We make use of a new high resolution dataset of ice elevation change that captures ice-mass loss north of 66°S to first show that non-linear uplift of the Palmer GPS station since 2002 cannot be explained by an elastic-only signal. We next apply a viscoelastic model with linear Maxwell rheology to predict uplift since 1995. We vary the thickness of the elastic lithosphere and upper mantle viscosity, and test the fit to the Palmer GPS time series showing that the observations may be closely fit with an upper mantle viscosity of less than 2 x 1018 Pa s, much lower than previously modelled. Comparison with vertical velocities from six GPS stations deployed after 2009 (the LARISSA network) verifies the results from the model. Subtracting the modelled uplift rate from the GPS gives insight into glacier loading changes over the late Holocene.
Despite the excellent model-data agreement, wider geophysical and laboratory studies suggest that a more complex rheological model may be required to correctly interpret the observed deformation. We report on investigations into adopting a linear Burgers or power-law rheology, and discuss the constraints placed on such models by this dataset and the potential implication for understanding grounding line stabilization through solid earth rebound.