Thermal structure of the Amery Ice Shelf from borehole observations and simulations

Abstract. The Amery Ice Shelf (AIS), East Antarctica, has a layered structure, due to the presence of both meteoric and marine ice. In this study, the thermal structure of the AIS and its spatial pattern are evaluated and analysed through borehole observations and numerical simulations. In the area with marine ice, a near-isothermal basal layer up to 120 m thick is observed, which closely conforms to the pressure-dependent freezing temperature of seawater. In the area experiencing basal melting, large temperature gradients, up to −0.36 °C m−1, are observed at the base. Three-dimensional (3-D) steady-state temperature simulations with four different basal mass balance datasets reveal a high sensitivity of ice-shelf thermal structure to the distribution of basal mass balance. We also construct a one-dimensional (1-D) temperature column model to simulate the process of ice columns moving along flowlines with time-evolving boundary conditions, which achieves slightly better agreement with borehole observations than the 3-D simulations. Our simulations reveal internal cold ice advected from higher elevations by the AIS’s tributary glaciers, warming downstream along the ice flow, and we suggest the thermal structures dominated by the cold core ice may commonly exist among Antarctic ice shelves. For the marine ice, the porous structure of its lower layer and interactions with ocean below determine the local thermal regime and give rise to the near-isothermal phenomenon. The limitations in our simulations identify the need for ice shelf/ocean coupled models with improved thermodynamics and more comprehensive evaluation of boundary conditions. Given the temperature dependence of ice rheology, the depth-averaged ice stiffness factor B(Th) derived from the simulated temperature field is presented to quantify the influence of the temperature distribution on ice shelf dynamics. The full 3-D field of this factor will assist as an input to future modelling studies.