Coupling framework (1.0) for the Úa (2023b) ice sheet model and the FESOM-1.4 <i>z</i> -coordinate ocean model in an Antarctic domain

Abstract. The rate at which the Antarctic ice sheet loses mass is to a large degree controlled by ice–ocean interactions underneath small ice shelves, with the most sensitive regions concentrated in even smaller areas near grounding lines and local pinning points. Sufficient horizontal resolution is key to resolving critical ice–ocean processes in these regions but difficult to afford in large-scale models used to predict the coupled response of the entire Antarctic ice sheet and the global ocean to climate change. In this study we describe the implementation of a framework that couples the ice sheet flow model Úa with the Finite Element Sea Ice Ocean Model (FESOM-1.4) in a configuration using depth-dependent vertical coordinates. The novelty of this approach is the use of horizontally unstructured grids in both model components, allowing us to resolve critical processes directly, while keeping computational demands within the range of feasibility. We use the Marine Ice Sheet Ocean Model Intercomparison Project (MISOMIP) framework to verify that ice retreat and readvance are reliably simulated, and inaccuracies in mass, heat and salt conservation are small compared to the forcing signal. Further, we demonstrate the capabilities of our approach for a global ocean–Antarctic ice sheet domain. In a 39-year hindcast simulation (1979–2018) we resolve retreat behaviour of Pine Island Glacier, a known challenge for coarser-resolution models. We conclude that Úa–FESOM is well suited to improve predictions of the Antarctic ice sheet evolution over centennial timescales.