A surprisingly capable minimal bonded discrete element model for sea ice

Current climate models simulate sea ice as a continuous medium despite the discontinuities  arising from fracturing and floe-floe interactions. Discrete element models (DEMs) can directly resolve these discontinuities, making them valuable tools for understanding subgrid-scale sea ice dynamics. However, DEMs are typically limited by their high computational cost and large number of unconstrained microscale parameters, which hinder their validation and interpretability. In this work, we develop, calibrate, and evaluate the performance of a low-complexity, two-dimensional bonded DEM for process-based studies of ice flow and fracture. We modify the linear bonded particle model implemented in the molecular dynamics software, LAMMPS, to prevent failure in compression because unfractured sea ice is significantly stronger under compression than tension. Unlike most DEMs, our model does not explicitly integrate angular momentum, which halves the number of computations required for each particle. With this simplification, particles are frictionless and bonds do not break under torques. From simple shear experiments, we show that bonded elements behave elastically prior to failure, with an effective elastic modulus that scales linearly with the inter-particle bond stiffness. These experiments also illustrate that the ice deformation is localized in both space and time, in agreement with observations. Using a canonical geometry idealizing sea ice flow through the Nares Strait, we demonstrate that the model can reproduce ice arch formation and collapse previously observed in higher-complexity models. The small input parameter space can be explored with ensemble runs that would be infeasible for a higher-complexity model. We find that our model can represent four possible outcomes: viscous flow, ice arch formation and collapse, stable ice arch formation, and no fracturing. These regimes collapse onto a single control parameter given by the product of bond stiffness and critical strain. Despite neglecting key processes such as ridging and friction, our model reasonably represents short-timescale, discrete sea ice dynamics with fewer parameters to calibrate and lower computational cost than higher-complexity DEMs.