Quantitative analysis of Mars surface radar reflectivity at 20 MHz

The radar surface echoes taken from SHARAD observations are extracted to obtain a reflectivity map covering almost half of the martian surface, and then compared to available roughness maps. Then, we used a 2-step method, based on a stochastic description of reflectivity, in order to (i) separate the coherent/incoherent components of the signal by means of a probability density function fitting of amplitude distributions, and (ii) express these components with respect to roughness/permittivity values by adapting common backscattering models to the nadir case. Scattering is found to be the most important process dominating reflectivity over the martian terrains. Reflectivity is best correlated with roughness maps derived from the slope parameter. The stochastic behavior of reflectivity is confirmed by the excellent agreement of our models and demonstrates the low probability of retrieving permittivity from scattering surfaces. The non-stationary nature of most martian terrains is consolidated by a very good fit to K-distributions, justifying the use of the fractal theory for describing surface roughness on Mars. A few slightly-rough regions exhibit both a coherent and incoherent component. When surface correlation length is non-significant within the resolution cell, we show that the RMS roughness at a decameters baseline can be derived with sub-decimeter sensitivity without prior signal calibration, improving the capabilities of SHARAD in determining surface roughness with a view to landing site selection. Sets of derived dielectric constants are obtained and analyzed with regard to the reference signal used for calibration. Given the very different implications of each set to the composition of martian terrains, we emphasize the importance of good signal calibration as a major issue for SHARAD and the next interplanetary radar missions. (C) 2012 Elsevier Inc. All rights reserved.