A method for surface roughness parameter estimation in passive microwave remote sensing

Surface roughness parameter is an important factor and obstacle for retrieving soil moisture in passive microwave remote sensing. Two statistical parameters, root mean square (RMS) height (s) and correlation length (l), are designed for describing the roughness of a randomly rough surface. The roughness parameter measured by traditional way is independence of frequency, soil moisture and soil heterogeneity and just the “geometric” roughness of random surface. This “geometric” roughness can not fully explain the scattered thermal radiation by the earth’s surface. The relationship between “geometric” roughness and integrated roughness (contain both “geometric” roughness and “dielectric” roughness) is linked by empirical coefficient. In view of this problem, this paper presents a method for estimating integrated surface roughness from radiometer sampling data at different frequencies, which mainly based on the flourier relationship between power spectral density distribution and spatial autocorrelation function. We can obtain integrated surface roughness at different frequencies by this method. Besides “geometric” roughness, this integrated surface roughness not only contains “dielectric” roughness but also includes frequency dependence. Combined with Q/H model the polarization coupling coefficient can also be obtained for both H and V polarization. Meanwhile, the simulated numerical results show that radiometer with a sensitivity of 0.1 K can distinguish the different surface roughness and the change of roughness with frequency for the same rough surface. This confirms the feasibility of radiometer sampling method for estimating the surface roughness theoretically. This method overcomes the problem of “dielectric” roughness measurement to some extent and can achieve the integrated surface roughness within a microwave pixel which can serve soil moisture inversion better than the “geometric” roughness.