PhD Defense

The ongoing global warming raises the question of changing climate variability. In addition to modelling experiments, establishing the link between different past climate mean states and high frequency variability may be a precious source of information to understand present day climate observations and predict the future.

The EPICA Dome C (EDC) ice core has already permitted to reconstruct past temperature variations in East Antarctica over the last 800 thousands years (800 ky), thanks to measurements of water stable isotopes. In order to improve the depth resolution, originally obtained from 55 cm “bag samples”, the EDC core has then been continuously cut every 11 cm, providing new high resolution samples: “the fine samples”. If bag samples have allowed to cover a long climatic period, their low temporal resolution prevents the detailed study of past temperature changes, especially at sub-millennial scale. The measurements of fine samples thus promise to improve the temporal resolution by a factor of 5 for past periods of interest, and therefore to better document their climatic variability.

Here, we have focused on four interglacial periods. MIS 5 and 7, respectively occurring at ~125 and 240 ky ago, were chosen for their respective pecularities which are a strong intensity for the first one and a very short duration for the second one. MIS 11 and 19 were chosen for their orbital configuration, close to the present day one. Thanks to our new high resolution measurements, MIS 5, 7, 11 and 19 are now documented at respective temporal resolutions of 20, 30, 50 and 130 years. MIS 19 results are quite frustrating because they do not reveal any additional information about sub-millennial climate variability during this interglacial period. This finding reveals an unexpected large impact of isotopic diffusion, a physical process that progressively erases climatic information within the ice. This process is particularly enhanced in the deepest part of the EDC core because of a long residence time in warm ice (>-10°C). These results are of prime importance for the IPICS (International Partnership for Ice Core Science) “oldest ice” project, which aims to find the best Antarctic deep drilling site to go further than 1 million years ago.

The added value of measuring fine samples for characterizing sub-millennial climate variability during past interglacials is fully confirmed regarding MIS 5, 7 and 11 results. Each of these three interglacials indeed records a climate variability at multi-centennial to multi-decadal scale, as it was previously demonstrated only for our present interglacial: the Holocene or MIS 1. This variability is characterized by changes in amplitude which can be related to long-term trends. The increasing variability during cooling phases could be an early expression of processes driving glacial millennial variability. Within these interglacials, variability appears to be significantly linked with the rate of change of long term trends, with deceleration being linked with smaller variability (and the opposite for acceleration). Spectral analyses reveal that changes in variance coincide with changes in the dominant periodicities. This result highlights the parallel between variability changes and the establishment of new dynamic climatic structures.

So far, it remains difficult to clearly identify the climate mechanisms responsible for such variability changes during interglacial periods, because of: (i) dating uncertainties that impact the frequency analyses; and (ii) a lack of available records at similar resolutions, which are needed to document past natural forcings (such as changes in solar or volcanic activties) and internal feedbacks (e.g. oceanic or atmospheric circulation reorganisations). Our dataset is expected to provide precious information for comparison with transient Earth system modelling over several interglacials.