exchange flux when searching its way upstream. The refuge attractiveness can be optimized by testing different groyne orientations, creating an expanded velocity field close to the exit and the entrance. Important is a high velocity field leaving the refuge at its lower end but also a backwater zone near the groyne. The high velocity field attracts the fish and the close backwater zone allows the fish to enter the shelter. For the best configuration, more than 80% of the fish found the refuge by swimming mainly from downstream, 20 minutes after the beginning of hydropeaking.