The aim of the modelling component of the NEGIS project is to model the long-term retreat of the NEGIS, from the continental shelf during the LGM until present-day. Using both onshore and offshore geomorphological and geological data to validate and calibrate the model, the project aims to develop a spatially robust understanding of the ice streams response to climate forcing and changing ice stream dynamics over 100 – 1000yr timescales.
This modelling study will be undertaken using the 3D BISICLES ice sheet model, which incorporates an adaptive mesh, thereby allowing the most dynamic regions of the ice stream, such as the grounding line and ice shelf, to be resolved at extremely high resolutions >1km, while maintaining lower resolutions closer to the ice summit where less activity is typically observed. By using this approach, critical components of the ice stream can be simulated at extremely high resolutions over 1000yr timescales while requiring relatively low computational power.
The input data utilised by the model during the initial setup comes primarily from the latest version of the BedMachine Database (v3) (Morlighem et al. 2017). This dataset incorporates present-day ice thickness, bed topography, and bathymetry for the whole of Greenland at 150m resolution. Ice velocity data is taken from Mouginot et al. (2017), which maps velocity from a range of satellite sources (Landsat-8, Sentinel-1, RADARSAT-2) at resolutions of ~15m for the whole of Greenland. Additional offshore data encompassing the limits of the eastern continental shelf is taken from the IBCAO database (Jakobsson et al. 2012).
Once the chosen region has been extracted from all input datasets, an inverse problem is run within the model to produce an initial map of basal friction. This is required to produce basal traction and velocity coefficients that match the thickness and velocity of the input data within the chosen domain. Air temperature within the model is also spun-up to produce an initial temperature that is in equilibrium with the underlying ice and vice versa. This provides an initial temperature which reflects the underlying ice, while spinning up the temperature prevents an initial ‘shock’ at the start of subsequent model runs, where the model reacts rapidly and unrealistically to new input data.
Once initial conditions have been produced, a series of parameters within the model must also be tuned to reflect a number of ice dynamics. In particular, attention is given to the surface and basal fluxes. In addition, a number of different calving laws are tested and tuned within the model, with the aim of producing an accurate pattern of ice shelf and grounding line migration. These calving laws range from relatively simple thickness based laws, to the physically based Benn calving law (Nick et al. 2017), which incorporates additional properties such as longitudinal strain and crevasses. When suitable tuning parameters have been combined with the BISICLES thermodynamic component, the model is run for several thousand years to produce an ice stream that advances to the continental shelf. From here, an array of forcing records from ice cores are applied to the model to simulate a range of retreat scenarios, tuned using the aforementioned tuning parameters to reflect geomorphological and geological data constraining ice stream history.
The final result will be a series of scenarios documenting the advance and subsequent retreat of the NEGIS throughout the LGM. This will build upon point specific chronologies taken from the NEGIS project and provide a long-term history for the ice stream. In addition, the completion of a sensitivity analysis on the chosen tuning parameters will provide an understanding of the key controls behind the ice dynamics of the NEGIS, and thereby provide a greater understanding of long-term ice stream behaviour as a whole.