High-head storage hydropower plants mainly operate their turbines during periods of high energy demand. The sudden starting and stopping of turbines (hydropeaking) lead to highly unsteady flow in channels and rivers. Besides hydropeaking, other anthropogenic actions and natural events such as sluice gate operations, flushing of reservoirs, debris jam and break up, ice jam and break up, sudden stopping and starting of turbines of runoff river hydropower plants, flashfloods or dambreaks can also cause highly unsteady flow. From an ecological point of view, hydropeaking consists in a non-natural disturbance of the flow regime. Possible mitigation measures aiming to reduce the effect of hydropeaking in a river downstream of the powerhouse can be divided into the installation of detention basins and the improvement of the river morphology. Morphological measures such as macro-roughness at banks might increase the flow resistance as well as the passive retention, which both increase the natural retention capacity of rivers and thus modify the form of the surge wave. In prismatic and nearly prismatic channels, highly unsteady flow conditions can be calculated using the elementary surge wave theory or numerical methods based on the Saint-Venant equations, respectively. For channels with large-scale roughness at the side walls, which may occur by the arrangement of particular morphological measures, no systematic experimental investigations have been done so far on the propagation of surge waves including downstream water-depth and flow velocity. The aim of this research project was to study how macro-roughness elements at the channel banks influence the unsteady flow conditions due to surge waves. 41 configurations of macro-rough banks and various discharges have been tested in a 40 m long flume with a bed slope of 1.14‰. A special experimental setup has been designed which is able to generate surge waves characterized by different discharge ratios. The first step of the experimental investigations focused on the determination of the flow resistance under steady flow conditions caused by large scale roughness elements at the channel banks, namely rectangular cavities (depressions). In a second step, positive and negative surge waves induced at the channel entrance have been tested in the same geometries. Five different discharge scenarios have been considered for each geometry. For comparison, experiments have also been performed in a prismatic reference channel without macro-roughness elements at the banks. The analysis of the experiments for steady flow conditions results in the following conclusions: The total head-loss is governed by different phenomena such as vertical mixing layers, wake-zones, recirculation gyres, coherent structures and skin friction. The flow resistance is significantly increased in the macro-rough configurations due to the disturbance of the bank geometry. Three different approaches have been considered in order to relate the additional, macro-rough flow resistance fMR to the forms of the banks by empirical formulas. The first approach is based on a powerlaw optimization, the second one on a model based on form drag and the third one uses the Evolutionary Polynomial Regression method. By separating the observed flow conditions in a square grooved, a reattachment and a normal recirculating flow type, the developed macro-rough flow resistance formulas are in good agreement with the laboratory experiments. Formulas of the second approach are suggested for practical applications in river engineering. Water body oscillations in cavities have been observed in axi-symmetric macrorough configurations. They lead to water-surface oscillations and transverse velocity components. Peaking excitations with heavy oscillations occur for Strouhal numbers values at 0.42 and at 0.84. The analysis of the experiments with unsteady flow, namely surge waves, shows: The elementary surge wave theory for the prismatic reference configuration including secondary waves and wave breaking could be verified in the laboratory flume. The absolute surge wave celerity Vw and the celerity c are lower for the macro-rough configurations. The decrease of the absolute surge wave celerity Vw is mainly due to the increased flow resistance and lies between 5% and 25% for both, positive and negative waves coming from upstream. Due to the dispersive character, the positive and negative surges from upstream are characterized by a sudden change (front), followed by a progressive change (body) of the water level. The sudden change can be clearly distinguished from the progressive change in the prismatic channel. In the channel with macro rough banks the separation is often hardly visible. The front height decreases between 5% to 25% along the laboratory flume for the prismatic configuration. In the macro-rough channel, the decrease can reach 70%. Behind the wave front of a positive surge, the water level continuous rising to a level which is higher than in the prismatic channel due to the channel bank macro-roughness. The decrease of the height of the surge wave front along the channel could be approached fairly good with an empirical relationship applicable for prismatic and macro-rough configurations. Numerical simulations based on the boundary conditions of the experiments have been done for steady flow tests (2-D simulations). They reveal that the integration of the turbulent stresses is required for the reproduction of the steady flow tests. Test cases with unsteady flow (1-D simulations) show that surge waves in rivers can be computed by integrating the macro-rough flow resistance in the source terms as well as by a modification of the continuity equation taking into account the passive retention.