Atmosphere Ocean Science Colloquium
Modeling emergent Arctic sea ice dynamics using small-scale discrete element methods within macro-scale equations
Speaker: Andrew Davis, NYU
Location: Warren Weaver Hall 1302
Date: Wednesday, March 11, 2020, 3:30 p.m.
Arctic sea ice extent has rapidly declined in recent decades. Despite studies that show significant responses in global climate to the changing Arctic, unknown physics, uncertain model parameters, and computational limitations make predictive sea ice models elusive. Sea ice is composed of discrete floes, deformable bodies whose small-scale interactions lead to the emergent Arctic basin scale dynamics. Discrete element models (DEMs) for granular media tracks each individual particle using contact laws to model the interaction, bonding, and fracture of adjacent floes. Our approach allows for mixed contact laws: particle-particle interactions may approximate constitutive laws in a peridynamic continuum or simulate viscoelastic contact between rigid bodies. We explore how model parameters such as particle geometry along with different contact laws affect the macro-scale dynamics. Additionally, computational limitations make explicitly modeling the hundreds of thousands of ice floes in the Arctic computationally prohibitive. We, therefore, derive macro-scale equations that model the basin scale dynamics and are computationally tractable because they only require small-scale DEMs. Furthermore, we show that under appropriate macro-scale model refinement that we can recover the granular nature of sea ice.