Radiative Transfer Model for Cross-Polarization of X-&C-band Radar Observations on Remote Sensing of SWE of Mountainous Snow
Abstract
Recently there are interests in using C band Sentinel 1 radar observations to infer snow depth and snow water equivalent (SWE) for mountainous snow. The Sentinel 1 data show sensitivity at cross-polarization for snowpack larger than 1 meter which contradicts traditional theories that snow volume scattering at C band is small and much lower than surface scattering by the snow/soil rough interface.
For mountain snow, due to the sequential periods of snow accumulation, ice grains in snow are densely packed and can adhere to form aggregates or clusters. In the past, snow microstructure was modeled by spherical grains with single grain size as well as multiple sizes with different distributions. The stickiness was introduced to facilitate aggregation of spheres. In this paper, we use the bicontinuous model in which the amount of aggregation can be controlled. The aggregates are irregular shape that can increase the amount of cross polarization that can be much larger than the prior stickiness approach. Aggregates which are densely packed also have larger diameter than the single ice grain. The effects of aggregates can account for stronger than expected co- and cross- polarization volume scattering at C and X band. It is important to apply the forward model to analyze the radar signatures of snow with aggregates. Although the C band co-polarization volume scattering is increased, it can be still be lower than the co-polarization of rough surface scattering. However, the C band cross polarization is larger than that of surface scattering. In this paper, we propose a RT model for snow with dense aggregates to analyze the radar backscatter at C and X band. The effects of densely packed aggregates of ice grains has been modeled with the bicontinuous media. Scattering properties of snow with aggregates are fully coherently computed with the numerical solution of Maxwell equation in 3D (NMM3D), including extinction coefficients, phase matrix, and scattering albedo. By substituting these scattering properties into the RT equation, the backscattering of snowpack is obtained. In this paper, the C band Sentinel 1 data and SnowSAR (X and Ku band) from SnowEx 2017 are analyzed based on co-located ground measurements. The RT model is then calibrated and validated with radar observations. The snowpack with aggregates has larger backscattering for both co- and cross-polarization, which are in good agreement with radar data. Especially, snow volume scattering of C and X band at cross-polarization is larger than the scattering from underlying ground, showing sensitivity to SWE.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2020
- Bibcode:
- 2020AGUFMC007...01Z
- Keywords:
-
- 0736 Snow;
- CRYOSPHERE;
- 0740 Snowmelt;
- CRYOSPHERE;
- 0794 Instruments and techniques;
- CRYOSPHERE;
- 1863 Snow and ice;
- HYDROLOGY