UiT the Arctic University of Norway is hiring a PhD fellow in geophysics (sea ice dynamics and data analysis).
- Organisation: UiT the Arctic University of Norway, Faculty of Science and Technology, Department of Geosciences
- Location: Tromsø, Norway
- Duration: 4 years temporary position
- Application deadline: 1 August 2017
About the PhD project:
The PhD project shall focus on the physical, mathematical and numerical modeling of flexure phenomena in sea ice and statistical analysis of relevant data. Sea ice can be modeled as a thin elastic or viscoelastic plate floating on top of an incompressible fluid. The ice motion itself contains information about the mechanical and physical properties of the ice, the sea, the seafloor, the atmosphere, and the overall geometry. Hence, measurement, analysis and interpretation of the ice motion may provide crucial information about important geophysical parameters of interest to various geophysical applications ranging from climate research to resource exploration.
The basic difficulty when connecting data from e.g. geophones situated on the ice to the actual ice motion, is that there is a plethora of different propagating wave modes that may co-exist and overlap in space, time, frequency and wavenumber. In addition, some of the modes are dispersive, while others are non-dispersive. The richness of co-existing modes calls for better models and improved statistical analysis tools. This PhD project will address this challenge by developing a complete catalog of possible wave modes, with a comprehensive description of their behavior in all possible domains. The modeling work will start from first principles, and the complexity of some of these wave models will require that numerical codes are developed and run on high-performance computational massively parallel computers.
A major contribution from this project will be to develop a suit of automatic software tools to separate the various wave modes through the use of state-of-the-art pattern recognition methods. The candidate shall develop optimal attributes for the pattern recognition. It is very likely that suitable attributes will be developed through time-frequency/space-time domains. The candidate shall develop variants of empirical mode decomposition and subspace projection methods suitable for mode separation of ice related modes. The basic idea is, thus, to connect the wave models with the statistical pattern recognition
To qualify the pattern recognition based wave mode separators, the candidate will use her/his own numerically simulated data, in addition to high-quality experimental data that we will obtain from collaborators at the University of Bergen.
The PhD project involves collaboration with a research group at the University of Bergen, and researchers at Stanford University in California.