Master thesis defense by Thea Quistgaard

Title: "Laki to Tambora: Signal Restoration and Pattern Recognition in Ultra High Resolution Volcanic and Isotopic Signals"

Abstract: "A difficulty when studying data from ice cores, is the processes of densification and diffusion throughout the firn column. These processes attenuates some ofthe signal of the measured data, i.e. the water isotopic ratios or the chemical compositions of the ice. This work focuses on restoring the most likely signal in a given depth section of an ice core, through a method referred to as ’back-diffusion’. This method attempts to restore a signal by back diffusing it with an estimated diffusion length, σ. This diffusion length is affected by a number of different parameters, especially interesting by the temperature at deposition. So to give a qualified estimate on the temperature at a given time, it is of utmost importance to estimate the diffusion length as accurate and precise as possible. The restoration is achieved through a number of different computational methods, with a specific focus on a constrained optimization routine. This routine assumes a number of constraint over the depth series, especially the number of years expected in the section. This dating is made through Electrical Conductivity Measurements, where the two historically well documented volcanic eruptions of Laki (1783) and Tambora (1815) are visible. The method is tested on five different ice cores, from the Alphabet ice core series, all shallow cores drilled in the vicinity of the Crête ice core. The method has room for improvement, especially some of the simpler assumptions like the constraints and the optimization routine could benefit from further development. Moreover, the work carried out in this thesis leaves room for additional examination and development, but lays a good foundation to be used as a stepping stone in future research."

Supervisor: Vasileios Gkinis

Censor: Poul G. Hjorth

Participate by using the zoom link: https://ucph-ku.zoom.us/j/62483891593