Synthetic aperture radar (SAR) produces high-resolution images of the Earth's surface (and subsurface, under some conditions) by coherently processing the returns from a pulsed radar. It is an active sensor, hence is not dependent on natural illumination for its operation, and the comparative transparency of the atmosphere at centimetre wavelengths allows penetration of clouds and haze. SAR therefore offers considerable potential as an all-weather, all-time-of-day sensor, as long as propagation effects do not destroy signal coherence. A particularly useful way of viewing SAR performance is in terms of the synthetic antenna gain pattern, and performance measures derived from this pattern (e.g. spatial resolution, sidelobe energy). Perturbations of the phase due to irregularities in the refractive index structure of the ionosphere and troposphere lead to changes in the system performance. We have used a phase screen model to generate phase perturbations of the form expected due to propagation through a disturbed ionosphere. Simulations have been undertaken and used to assess the performance of a variety of SAR systems under particular geophysical conditions. These simulations indicate that at the longer wavelengths (C band and above), severe disturbances may greatly reduce image contrast, to the extent of destroying the image. Geometric distortions will also be introduced.