Precipitation Effects for X- and Ka-band SAR

Space-borne Synthetic Aperture Radar (SAR) imaging is often considered to possess both day/night and all weather operational capabilities. Whereas the rst argument is true since we are dealing with an active sensor; the second does not hold in cases for which the operating frequencies are above 3 GHz. Indeed, the SAR performance can be signi cantly affected by atmospheric effects (losses), especially at unfavourable weather conditions. The principal reasons for the restriction on the use of these higher frequencies can be found in clear air losses (water vapour and oxygen), cloud attenuation and attenuation due to precipitation, primarily rain. 1 Tropospheric Effects The troposphere, as the lowest part of the Earth's atmosphere, reaches from the surface to approximately 12 km above ground and causes, amongst other effects, attenuation of traversing signals due to hydrometeors (rain, snow, hail), atmospheric gases, fog and clouds [2]. Except at low elevation angles, the attenuation of frequencies below 1 GHz is negligible. Insigni cant contributions to the attenuation will be obtained for frequencies up to 10 GHz due to fog and non-precipitating clouds. However the transmission spectrum exhibits peaks for frequencies around 22 GHz and 60 GHz due to molecular resonances from gases i.e. water vapour and oxygen. Whereas absorption effects due to atmospheric gases are present constantly and everywhere, attenuation due to condensed water in the form of precipitation, clouds and fog is infrequent and is limited to certain areas. Attenuation consists of two physical processes: the reduction of the wave's energy due to the heating of the water particles and, the scattering of energy away from the main direction of propagation. Further details on the assessment of atmospheric effects on SAR images can be found in [3]. 1.1 Modelling of Attenuation and Backscattering in SAR images For the modelling of the attenuation and backscattering effects in SAR images, Figure 1 is clarifying the underlying geometry [4]. The diagram provided at the bottom of Figure 1 shows the qualitative variation of the normalised radar cross section (NRCS) due to the idealised rain cell. The detailed modelling and calculation is provided in the following two sections. One of the major problems affecting microwave and millimetre wave bands for terrestrial and space-borne radars is the attenuation through rain [2]. A convenient way to describe the rain intensity is the so called rainfall-rate or rainrate given in millimetres per hour. This quantity refers to a certain ux of rain towards the surface of the Earth and may be measured e.g. by gauges or weather radars. A widely acceped empirical relation of the form