We reconstructed sea surface temperatures for the Oligocene and Miocene periods (34–11 Ma) based on archaeal lipids from a site close to the Wilkes Land coast, Antarctica. Our record suggests generally warm to temperate surface waters: on average 17 °C. Based on the lithology, glacial and interglacial temperatures could be distinguished, showing an average 3 °C offset. The long-term temperature trend resembles the benthic δ¹⁸O stack, which may have implications for ice volume reconstructions.

In the context of a warming climate, mountain environments are especially vulnerable to a change in the risk pattern. Our study focuses on the past evolution of wet avalanches, likely triggered by warmer temperatures destabilizing the snow cover. In the last 3300 years we observed an increase of wet avalanche occurrence related to human activities, intensifying pressure on forest cover, as well as favorable climate conditions such as warmer temperatures coinciding with retreating glacier phases.

Recording the precise timing and sequence of events is essential for understanding rapid climate changes and improving climate model predictive skills. Here, we precisely assess the relative timing between ocean and atmospheric changes, both recorded in the same deep-sea core over the last 45 kyr. We show that decreased mid-depth water mass transport in the western equatorial Atlantic preceded increased rainfall over the adjacent continent by 120 to 980 yr, depending on the type of climate event.

Understanding the causes of past climatic changes plays a major role in deciphering monsoon variability. This study aims to understand past changes in siliceous productivity in the Somali upwelling area, as well as the palaeo-upwelling strength and its relation with southwest monsoon (SWM) rainfall. It is found that the Somali upwelling has historically had a negative impact on SML rainfall; this has important implications as previous studies predict a future increase in the Somali upwelling.

We try to test and evaluate an approach for using two main sources of information on the climate of the past: climate model simulations and proxies. This is done via data assimilation (DA), a method that blends these two sources of information in an intelligent way. However, DA and climate models are computationally very expensive. Here, we tested the ability of a computationally affordable DA to reconstruct high-resolution climate fields.