The 250 ka centred on the Oligocene/Miocene boundary (23.03 Ma) were witness to a transient glaciation of Antarctica known as the Mi-1 event. Although a number of deep-sea and Antarctic sedimentary records exist from this period and suggest some relationship between ice volume and obliquity, the rapid rate of change in Antarctic ice volume coupled with the relatively low resolution (~1,000–5,000 years) of the available records means that the details and drivers of this event are still unclear. There is little to no evidence from the Southern Hemisphere mid-latitudes, and thus little knowledge of how the glaciation affected the climate further afield.
The Foulden Maar diatomite is a maar-lake deposit from the South Island of New Zealand. The deposit is finely laminated, and pollen evidence indicates a date in the early Miocene. I use the presence of orbital cycles within the physical properties records to show that individual pairs of light/dark laminae represent single years, and that the deposit covers ~100 ka. I then combine this information with 40Ar/39Ar dates from the base of the core and a magnetic reversal at ~106 m to show that the diatomite was being deposited during the period ~23.03–22.93 Ma, the peak and initial deglaciation phase of Mi-1.
The density and reflectance of the Foulden Maar deposit covary with Antarctic ice volume as inferred from Ocean Drilling Program (ODP) cores. During periods of greater Antarctic ice volume, conditions were cooler and drier at 45°S. Tropical influence on the site can be inferred from the presence of a half-precession cycle and significant variability in the El Niño–Southern Oscillation (ENSO) band. The ENSO-band variability persists throughout the depositional period, indicating that the increasing warmth during the Mi-1 deglaciation does not result in permanent El Niño-like conditions. While the physical properties measurements imply increasing warmth throughout the depositional period, I also infer four millennial-scale cool reversals.
The stomatal index data indicate a rapid increase in atmospheric carbon dioxide (pCO2) from <2x pre-industrial atmosopheric levels (PAL) to >3xPAL around 22.95 Ma, followed by a more gradual decrease back to around 2xPAL over the rest of the depositional period. This carbon dioxide pulse coincides with the initiation of rapid deglaciation in Antarctica, and suggests that pCO2 changes may have been connected to the deglaciation phase of Mi-1.
I use a number of techniques and lines of evidence for both the age model and the cyclostratigraphic parts of this study. Where only one technique is possible, evidence to support my conclusions is drawn from other sources. However, some of the stomatal index evidence is not statistically reliable, and more data is required before firm conclusions can be drawn.
A transient increase in pCO2 at ~22.95 Ma is correlated with the initiation of the deglaciation phase of Mi-1. Decreased Antarctic ice volume corresponded to warmer, more humid conditions throughout the region, suggesting coupling within the Southern Hemisphere system. The effects of ENSO variability were also widespread, and were not suppressed by increasing climatic warmth. The warming trend was interrupted by four abrupt millennial-scale cooling events, which may indicate that such events are characteristic of deglacial periods. This investigation shows that during the Mi-1 deglaciation, Southern Hemisphere climate was dominantly controlled by Antarctic ice volume, and thus ultimately by pCO2 and obliquity.
