Ice core methane and carbon monoxide as tracers of the terrestrial biosphere since the pre-industrial

Abstract

Ice core trace gas records have defined our understanding of carbon cycle feedbacks over the last 800,000 years. Yet, due to the slow process by which snow transforms into ice, few Antarctic ice core records have been able to capture highly resolved atmospheric information over the most recent centuries. Instead, low-resolution measurements of firn air have often been used to bridge the gap between the end of ice core gas records and the start of direct atmospheric observations. Explanations for suspected periods of multidecadal trace gas variability across the 19th and 20th centuries—superimposed on industrially driven increases in their concentration— have therefore remained challenging. In this thesis, multiple high snow accumulation Antarctic ice cores are utilised to continuously reconstruct the mixing ratios of atmospheric methane (CH4) and carbon monoxide (CO) over the last 200 years. An additional 53-year record of methane stable carbon isotope (δ13C-CH4) is also presented from the mid-20th century. These records provide important replication of the limited number of Antarctic trace gas records available for this period, and, with their high temporal resolution, offer new opportunities to explore the drivers of multidecadal trace gas variability since industrialisation. After providing the necessary context in Chapter 1, Chapter 2 presents an in-depth characterisation of the trace gas continuous flow analysis (CFA) system developed for the simultaneous detection of CH4 and CO in ice at the British Antarctic Survey (BAS). Chapter 3 presents three new high-resolution continuous CH4 records spanning from 1820 to 1995 CE. Data are directly compared to the seminal Law Dome ice and firn air record where a high degree of coherence, including a slowdown in CH4 growth rate in the mid 1940s and peak growth rates of up to 20 ppb yr-1 in the 1970s, attests to the fidelity of both works. A novel synchronisation algorithm utilising Bayesian inference is then used to fine-tune each gas age chronology. The CH4-synchronised chronologies are applied to the simultaneously derived CO records in Chapter 4. After accounting for a suspected mode of quasi-annual in situ production, the records are used to confirm a period of early 20th century atmospheric CO stabilisation first suggested in recent firn air modelling studies but notably absent from the output of state-of-the-art chemistry climate models. Using an atmospheric box model, this mode of variability is hypothesised to relate to human-induced suppression of natural fire regimes following rapid changes in 20th century land use and agricultural practices across the Southern Hemisphere. Chapter 5 focuses on the fidelity and implications of a new 53-year record of atmospheric δ13C-CH4 from 1941 to 1994 CE that suggests a greater rate of isotopic enrichment than previously reported. Mass balance calculations support previous top- down studies which indicate that historical anthropogenic fossil fuel emissions have been largely underestimated by bottom-up approaches. Then, to explain a plateau in isotopic enrichment in 1970 CE, biogenic emissions may have declined by up to 20 Tg yr-1 from 1940 to 1960 CE. If wetlands are assumed to be the source of these emissions, then this trend contradicts expectations based on natural climate drivers and instead may align with the historical pattern of anthropogenic wetland drainage. Ultimately, the unprecedented scale and rate of post-industrial land use change is proposed as a key factor in driving multidecadal trace gas variability for both CH4 and CO across the 20th century, disrupting or even decoupling natural terrestrial carbon-cycle feedbacks. Chapter 6 then outlines a preliminary investigation into new trace gas CFA methods while other notable contributions made during this work