Radar-Inference of the Basal Properties and Englacial Temperature of the Greenland Ice Sheet

Abstract:

Measurements of the basal properties (primarily the presence of basal melting, and bed roughness) are important for defining the lower boundary condition for thermomechanical models of ice sheets. Additionally, constraining the englacial temperature provides a test of the steady state solution of such models. Whilst direct measurements of the basal properties and temperature can only be made at a limited number of borehole sights, radio echo sounding provides a means to infer the spatial distribution of both throughout entire ice sheets. The radar-inference of basal melt is possible due to a wet basal interface having a ~10-15 decibel higher reflection coefficient than an ice-bedrock interface, and the radar-inference of englacial temperature is possible due to there being an exponential Arrhenius relationship with the radar attenuation rate.

In our study we use ~10 years of radio echo sounding data to map the spatial distribution of basal melt and depth-averaged temperature for the Greenland ice sheet. A necessary precursor to our investigation has been the development of a refined algorithm to infer the radar attention rate using the variation in bed-returned power with ice thickness. The algorithm is conditioned using a prior Arrhenius model calculation of the attenuation rate, and enables sample regions to be selected that are not significantly biased by an inhomogeneous bed. We demonstrate that, for the first time, this algorithm approaches the necessary accuracy to distinguish basal melt directly, with the uncertainty for the radar-inferred attenuation loss less than the decibel separation in reflection coefficients for wet and dry beds. We further cross-validate our method by using the radar-inferred attenuation loss as a constraint to predict the spatial distribution of geothermal heat flux, and to reconstruct temperature profiles that are closer to borehole measurements than zeroth order thermomechanical ice sheet models.