Ice cores are considered the gold standard for recording past climate and biogeochemical changes. However, gas records derived from ice core analysis have until now been largely limited to centennial and longer timescales because sufficient temporal resolution and analytical precision have been lacking, except during rare times when atmospheric concentrations changed rapidly. In this thesis I used a newly developed methane measurement line to make high-resolution, high-precision measurements of methane during the late Holocene (2800 years BP to present). This new measurement line is capable of an analytical precision of < 3 ppb using ~120 g samples whereas the previous highest resolution measurements attained a precision of ± 4.1 ppb using 500-1500g samples [MacFarling Meure et al., 2006]. The reduced sample size requirements as well as automation of a significant portion of the analysis process have enabled me to make >1500 discrete ice core methane measurements and construct the highest resolution records of methane available over the late Holocene. Ice core samples came from the recently completed West Antarctic Ice Sheet (WAIS) Divide ice core which has as one of its primary scientific objectives to produce the highest resolution records of greenhouse gases, and from the Greenland Ice Sheet Project (GISP2) ice core which is a proven paleoclimate archive. My thesis has the following three components.
I first used a shallow ice core from WAIS Divide (WDC05A) to produce a 1000 year long methane record with a ~9 year temporal resolution. This record confirmed the existence of multidecadal scale variations that were first observed in the Law Dome, Antarctica ice core. I then explored a range of paleoclimate archives for possible mechanistic connections with methane concentrations on multidecadal timescales. In addition, I present a detailed description of the analytical methods used to obtain high-precision measurements of methane including the effects of solubility and a new chronology for the WDC05A ice core. I found that, in general, the correlations with paleoclimate proxies for temperature and precipitation were low over a range of geographic regions. Of these, the highest correlations were found from 1400-1600 C.E. during the onset of the Little Ice Age and with a drought index in the headwater region of the major East Asian rivers. Large population losses in Asia and the Americas are also coincident with methane concentration decreases indicating that anthropogenic activities may have been impacting multidecadal scale methane variability.
In the second component I extended the WAIS Divide record back to 2800 years B.P. and also measured methane from GISP2D over this time interval. These records allowed me to examine the methane Inter-Polar Difference (IPD) which is created by greater northern hemispheric sources. The IPD provides an important constraint on changes in the latitudinal distribution of sources. We used this constraint and an 8-box global methane chemical transport model to examine the Early Anthropogenic Hypothesis which posits that humans began influencing climate thousands of years ago by increasing greenhouse gas emissions and preventing the onset of the next ice age. I found that most of the increase in methane sources over this time came from tropical regions with a smaller contribution coming from the extratropical northern hemisphere. Based on previous modeling estimates of natural methane source changes, I found that the increase in the southern hemisphere tropical methane emissions was likely natural and that the northern hemispheric increase in methane emissions was likely due to anthropogenic activities. These results also provide new constraints on the total magnitude of pre-industrial anthropogenic methane emissions, which I found to be between the high and low estimates that have been previously published in the literature.
For the final component of my thesis I assembled a coalition of scientists to investigate the effects of layering on the process of air enclosure in ice at WAIS Divide. Air bubbles are trapped in ice 60-100m below the surface of an ice sheet as snow compacts into solid ice in a region that is known as the Lock-In Zone (LIZ). The details of this process are not known and in the absence of direct measurements previous researchers have assumed it to be a smooth process. This project utilized high-resolution methane and air content measurements as well as density of ice, δ¹⁵N of N₂, and bubble number density measurements to show that air entrapment is affected by high frequency (mm scale) layering in the density of ice within the LIZ. I show that previous parameterizations of the bubble closure process in firn models have not accounted for this variability and present a new parameterization which does. This has implications for interpreting rapid changes in trace gases measured in ice cores since variable bubble closure will impact the smoothing of those records. In particular it is essential to understand the details of this process as new high resolution ice core records from Antarctica and Greenland examine the relative timing between greenhouse gases and rapid climate changes.