Atmospheric rivers (ARs) move vast amounts of water through the atmosphere and often cause weather extremes, yet they are usually studied as regional events. Using 84 years of mapped AR trajectories, we reveal the global "roadmap" of ARs, a transport network of high-activity hubs, sparse atmospheric highways & hierarchical basins. Our approach shows how water vapor is systematically channelled through an atmospheric transport network, offering new ways to study changes in the global water cycle.

We study when anthropogenic signal becomes detectable in the global land carbon sink, which has risen since the 1950s due to CO₂ fertilization and mid- to high-latitude warming. The signal emerges earlier at the global than at regional scales. Future scenarios (2016–2100) take longer to detect than the historical period (1851–2014) because the signal is weaker relative to larger natural variability. Removing circulation-induced variability with dynamical adjustment shortens the detection time.

The impacts of climate extremes such as heatwaves and droughts can be made worse when they happen at the same time. Dheed is a global database of dry and hot compound extreme events from 1950 to 2023. It can be combined with other data to study the impacts of those events on terrestrial ecosystems, specific species or human societies. Dheed's analysis confirms that extremely dry and hot days have become more common on all continents in recent decades, especially in Europe and Africa.

We investigated how machine learning can forecast extreme vegetation responses to weather. Examining four models, no single one stood out as the best, though "echo state networks" showed minor advantages. Our results indicate that while these tools are able to generally model vegetation states, they face challenges under extreme conditions. This underlines the potential of artificial intelligence in ecosystem modeling, also pinpointing areas that need further research.