Data Assimilation Development and Arctic Sea Ice Changes
Description
The assimilation of data into physically-based computational models has emerged as one of the central themes of research in applied mathematics over recent years. The development of effective and efficient data assimilation (DA) schemes raises a number of fundamental challenges. This proposal will pick up this challenge with two contrasting, but interrelated directions of research:
A. Designing coherent structure based approaches to ocean DA. B. Developing DA schemes designed specifically to work with next generation sea-ice models.
The first direction (A) constitutes a new approach to dealing with the fundamental conundrum of DA, namely that the two basic strategies, coming from optimization and statistics, work well for high-dimensional and nonlinear systems respectively. But neither approach works effectively for both. The new approach being developed here will initially be focused on the all-important issue of assimilating data from Lagrangian observations of the ocean. The idea rests on the extensive past work, supported by ONR, on Lagrangian Coherent Structures.
The goal of the second direction (B) will be to bring advanced DA schemes to bear on the problem of determining how and when Arctic sea-ice will break up. The changes precipitated by a warming Arctic Ocean are already evident and it will be critical in the coming years to understand, and be able to predict, the location of open water and the presence of degrading ice. It is only the next generation sea-ice models, which account for material properties of the ice, that will be in apposition to properly capture these events. At the same time, data from observations will be critical to guide and correct these models. DA for these models, however, poses a fundamental challenge as the underlying numerical solvers are both Lagrangian and adaptive-mesh based.
The full picture of Arctic sea-ice brings up other issues attendant to the overall dynamics of the Arctic. These include the following topics to be considered as part of this overall effort: (i) the degrading of the ice pack by waves in the marginal ice zone, (ii) how much cold water is transported at depth away from the Arctic, which is believed to be measurable as outflow through a small number of chokepoints surrounding the Arctic basin, and (iii how warm water, originating in the tropics, is transported northward and such meteorological features as hurricanes are thought to play a key role.
RENCI's Role
The project is led by RENCI personnel Christopher Jones, Domain Scientist in Data Assimilation.