Inferring the crystal orientation fabric of the Northeast Greenland Ice Stream using polarimetric radar data

The crystal orientation fabric (COF) of glacial ice strongly influences its mechanical properties and evolves with flow. Large-scale ice flow models currently neglect COF evolution, but including it is becoming increasingly feasible. However, observations are still very sparse and often depth-averaged or point measurements. Radars, and especially polarimetric radars, are sensitive to the COF due to the birefringence of ice and provide a relatively easy way to collect observations that can be used to infer the anisotropy and orientation of the COF. In this study, we formulate the problem of inferring the COF from polarimetric radar data as an inverse problem to derive depth-resolved horizontal COF anisotropy. The method is applied to polarimetric radar data from the Northeast Greenland Ice Stream (NEGIS), where previous methods have struggled due to the high anisotropy. The method relies on an iterative linearization of the Fujita radio-wave depolarization matrix model to estimate COF orientation and scattering anisotropies. It also employs a linear maximum likelihood solution to derive eigenvalue differences from travel-time anisotropies. The inversions generally recreate the observed power anomalies and reveal a strong increase in horizontal anisotropy at shallow depths in NEGIS, followed by a rapid decrease near the ice stream base, likely due to recrystallization processes. The inversion also shows a near flow-aligned COF close to the onset of the ice stream, with increasing misalignment along a 30 km flowline downstream.