The Utrecht Finite Volume Ice-Sheet Model: UFEMISM (version 1.0)

Abstract. Improving our confidence in future projections of sea-level rise requires models that can simulate ice-sheet evolution both the and geological past. A physically accurate treatment large changes geometry a proper processes near margin, like grounding line dynamics, which turn high spatial resolution specific region, so small-scale topographical features are resolved. This leads to demand for computationally efficient models, where such be feasibly applied simulations 105–107 years duration. Here, we present evaluate new model solves hybrid SIA–SSA approximation stress balance, including heuristic rule grounding-line flux. is done on dynamic adaptive mesh adapted modelled during simulation. Mesh configured fine only at specified areas, as calving front or line, well point locations ice-core drill sites. strongly reduces number grid points equations need solved, increasing computational efficiency. allows resolve small geometrical features, outlet glaciers sub-shelf pinning points, significantly affect large-scale dynamics. We show reproduces analytical solutions intercomparison benchmarks schematic configurations, indicating numerical approach valid. Because unstructured triangular mesh, vertices increases less rapidly with than square-grid model, greatly reducing required computation time resolutions. simulation all four continental ice sheets an entire 120 kyr glacial cycle, 4 km expected take 100–200 wall clock hours 16-core system (1600–3200 core hours), implying this used high-resolution palaeo-ice-sheet simulations.