What causes GPR signal scattering in temperate ice?

Abstract

<p>Ground penetrating radar (GPR) has been extensively used in glaciology to characterize englacial and subglacial properties such as ice thickness, liquid water content or water drainage pathways. However, GPR radargrams in temperate ice often appear blurred by undesirable noise, mostly appearing as highly scattered signal and strong attenuation. It is often suggested that this noise originates from englacial water inclusions, since water and ice  have a large contrast in their di-electric permittivity. Here, we use the open-access software gprMax to forward-model a GPR signal in temperate ice and to quantitatively assess the role that water scatterers have as noise source. In particular, we explore how different liquid water contents (LWC) and water-inclusions size affect the GPR signal, and show that their effect is much larger than the potential presence of a wet snowpack or a heterogenous distribution of ice permittivity. This confirms previous hypotheses suggesting that water inclusions within the ice are the main factor for the typical noise seen in GPR data for temperate ice. For simulations referring to 25 MHz GPR antennas, we find that a bulk LWC  equal or superior to 0.2 %, associated to decimeters-scale water inclusions, already constitute a limit in bedrock detectability for ice thicknesses larger than 100 m. These values of LWC and water-inclusion size are typical for Alpine field settings, clarifying why noisy signals are often encountered in such environments. Our numerical simulations for different LWC values also suggest than distinguishing between cold and temperate ice based on GPR scattering might be more difficult than previously assumed.</p>