The formation of scour holes is a vital thread of retaining the water infrastructure of rivers. Nowadays, protective filters are used to prevent washing away of the underlying material. Rijkswaterstaat started the program 'Stroomlijn’, in which trees (logs) are cut from the floodplains and placed in a sheltered zone of the river to test whether this would increase the flora and fauna. After some years, the logs might be re-used as bottom protection. This kind of bottom protection is formed by placing multiple logs next to each other, which together form a log carpet, and placing multiple log carpets on top of each other. This is called a log filter. Until now, no research has been carried out into the use of logs as bottom protection. The main function of a log filter is to reduce the hydraulic load so that the erosive capacity of the grains of the bottom will not be reached. Since a log filter has the same main function as a geometrically open filter, this research focuses whether the hydraulic load reduction of a log filter could be described by the current theory of granular open filter. In granular open filter theory, the hydraulic load is defined as the pore flow velocity and turbulent kinetic energy inside the filter layer. The hydraulic load is caused by the influence of the energy slope and the influence of the flow over the filter structure. The pore flow velocity and turbulent kinetic energy caused by the energy slope is always present, regardless of how thick the layer is. The influence of flow over the filter structure decreases exponentially and depends on the load damping length. In order to provide the hydraulic load reduction of a filter consisting of logs, experiments have been conducted in a flume of The Fluid Mechanics Laboratory at the Delft University of Technology. A filter consisting of logs has been mimicked and tested. The hydraulic load, i.e. increasing the depth average flow velocity and discharge, was increased until erosion occurred. During each increasing step, flow velocities above and inside the filter layer were measured. From the experiments, it is concluded that at least two layers are needed in order to ensure a significant decrease of the pore flow velocity and turbulent kinetic energy. Adding more than two layers will not lead to further reduction of the pore velocity or turbulent kinetic energy. The pore flow and turbulent kinetic energy beneath layer 2 are dominantly caused due to the influence of the energy slope. Hence, the equation describing the influence of the energy slope in a granular open filter is also appropriate in case of a log filter. However, the influence of the flow over the filter structure on the pore flow velocity and turbulent kinetic energy in a log filter does not decrease exponentially over depth, whereas in a granular filter it does. Therefore, we concluded that the current open granular filter models which describes the exponential decay of the pore flow velocity and turbulent kinetic energy due to the influence of the flow over the filter construction could not be applied in a log filter. In addition, we concluded that the flow in a log filter is very anisotropic. The flow characteristics differ in front, halfway and behind the log. In front of the log, stagnation is present, which leads to a vertical jet flow directed downwards. This flow causes erosion in front of the logs in the first layer. The erosion process in case of a log filter differs from granular open filters, which stated that erosion is predominantly caused by the hydraulic load, i.e. pore flow velocity and turbulent kinetic energy, in the filter and the bed pressure fluctuations due to low-frequency vortices. We also concluded that at locations where the porosity is higher, erosion occurred at lower hydraulic load conditions, i.e. discharge and depth-averaged flow velocity. This indicates that the porosity is an important variable when logs are used as bottom protection.