Temperature reconstruction based on stable nitrogen and argon isotope measurements on Greenland ice cores

• Main Author: Döring, Michael
• Other Authors: Leuenberger, Markus Christian
• Format: Doctoral or Postdoctoral Thesis
• Language: English
• Published: 2020
• Subjects: 530 Physics; Greenland; Greenland ice cores
• Online Access: https://boris.unibe.ch/197938/1/doering20phd.pdf; https://boris.unibe.ch/197938/

Stable isotopes of the inert gases nitrogen (δ15N) and argon (δ40Ar) measured on ancient air extracted from polar ice cores provides the opportunity to reconstruct past temperature variations. Thereby, the temperature reconstruction is conducted by inverting firn densification and heat diffusion models to fit the δ15N and δ40Ar data. The main focus of this work is to develop automated procedures for solving that inverse problem. Automated fitting procedures have the advantage that manual tuning of parameters is avoided which leads to a better reproducibility and comparability between different studies. Furthermore, the use of automated algorithm saves working hours, which should not be underestimated. Another aim of this thesis is the reconstruction of the Holocene temperature history of Greenland summit using one of the automated methods developed in this study. First, an introduction into the topic is given in Chapter 1 containing a motivation of Holocene climate reconstructions, and an overview about a variety of reconstruction methods and studies. In section 1.2 the physical fundamentals needed for the understanding of this work are presented and in section 1.3 an overview about different methods of firn model inversion techniques are provided. Chapter 2 presents the automated firn model inversion methods developed in this work. In section 2.1 the suitability of a 2-dimensional polynomial transfer function for describing a firn model is investigated for using the model in static or dynamic behaviour. This was done in order to visualize the non-linearity of the problem. Section 2.2 presents an algorithm for the automation of the δ18Oice calibration technique. In sections 2.3 and 2.4 the fundamentals and implementations of a Monte-Carlo based automation technique independent from δ18Oice is described and evaluated on synthetic data for Holocene reconstructions and on Glacial data. In chapter 3 the results of using the Monte-Carlo based algorithm (sections 2.3 and 2.4) for fitting already published GISP2 ...