Uppermost crustal structure regulates the flow of the Greenland Ice Sheet

Abstract

The flow of the Greenland Ice Sheet is controlled by subglacial processes and conditions that depend on the geological provenance and temperature of the crust beneath it, neither of which are adequately known. Here we present a seismic velocity model of the uppermost 5 km of the Greenlandic crust. We show that slow velocities in the upper crust tend to be associated with major outlet glaciers along the ice-sheet margin, and elevated geothermal heat flux along the Iceland hotspot track inland. Outlet glaciers particularly susceptible to basal slip over deformable subglacial sediments include Jakobshavn, Helheim and Kangerdlussuaq, while geothermal warming and softening of basal ice may affect the onset of faster ice flow at Petermann Glacier and the Northeast Greenland Ice Stream. Interactions with the solid earth therefore control the past, present and future dynamics of the Greenland Ice Sheet and must be adequately explored and implemented in ice sheet models.

Extraction and applications of Rayleigh wave ellipticity in polar regions

“…The elliptcity curves were then inverted using the Monte-Carlo based Neighbourhood Algorithm (Sambridge, 1999) to generate 1-D seismic shear-wave velocity (Vs) profiles directly beneath each station. Jones and others (2021) showed that low seismic velocities correlate with large glaciers of Jakobshavn, Helheim and Kangerdlussuaq and regions of elevated heat flux as proposed by Martos and others (2018). Jones and others (2021) inferred that the low seismic velocities within the interior of the ice sheet are the subglacial expression of the Iceland hotspot track that traversed Greenland ∼80−50 Ma and controls the onset of fast ice flow for the North East Greenland Ice Stream and Petermann Glacier.…”

“…The method assumes that the Rayleigh wave particle motion is retrograde permitting the calculation of the back azimuth of the signal from the station to the source (Schimmel and others, 2011;Sergeant and others, 2013;Berbellini and others, 2019). Finally, the ellipticity of the extracted Rayleigh waves is estimated as a function of frequency and can subsequently be used in inversions for subsurface structure (e.g., Berbellini and others, 2019;Jones and others, 2021) 3. Current applications…”

“…Unique to the study of Jones and others (2021) was the use of ellipticity to characterise the subglacial geology as opposed to the ice-sheet seismic velocity or thickness as done in previous studies using H/V ratios (e.g., Lévêque and others, 2010;Picotti and others, 2017;Yan and others, 2018;Köhler and others, 2019;Preiswerk and others, 2019;Yan and others, 2020). Moreover, the study of Jones and others (2021) performed the calculation of horizontal-to-vertical amplitude ratios on extracted Rayleigh waves as a function of frequency to produce ellipticity curves as opposed to the traditional approach which calculates the spectral H/V ratio from windows of ambient noise.…”

“…Once estimated, the ellipticity measurements can be inverted for the 1-D vertical profiles of seismic velocities (Vp and Vs) and density directly beneath each station. The non-linear inverse problem is solved by minimising the difference between the measured ellipticity and theoretical estimates which are usually computed using the propagator method (e.g, Aki and Richards, 2002) or normal mode approaches (e.g., Berbellini and others, 2019;Jones and others, 2021). Alternatively, approaches based on a diffuse seismic wavefield can incorporate other seismic phases into the forward modelling allowing for H/V measurements made using the spectral amplitude method of Nakamura (1989) to be inverted (e.g., García-Jerez and others, 2016;Yan and others, 2018;Köhler and others, 2019;Preiswerk and others, 2019).…”

“…Alternatively, approaches based on a diffuse seismic wavefield can incorporate other seismic phases into the forward modelling allowing for H/V measurements made using the spectral amplitude method of Nakamura (1989) to be inverted (e.g., García-Jerez and others, 2016;Yan and others, 2018;Köhler and others, 2019;Preiswerk and others, 2019). Due to the non-linear nature of the inversion, a Monte Carlo approach is typically used to effectively sample the parameter space (e.g., Berbellini and others, 2019;Jones and others, 2021).…”