Master Thesis defense by Caroline Juelsholt

Title: Assessing the issue of the water isotope signal loss in the Beyond EPICA Oldest Ice core - A high-resolution data perspective

Abstract:

The Mid-Pleistocene Transition (MPT) took place between 1.25 million years ago and 700,000 years ago. This period marks the transition between 41,000 year glacial cycles to 100,000 year glacial cycles. The Beyond EPICA Oldest Ice (BEOI) project has provided a new deep ice core record spanning at least 1.2 million years which will be used to investigate the MPT. One of the pillars of BEOI is to measure the isotopic composition of the ice core. Stable water isotopes from ice cores have long been used as a proxy for past temperatures. However, the isotope signal is lost over time due to molecular diffusion making climate reconstruction difficult. Quantifying this signal loss is essential for determining how much information is preserved in the BEOI core. In this thesis, the challenge of quantifying the diffusion-driven attenuation of the water isotope signal in the BEOI core is addressed utilising 2.5 cm high-resolution discrete water isotope samples from seven different sections of the core. Red- and white noise power models are fitted to the power spectrums of δ¹⁷O, δ¹⁸O and δD to estimate diffusion lengths and noise levels. 

The dataset presented in this thesis contains 6160 water isotope samples, where 3080 samples were measured at the University of Copenhagen. All isotope values have been calibrated to the VSMOW/SLAP scale using local water isotope standards, and newly calibrated values for all local standards are presented. In addition, the second ever δ¹⁷O estimate is presented for the International Atomic Energy Agency reference material Greenland Summit Precipitation (GRESP) at -17.74 ± 0.02 ‰. 

The estimated diffusion lengths from the red- and white noise models broadly follows the same trend with the red noise estimates generally being 5-10 % larger than the white noise estimates. The diffusion lengths vary from around 6 cm in the shallower sections between 120 m and 360 m to 15-20 cm in the very deep sections of the core. Furthermore, the results indicate a strong signal attenuation with only millennial scale variability being preserved at the deepest parts of the BEOI core. The results from the deepest section of the core does not follow the expected increase in diffusion lengths with depth suggesting a disturbed stratigraphy below 2475 m. Further analyses are needed to estimate more accurately at which depth the stratigraphy is no longer continuously preserved.  

Supervisor: Vasileios Gkinis