Upper mantle viscosity structure and lithospheric thickness of Antarctica inferred from recent seismic models

Abstract

<p>Upper mantle viscosity structure and lithospheric thickness control the solid Earth response to variations in ice sheet loading. These parameters vary significantly across Antarctica, leading to strong regional differences in the timescale of glacial isostatic adjustment (GIA), with important implications for ice sheet models.  We estimate upper mantle viscosity structure and lithospheric thickness using two new seismic models for Antarctica, which take advantage of temporary broadband seismic stations deployed across Antarctica over the past 18 years. Shen et al. [2018] use receiver functions and Rayleigh wave velocities from earthquakes and ambient noise to develop a higher resolution model for the upper 200 km beneath Central and West Antarctica, where most of the seismic stations have been deployed. Lloyd et al [2019] use full waveform adjoint tomography to invert three-component earthquake seismograms for a radially anisotropic model covering Antarctica and adjacent oceanic regions to 800 km depth. We estimate the mantle viscosity structure from seismic structure using laboratory-derived relationships between seismic velocity, temperature, and rheology. Choice of parameters for this mapping is guided in part by recent regional estimates of mantle viscosity from geodetic measurements. We also describe and compare several different methods of estimating lithospheric thickness from seismic constraints.</p><p>The mantle viscosity estimates indicate regional variations of several orders of magnitude, with extremely low viscosity (< 10<sup>19</sup> Pa s) beneath the Amundsen Sea Embayment (ASE) and the Antarctic Peninsula, consistent with estimates from GIA models constrained by GPS data.  Lithospheric thickness is also highly variable, ranging from around 60 km in parts of West Antarctica to greater than 200 km beneath central East Antarctica. In East Antarctica, several prominent regions such as Dronning Maude Land and the Lambert Graben show much thinner lithosphere, consistent with Phanerozoic tectonic activity and lithospheric disruption. Thin lithosphere and low viscosity between the ASE and the Antarctic Peninsula likely result from the thermal effects of the slab window as the Phoenix-Antarctic plate boundary migrated northward during the Cenozoic. Low viscosity regions beneath the ASE and Marie Byrd Land coast connect to an offshore anomaly at depths of ~ 250 km, suggesting larger-scale thermal and geodynamic processes that may be linked to the initial Cretaceous rifting of New Zealand and Antarctica. Low mantle viscosity results in a characteristic GIA time scale on the order of several hundred years, such that isostatic adjustment occurs on the same time scale as grounding line retreat.  Thus the associated rebound may lessen the effect of the marine ice sheet instability proposed for the ASE region. </p>