New North Atlantic palaeo-temperature reconstruction from terrestrial sedimentary archives: interrogating climates of the past with tools of today

Allison, Matthew ORCID: 0000-0003-1999-9814 (2026). New North Atlantic palaeo-temperature reconstruction from terrestrial sedimentary archives: interrogating climates of the past with tools of today. University of Birmingham. Ph.D.

Preview

Allison2025PhD.pdf
Text - Accepted Version
Available under License Creative Commons Attribution.
Download (3MB) | Preview

Abstract

This thesis investigates the preservation, reliability, and application of lipid biomarker archives for reconstructing past climates, with a focus on the late Miocene–early Pliocene Barmur Group (Tjörnes Peninsula, NE Iceland). First, an updated stratigraphic and structural model of the Barmur Group is developed using drone-based virtual outcrop mapping, refining the composite thickness to 442 m. This revised framework places further constraints on the age of the first opening of the Bering Strait, placing it c. 5.0 Ma, establishing a robust foundation for subsequent palaeoenvironmental reconstructions. Second, the Gaussian Process Emulator (GPE) - OPTiMAL, is developed further. The novel Contextualised Distance (CD) metric is introduced to quantify the analogue status of ancient biomarker distributions relative to modern calibration datasets, demonstrating that the isoGDGT system appears to satisfy the uniformitarian assumption, despite transient deviations during greenhouse climates. This GPE framework is applied to a 200 Ma record of isoGDGT data, the largest to date. Third, the Barmur Group lignites are interrogated to test the resilience of brGDGTs, a terrestrial biomarker, under high geothermal gradients. Burial and thermal modelling, integrated with vitrinite reflectance and supporting biomarker indices, reveal systematic diagenetic alteration of recovered brGDGTs yet preservation of interpretable palaeoclimatic signals. Finally, a combined machine learning workflow is employed in which a Gaussian Process Classifier predicts depositional setting of brGDGT assemblages, enabling targeted palaeotemperature reconstructions with a depositionally-specific GPE. These models reduce predictive uncertainty relative to conventional MBT/CBT indices and provide realistic estimates of warmest month air temperature and in situ water temperature, consistent with independent palaeoclimate proxies and regional records. Together, this thesis advances the methodological foundations of biomarker-based palaeothermometry and demonstrates the potential of Iceland’s Barmur Group to bridge laboratory-scale experiments with geological archives of climate change.

Type of Work:

Thesis (Doctorates > Ph.D.)

Award Type:

Doctorates > Ph.D.

Supervisor(s):