Optimal conditions for wet snow detection using RADARSAT SAR data

This paper presents simulation results of the backscattering coefficient, in order to discriminate between wet snow and dry snow covers sensed at 5.3 GHz by the RADARSAT Synthetic Aperture Radar (SAR) sensor. Snow-field measurements coinciding with the RADARSAT SAR overpasses are used to explore and set out optimal conditions for wet snow detection, as a function of the sensor incidence angles. The conditions concern wet snow surface characteristics, mainly the roughness represented by the surface slope m and the volumetric liquid water content, snwc (vol.%). Based on the 3-dB threshold value used in several wet snow detection algorithms, the results show that in order to be discriminated from dry snow covers, wet snow surfaces must be characterized as: (a) m≤0.058 and snwc≤1.1, if the sensor operates in the S1 mode (20-27° incidence angle range), and (b) m≤0.082 and snwc≤3.0, if the observations are made in the S7 mode (45-49° incidence angle range). For the identification of a very wet snow, it is also shown that the S7 mode of RADARSAT SAR sensor is more suitable than the S1 mode. The latter, however, provides better discrimination for low values of the snow liquid water content. Furthermore, for wet snow detection based on modeling, the present paper demonstrates the importance of using the appropriate methodology to assess the dielectric constant of the background medium.