Prediction of snow permittivity and air temperature using received snow backscatter values over Greenland

The dry snow zone is the largest component of the Greenland ice sheet and is identified as the region that experiences no annual melt. Therefore, radar backscatter (σ) is expected to be relatively constant over time within the dry snow. However, annual variation was discovered in QuikSCAT data. This paper tests the hypothesis that annual variation in backscatter is caused by changes in permittivity. To do this, a model is provided relating permittivity to backscatter. Using this model, we test if received backscatter values can predict the permittivity and temperature of the corresponding snow. Both ML and MAP estimators are employed, and MAP is shown to have superior performance for the selected values of SNR. However, neither estimator is shown to consistently predict permittivity.