In the comcept of GPS-based bistatic radar, which is being explored to support terrain awareness and aircraft altimetry, ther reflection of the GPS signal from the surface of the earth plays an important role. To investigate the characteristics of these signal reflections, flight test data was processed to obtain a height estimate. Local terrain elevation data was added to this height estimate, making it possible to compare the result to the thruth reference from the flight. A positive bias of approximately 100 meters was found. It was suspected that the cause for this bias is that the reflection model that is used is too simple. The model assumes specular reflection on a terrain shape described by triangles between three elevation samples from a terrain elevation dataset. However, because of for example terrain roughness, diffuse reflection may occur in reality, causing the correlation peaks from the reflections to merge into one blob, making detection of the earliest reflection hard to impossible. As a result, the detected correlation peak will be later, leading to a larger estimate of the distance, and thus to a positive bias in the height. A simulation has been developed to investigate the effects of diffuse reflection. Results from this simulation show that even for moderate terrain profiles, the reflected beams spread out over an area, multiple times larger than the area the terrain profile represents. This confirms the suspicion that the reflection model that is used is too simple. Before the concept of GPS-based bistatic radar can be used for integrity monitoring or terrain awareness, it must be possible to estimate this a priori unknown bias.@en