Analysis and Prediction of Sea Ice Evolution using Koopman Mode Decomposition Techniques
Abstract
The program goal is analysis of sea ice dynamical behavior using Koopman Mode Decomposition (KMD) techniques. The work in the program's first month consisted of improvements to data processing code, inclusion of additional arctic sea ice concentration data, and identification of additional data sets of interest. We used recently developed improved KMD algorithms and windowing over 5, 10, 15, 20, and 30 year time periods to capture trends with shorter time scales that are less apparent in the full 39 year data set. The expected annual variation in sea ice concentration was clearly visible in the resulting Koopman spectra as well as a potentially novel slowly decaying mode that indicates a possible change in dynamical behavior beginning in the 1990's. Future work will include analysis of southern hemisphere sea ice concentration data for comparison, analysis of the spatial modes corresponding to eigenvalues of interest, and initial analysis of sea ice thickness data.
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 30, 2018
- Accession Number
- AD1051509
Entities
People
- Maria Fonoberova