Ice-sheet/ocean interaction in narrow fjords around Greenland is one of the key outstanding challenges in modeling studies of climate change and sea level rise. The runoff from Greenland's ice-sheet (GiS) is a significant factor in the regional ocean dynamics and global sea-level rise, yet present-day climate models are not able to resolve fine-scale processes in the fjords without prohibitive computational costs. This is due to orders of magnitude difference of spatial scales between the open ocean (~1000km) and fjord (<1km) as well as complicated bathymetry and coastline between the ocean outside fjord and a marine-terminating glacier.

The goal of the NUMO project is to develop a non-hydrostatic ocean model with an unstructured grid and local non-conforming mesh refinement, able to resolve fine scale ice-sheet / ocean interactions, the 3-dimensional circulation within Greenland fjords and exchanges with the ocean outside. As a proof-of-concept, we focus on Sermilik Fjord and its interaction with Helheim glacier and the sub-polar North Atlantic Ocean. The fine scale and most critical processes at an ice-sheet / ocean interface are modeled using a high-resolution, locally refined mesh. An unstructured mesh is used to realistically represent the geometry of the fjord, while in the areas of particular importance (i.e. glacier front) the resolution is increased by non-conforming mesh refinement. This approach allows us to transition between coarse and fine mesh quickly and robustly, ensuring relatively small overall computational cost of the simulation. Boundary conditions are prescribed from observations and passed to the model via CPL7 coupler. The model is designed such that it can be easily coupled with existing Earth System Models (ESMs) through a flux coupler. We present results of NUMO on standard ocean test cases, demonstrate the unstructured mesh capability and discuss the approach to high-resolution Sermilik Fjord simulation. The long-term goal is to simulate all Greenland’s fjords and adjacent coastal ocean and couple this simulation to a regional or global ESM.