Antarctica’s geothermal heat flow and its glacial isostatic adjustment response are critical to understand ice sheet stability. These demand a knowledge of the temperature of the Antarctic lithosphere, but challenges remain in resolving mantle thermomechanical properties. Here we use a two-stage process to resolve mantle temperature and composition. First, we derive an optimized relationship between shear-wave velocity and temperature, density, and composition, constrained by temperature, attenuation, and the average viscosity of the oceanic upper mantle. Applying this conversion relationship to a seismic shear-wave velocity model from adjoint tomography (ANT-20) yields an estimate of mantle density and temperature. In the second stage we apply a 3D finite-element-method gravity inversion to correct density distribution for the Antarctic lithosphere. The updated density structure further constrains the composition and so the temperature of the lithospheric mantle. Factoring in changes in mantle composition, areas with depleted mantle require a higher temperature than the initial estimate to fit the seismic velocity and density structure. Compared with a primitive mantle, the temperature in the depleted mantle is increased by up to 200 °C. From the updated temperature field, changes to the lithosphere thickness, mantle viscosity, and geothermal heat flow are defined: in East Antarctica, low viscosity area is largely unchanged (<1023 Pa s), while the estimated lithosphere thickness must decrease by up to 150 km, and heat flow must increase by 3–10 mW/m2. Collectively, the effects of an increased mantle temperature estimate suggest that a more dynamic and climate-responsive East Antarctic Ice Sheet is possible.

Crustal structure controls geothermal heat flux which is a key basal boundary conditions for ice-sheet flow in Antarctica. The crustal thickness of Antarctica is well resolved at large-scale, but the smaller-scale structures and density variations in the crust remain poorly constrained. Using 3D gravity inversion constrained by seismic Moho estimates, we model crustal structure in Antarctica, resolving sedimentary basin thickness and density, crustal density and internal layering, and the Moho. Spatial variations in upper crustal density are mapped to radiogenic heat production using a petrophysically-defined mapping approach. Significant variations are observed averaging 1.2 to 1.6 µW/m3, and as high as 2 µW/m3 in West Antarctica. The crustal contribution to geothermal heat flow is similarly variable averaging 18 to 27 mW/m2 and could be up to 60 mW/m2. The mapped variations are significant for correctly representing heat flow in Antarctica.