Multiscale Homogenization for Sea Ice
Abstract
The precipitous loss of Arctic sea ice observed in the past few decades has far reaching impact on the polar marine environment as well as more broadly on Earth s climate system. This reduced ice cover also opens up new opportunities for navigation, exploration and extraction of energy and food resources, and increased human activity in general. Along with the Arctic opening up goes the potential for more interactions and competition between countries. Thus it is of strategic as well as scienti?c interest to improve projections of how the sea ice cover may evolve in the future, and to develop more rigorous representations of sea ice in predictive models of the climate system and marine environment. One of the fundamental challenges of modeling sea ice and climate is to account for important processes that occur on scales ?ner than the coarse grids of numerical models, and how they in uence larger scale e -ectie behavior of the system. This linkage of scales is particularly interesting, relevant ???? and challenging ???? for sea ice, as it exhibits composite structure on length scales ranging over many orders of magnitude. Millimeter scale brine inclusionsare laced throughout the ice, and coalesce to form meter scale channels through which fluid can flow. Sea ice has centimeter scale polycrystalline structure which helps determine its bulk fluid ow and mechanical properties.Convective brine ows with meter scale structure a -ect bulk thermal and nutrient transport. Sea ice oes ranging from a few centimeters to tens of kilometers form the microstructure of the sea ice pack, viewed from an aircraft or satellite asa time-evolving composite of ice and ocean. The Arctic sea ice surface in late spring and summer is a complex mosaic of snow and pools of melt water on meter to kilometer scales, which determine sea ice albedo, a critical parameter in climate models. Here we propose to develop powerful methods of homogenization, over a broad range of scales as indicated above, to accurately account for important sub????grid scale sea ice structures and processes in climate modeling. We consider several key issues critical to advancing predictive capability, where mathematics of homogenization and statistical physics can provide a rigorous framework for analysis and computation. They include: ? Advection diffusion processes. Develop bounds and spectralmeasure computations for advection enhanced diffusivity, such as the thermal conductivity of sea ice in the presence of convection, or the diffusion coe?cient of a tagged oe in the ice pack. Develop inverse methods for recovering diffusion and velocity ?fields in large scale, effective advection diffusion models of sea ice evolution. ? Ocean waves in sea ice. Develop bounds on the effective complex viscoelasticity of the sea ice layer for wave propagation in the the marginal ice zone (MIZ), via Stieltjes integrals with spectral measures which depend on oe geometry and configurations. ? Low order predictors. Explore the use of simplified partial differential equation models for analyzing and predicting fundamental characteristics of the sea ice pack, such as MIZ width and sea ice concentration in unobserved regions. ? Statistical physics of melt ponds. Analyze critical exponents of melt pond evolution as long range order or connectivity develops, such as the correlation length exponent. Develop an Ising model of melt pond geometry that incorporates ice-albedo feedback, and investigate other low order models of melt pond and albedo evolution.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jul 26, 2018
- Source ID
- N000141812552
Entities
People
- Kenneth M. Golden
Organizations
- Office of Naval Research
- United States Navy
- University of Utah