Scattering from Random Rough Surfaces at X and Ku band for Global Remote Sensing of Terrestrial Snow

Snow water equivalent (SWE) retrievals based on X and Ku band radar observations have been studying over several decades. The physical basis of such approach is that snow volume scattering at X and Ku tends to increase with SWE. However, for small to moderate SWE, radar signals at X and Ku band are affected by surface scattering from the underlying soil, which degrades the sensitivity. It is important to estimate the soil rough surface scattering and subtract out its contributions to enhance the correlation of radar observations and SWE. In this paper, we present results of soil surface scattering based on the numerical solutions of Maxwell's equations in 3D (NMM3D) full wave simulations. In the past, full wave simulations have been limited with
$kh < 1.5$
. In plains, rms heights of rough soil surfaces are up to 3 cm, which is
$kh= 10$
at 17.2 GHz for Ku band. In this paper, we present NMM3D results for
$kh$
up to 15. We simulated two types of rough surface: “constant ratio” which means the ratio of correlation length to rms height are between 5 to 20, and “limited correlation lengths” meaning correlation lengths limited to 10 cm, Simulation results show that that backscattering increase with
$kh$
and then saturate at
$kh=2.5$
for rough surfaces with constant ratio. However, for rough surfaces with correlation lengths limited to 10 cm, backscattering first increase with
$kh$
to reach the peak and then decrease.