Snow Mapping

9.4.2.1 Introduction

Estimation of snow cover and snow properties is important for input to hydrological applications such as modeling and forecasting runoff from snow melt as well as understanding changes in local, and regional climatic regimes. Typical snow parameters derived from radar data include, snow extent, snow water equivalence (SWE), and snow state (wet/dry).

The backscatter response from a snow covered surface is a function of numerous interrelated factors including the dielectric properties of snow, snow temperature, density, age, and snow structure. The backscatter received from a snow covered surface includes contributions from surface scattering at the air/snow interface, volume scattering from the snow layer, and scattering from the snow/ground interface. The extent to which backscatter is a function of surface scattering or volume scattering is governed by the properties of the snow. When a snow pack is dry (at a temperature less than 0°C) microwaves easily penetrate the snow (Figure 9-32) and the backscatter is largely a function of snow depth and snow density.

Figure 9-32
omega sub p = Penetration Depth (m)     mv = Volumetric Liquid Water Content (percent)

Figure 9-32. Snow penetration depth as a function of liquid water content and microwave frequency (from Ulaby et al 1986).

Depending on the microwave frequency and snow depth, backscatter from a dry snow pack may largely be a function of the ground surface characteristics underlying the snow pack due to the relative transparency of dry snow at microwave frequencies.

At C-band, wet snow is an absorber and dry snow is transparent making the estimation of SWE difficult. Polarimetry may help by providing additional information about the snow pack helping to improve the SWE estimate.

9.4.2.2 Polarimetric Signatures

One study [Sokol et al] used the C-SAR on the Environment Canada CV-580 to investigate the polarimetric properties of a snow pack. Figure 9-33 shows snow pit profiles for 4 dates during the winters of 1997-1998 and 1998-1999. Figure 9-34 shows Co-polarization Signatures derived from the C-SAR data for these sites on these 4 dates. It can be seen that on the wet snow date, March 6, the signature is indicative of a smooth surface pedestal height = 0.2) with little polarization dependence.

Figure 9-33

Figure 9-33. Snow pit profiles for 4 dates (From Sokol et al).

The polarization signature for the Dry Snow (December 1, 1997) shows a polarization signature with a higher pedestal (0.4) due to the penetration to the ground surface, which is rougher. In this case, the peak at the VV polarization is indicative of surface scattering. As the snow pack develops and horizontal ice layers form within it, the polarization signature changes (as seen on March 12, 1998). The surface appears rougher with a pedestal height of 0.6. Significant backscatter is seen to occur at the HH and VV polarizations. The polarization signature is significantly different on March 9, 1999 where significant backscatter is seen in the HH polarization with much less at VV.

These examples show that the polarimetric signatures make it possible to extract more information from imagery of the snow-pack. It is not yet clear how this might improve SWE estimation.

Figure 9-34. C-Band Co-polarization plots for selected snow packs derived from data acquired by the C-SAR on the Environment Canada CV-580 (from Sokol et al).

 

 

Whiz quiz
apple

Question: How does multi-polarization aid in wetland mapping?

The answer is...

 

 

Whiz quiz - answer

apple

Answer: The wetlands typically have vertically structured plants like rushes and sedges which can be separated using a combination of VV and HH responses. The cross-polarization can also help in delineating water from vegetated targets, especially under windy conditions during data acquisition.