Snow Depth Extraction Based on Polarimetry and Multipass SAR Interferometry
Leinss, Silvan1; Hajnsek, Irena2
1ETH Zürich, SWITZERLAND; 2DLR Oberpfaffenhofen, GERMANY

Currently the determination of snow water equivalent (SWE) and other snow parameters rely on a network of weather stations, and air- and space borne missions with resolutions on the km-scale. High resolution weather and precipitation models will depend on high resolution input data. Active radar systems, especially SAR-systems can provide high resolution datasets independent of daylight. For a sufficient interaction of microwave radiation with the snow cover, high radar frequencies are needed. The satellite formation TerraSAR-X and TanDEM-X build for the generation of a high resolution globe-covering digital elevation model (DEM) (Krieger 2007) provides X-band data at 9.65 GHz with a resolution on the meter scale. Every time, flying over a certain area two acquisitions are taken which allow a DEM-generation for each pass (single pass interferometry). Differential interferometry (D-InSAR) is possible between acquisitions of different passes and polarimetric analysis (PolSAR) is possible when acquisitions are taken in different polarization channels. Here, an overview over results for snow height and snow water equivalent determination by using polarimetry, differential SAR interferometry of multi-pass and single-pass acquisitions will be presented.
Polarimetric phase differences between HH and VV polarization are detectable and change over the winter season. The measured differences show a strong correlation with ground-measured data and allow a determination of snow-depth with an accuracy of ±10 cm. The layered structure of snow is assumed to be the reason for the phase differences. A model will be presented to explain these phase differences. D-InSAR is a known method to detect height changes (Gabriel, 1989) on the wavelength-scale (λ= 3 cm) by comparing the measured interferometric phase with a reference, here a synthetically calculated phase, based on the best available digital elevation model (DEM). Atmospheric disturbances cancel out in single-pass interferograms, but remain visible as long-range phase patterns in multi-pass interferograms. Still, they cannot explain small scale phase patterns, which correlate with local topographic features. These phase patterns are caused by the changing penetration depth of microwaves but also by height deformations, both resulting in different location of scattering centers. Various properties of the cryosphere change over time and affect the location of scattering centers. Soil freezing, water content of snow, snow height and snow water equivalent but also vegetation cover are discussed to explain the detected phase patterns.
While the phase of the multi pass coherence is sensitive to height changes, its absolute value is an indicator for snow-fall events as well as for melting events, because the dielectric properties of snow and the configuration of scatterers undergo significant changes. For snowfall events, the drop of coherence correlates with the height of fresh snow.
In single pass interferometry, due to zero temporal difference between the two acquisitions, the bistatic mode of the TanDEM-X formation provides a very high coherence and phase accuracy. Therefore, elevation changes on the sub-meter scale can be detected, by comparing two obtained DEMs. This might not work in early winter when the snow layer is very cold (not conducting) and thin, but with some content of liquid water in the snow, a height difference should be possible to detect. For validation, ground measurements are essential and have been acquired. The test sites Sodankylae in northern Finland and Churchill in Canada, MB, have been chosen as already intensive ground measurements were done there within the framework of the CoreH2O mission (Rott, 2010). Snow height, snow water equivalent, air temperature, soil moisture, wind speed and even snow profile data are available. For both test-sites exists a set of 20 - 25 TerraSAR-X multi-pass acquisitions and 6, resp. 8 single-pass TanDEM-X acquisitions during the winter 2011/2012 and currently more data is taken over the same test site for the winter 2012/2013.

REFERENCES
Gabriel, Andrew K. et. al. 1989: ''Mapping small elevation changes over large areas: Differential radar interferometry,'' Journal of Geophysical Research, vol. 94, no. B7, pp. 9183 - 9191.
Krieger, G. e al. 2007: ''TanDEM-X: A satellite formation for high-resolution SAR interferometry,'' IEEE Transactions on Geoscience and Remote Sensing, vol. 45, no 11, pp. 3317-3341.
Rott, Helmut et al. 2010: ''CoreH2O: Cold regions hydrology high-resolution observatory for snow and cold land processes,'' Proceedings of the IEEE, vol. 98, no. 5, pp. 752-765.