The use of airborne radar reflectometry to establish snow/firn density distribution on Devon Ice Cap, Canadian Arctic: A path to understanding complex heterogeneous internal layering patterns

Abstract:

The internal layer stratigraphy of polar ice sheets revealed by airborne radio-echo sounding (RES) contains valuable information about past ice sheet mass balance and dynamics. Internal layers in the Antarctic and Greenland ice sheets are considered to be isochrones and are continuous over several hundreds of kilometres. In contrast, internal layers in Canadian Arctic ice caps appear to be very heterogeneous and fragmentary, consisting of highly discontinuous layers that can be traced over only a few to several tens of kilometres.

Internal layers most likely relate to former ice surfaces (the upper few meters of snow/firn), the properties which are directly influenced by atmospheric conditions including the air temperature, precipitation rate, and prevailing wind pattern. We hypothesize that the heterogeneous and complex nature of layers in the Canadian Arctic results from highly variable snow and firn conditions at the surface. Characterizing surface properties such as variations in the snow/firn density from dry to wet snow/firn, as well as high-density shallow ice layers and lenses of refrozen water can help to elucidate the complex internal layer pattern in the Canadian Arctic ice caps.

Estimates of the snow/firn surface density and roughness can be derived from reflectance and scattering information using the surface radar returns from RES measurements. Here we present estimates of the surface snow/firn density distribution over Devon Ice Cap in the Canadian Arctic derived by the Radar Statistical Reconnaissance (RSR) methodology (Grima et al., 2014, Planetary & Space Sciences) using data collected by recent airborne radar sounding programs. The RSR generates estimates of the statistical distribution of surface echo amplitudes over defined areas along a survey transect. The derived distributions are best-fitted with a theoretical stochastic envelope, parameterized with the signal reflectance and scattering, in order to separate those two components. Finally, the signal reflectance and scattering components are used in a backscattering model to invert for the distribution of the surface density and roughness generating the observed surface echo amplitudes.