Using statistical methods and a set of ensemble climate models, we decompose the sources of Atlantic Meridional Overturning Circulation (AMOC) variance. Three distinct phases of physical variability are identified: from 1850 to 1990, internal variability dominates; from 1990 to 2050, dynamical adjustment related to AMOC decline takes over; after 2050, differences between forcing scenarios become dominant. Beyond these physical factors, model variability remains a major source of uncertainty.

Climate models serve as good guesses of how humans affect the climate, but they cannot explore all possible future scenarios of interest. We develop a method that can serve as a fast and cheap stand-in to evaluate likely changes in variables like surface temperature and relative humidity at a regional scale in arbitrary future climates. Crucially, our method captures relationships between different geographic areas and predicts both average values and likely ranges using a unified framework.

We analyze the duration of large-scale weather patterns and their link to near-surface temperatures during heatwaves in Central Europe for 1950–2023. Compared to non-heatwave days, a stronger link between them is found on heatwave days from May to September. We relate our results to typical long-lasting weather patterns known as weather regimes. In July and August, weather patterns last longer as west winds are often blocked by Scandinavian and European blocking regimes, inducing hot extremes.

In this study, we find that global freshwater storage on land (excluding Greenland and Antarctica) will likely decrease much more than raw models suggest – about 83 millimeters less water by 2100. This means that earlier studies may have overestimated future water availability. When we correct the models using observations, the results consistently show significant declines in freshwater, which could worsen water stress worldwide if demand continues to grow.

Geological evidence indicate persistent tropical sea-ice cover in the deep past, often called Snowball Earth. Using a climate model, we show here that clouds substantially cool down the tropics and facilitate the advance of sea-ice into lower latitudes. We identify a critical threshold temperature close to 0 °C from where cooling down the Earth is accelerated. This value can be used as a constraint on Earth's sensitivity to CO₂, as recent cold paleoclimates never entered Snowball Earth.

Differences in surface albedo and large-scale circulation are often considered to drive the contrasts between monsoons and deserts. However, using a radiation-circulation framework, we find that large-scale circulation serves primarily as a trigger for this contrast. It is the radiative feedbacks from water vapor and clouds, rather than surface albedo, that amplify the contrast between these climates.

When one tipping element collapses the likelihood of another collapsing may be significantly affected. Using our simplified network model, we find that on the whole these interactions destabilise the Earth system, both in the short term and at equilibrium, though the effects are most noticeable after the year 2100. We find that to minimise tipping risks, it is essential to keep temperatures as close as possible to 1.5 °C in the short term and below 1 °C in the longer run.

Adding a layer of reflective particles high in the atmosphere is one suggested way of cooling the planet and reducing the impacts of climate change. This technique might be less logistically difficult in the high latitudes, because the material could be released at lower altitude there. Here, we use new simulations in three earth system models to assess how this form of intervention, High-Latitude Low-Altitude Stratospheric Aerosol Injection (HiLLA-SAI), would impact the global climate.