Advances in Understanding the Sea Ice Floe Size Distribution

Sea ice is a critical component of the polar climate system that is tightly coupled to the ocean and atmosphere. It is highly heterogeneous, composed of discrete floes which range in size across space and time. In this thesis, I use a combination of modelling and observational approaches to investigate how different physical processes determine the distribution of sea ice floe sizes. I construct the first global model that simulates floe sizes arising from the interaction of different physical processes. Floe sizes are modified by lateral melt, lateral growth, freezing together of floes and wave-ice interactions. By grounding process descriptions in underlying physics, observations of individual processes can be used to constrain model parameters. In light of the sparseness of floe size observations, I developed a novel methodology to constrain previously-unobserved floe freezing processes from in-situ observations. Results from global coupled sea ice–ocean model simulations are used to quantify the relative impacts of different processes on spatial and seasonal variability in the floe size distribution, providing hypotheses that could be tested by observational campaigns in the future. Under transient historical forcing, the model suggests that the fragmentation of Arctic sea ice has significantly increased over the satellite era. I also seek to improve understanding of feedbacks between sea ice floe size and the polar climate system. A fragmented ice cover exposes more ice area on the sides of floes to the ocean than sheet ice, promoting lateral melt, which reduces surface albedo. Conducting a statistical analysis of current climate models shows that inclusion of a lateral melt parametrization improves simulation of sea ice concentration relative to observations. However, calculation of lateral melt using the model for prognostic simulation of the sub-grid-scale floe size distribution results in little or no enhancement of lateral melt at a hemispheric scale compared to a simple parametrization, although it is likely to be important at smaller spatial and shorter temporal scales. The new model opens up the possibility of coupling sea ice and ocean surface wave models and of including floe size dependence in other processes, such as form drag, sea ice dynamics, ocean eddies and ocean–atmosphere heat transfer, which may result in significant impacts for polar climate