The reconstruction of ground surface temperature for the last 500 years, obtained at the deepest mountain permafrost borehole of the world (Stelvio Pass, 3000 m a.s.l., Italian Alps), is presented here. The main difference with respect to MAAT reconstructions obtained through other proxy data for all of Europe relates to post Little Ice Age (LIA) events. Indeed at this site a stronger cooling of ca 1 °C between 1940 and 1989 and even a more abrupt warming between 1990 and 2011 was detected.

In this work, we apply a range of methods to measure the geochemical composition of the calcite from fossil shells of Pycnodonte vesicularis (so-called honeycomb oysters). The goal is to investigate how the composition of these shells reflect the environment in which the animals grew. Ultimately, we propose a methodology to check whether the shells of pycnodonte oysters are well-preserved and to reconstruct meaningful information about the seasonal changes in the past climate and environment.

The present numerical experiments undertaken by a coupled atmosphere–ocean model indicate that the uplift of the Tibetan Plateau alone could have been a potential driver for the reorganization of Pacific and Atlantic meridional overturning circulations between the late Eocene and early Oligocene. In other words, the Tibetan Plateau could play an important role in maintaining the current large-scale overturning circulation in the Atlantic and Pacific.

We present a novel approach for ice-core-based temperature reconstructions, which is based on gas-isotope data measured on enclosed air bubbles in ice cores. The processes of air movement and enclosure are highly temperature dependent due to heat diffusion in and densification of the snow and ice. Our method inverts a model, which describes these processes, to desired temperature histories. This paper examines the performance of our novel approach on different synthetic isotope-data scenarios.

The Eocene--Oligocene transition was a major cooling event 34 million years ago. Climate model studies of this transition have used low ocean resolution or topography that roughly approximates the time period. We present a new climate model simulation of the late Eocene, with higher ocean resolution and topography which is accurately designed for this time period. These features improve the ocean circulation and gateways which are thought to be important for this climate transition.

We studied the impact of the six largest tropical eruptions in reference to Australian precipitation, the Indian Ocean Dipole (IOD), and El Niño–Southern Oscillation (ENSO). Volcanic forcing increased the likelihood of El Niños and positive IODs (pIOD) and caused positive rainfall anomalies over north-west (NW) and south-east (SE) Australia. Larger sulfate loading caused more persistent pIOD and El Niños, enhanced precipitation over NW Australia, and dampened precipitation over SE Australia.

We infer the physical mechanisms by which the mid-Pliocene could have sustained a warm climate. We also provide a mid-Pliocene perspective on a range of climate sensitivities applicable on several timescales. Warming inferred on the basis of these sensitivity parameters is compared to forecasted levels of warming. This leads us to conclude that projections for 300–500 years into the future underestimate the potential for warming because they do not account for long-timescale feedback processes.

Subduction zones intersecting buried carbonate platforms liberate significant atmospheric CO₂ and have the potential to influence global climate. We model the spatio-temporal distribution of carbonate platform accumulation within a plate tectonic framework and use wavelet analysis to analyse linked behaviour between atmospheric CO₂ and carbonate-intersecting subduction zone (CISZ) lengths since the Devonian. We find that increasing CISZ lengths likely contributed to a warmer Palaeogene climate.

From the analysis of an Antarctic ice core we showed that during the mid- to late Holocene (6000–750 BP) the long-term fire activity increased with higher rates starting at ~ 4000 BP and, more surprisingly, peaked between 2500 and 1500 BP. The anomalous increase in biomass burning centered at about 2000 BP is due to a complex interaction between changes in atmospheric circulation and biomass availability, with the main contribution coming from southern South America.

This study highlights the counterintuitive result that continental ice sheets can also induce a warming, in particular in the Arctic region. The warming is explained by an increased northward heat transport, resulting from interactions between the atmospheric circulation and ice sheet topography. There is thus an important feedback between ice sheets and temperature, which can help to explain the differences in ice distribution between the Last Glacial Maximum and earlier glacial periods.

We combine North Atlantic sea-surface temperature reconstructions and global climate model simulations to study rapid glacial climate shifts (30–40 000 years ago). Pre-industrial climate boosts similar, albeit weaker, sea-surface temperature variability as the glacial period. However, in order to reproduce most of the amplitude of this variability, and to see temperature variability in Greenland similar to the ice-core record, although with a smaller amplitude, we need forced simulations.

Atmospheric general circulation models equipped with water stable isotopes are key tools to explore the links between climate variables and precipitation isotopic composition and thus to quantify past temperature changes using ice core records. Here, we evaluate the skills of ECHAM5-wiso to simulate the spatio-temporal characteristics of Antarctic climate and precipitation isotopic composition at the regional scale, thanks to a database of precipitation and ice core records.

The BARCAST climate field reconstruction method is tested using synthetic data experiments. It is demonstrated that the output reconstructions have altered statistical properties compared with the input data, but they are also not necessarily consistent with the model assumption of the reconstruction method. The conclusion is that the statistical properties of a reconstruction not only reflect the statistics of the real climate, but they may very well be affected by the manipulation of the data.