ESTIMATING AND ACCOUNTING FOR DIFFUSION IN DEEP ICE USING ADVANCED STATISTICAL METHODS
Supervisors: T. Laepple (AWI, D), V. Gkinis (Univ. of Copenhagen, DK)
Academic secondment: Univ. of Copenhagen (DK); Non-academic secondment: Risk Management Solutions (UK)
Abstract: The isotope record in the deep part of the oldest ice core is expected to be strongly affected by diffusion. Therefore, the isotopic record of millennial climate variations and potentially even glacial-interglacial changes is expected to be heavily smoothed. Accounting for the diffusion and mixing processes will thus be necessary to recover the amplitude of the climate variations (back-diffusion) as well as to provide information which climate variations were consistent to the measured isotopic variations (probabilistic reconstruction). The proposed PhD project will work on this challenge by using and developing advanced statistical techniques for the estimation of the diffusion length, the statistical characterization of the measurement process and deconvolution of the deep ice-core signal. It will complement the other PhD projects proposed in the ITN working on this question from a measurement or firn/ice- modeling perspective.
Keywords: climate reconstruction, signal recovery, statistics, firn and ice diffusion, water isotopes