Scientists from around the world have gathered together at the annual European Geosciences Union general assembly, to discuss current projects, working hypotheses and potential findings. There are nearly 18,000 in attendance this year and there is much to learn.

AMOC – the Atlantic Meridional Overturning Circulation - brings warmth to the north and cooler waters to the south in huge volumes. Climate modellers have expressed concern for its collapse (and subsequent weather chaos) as temperatures rise more generally, but others have said it is more resilient. But Stefan Rhamstorf has announced that extending the models past 2100 can show a different picture. The odds have shifted from 10% to more like 50/50 if the Paris climate target is missed.

Has such climate change ever occurred before? And if so, what drove it? Hana Jurikova and colleagues have been using novel techniques to detect a link between atmospheric CO2 levels and rapid climate change in the geological past, and explains how boron records in ancient brachiopods might give us a clue.

What of the 6.2 magnitude earthquake near Istanbul last week? Could it have been worse? Will the next one be the big one? Expert Patricia Martínez-Garzón of GFZ in Germany doesn’t quite allay the fears.

Could more lives be saved from landslides and flash floods if we could set up a warning system? Stefania Ursica hopes so, and has looked to animal behaviour to design a programme to scan networks of seismic monitoring stations’ output for the faint signals. Encoding different hunting and communication strategies – from nomadic whales to humming birds and bats, her new algorithm might be just the thing, though prediction will always be a different problem.

Presenter: Roland Pease Producer: Alex Mansfield with Sophie Ormiston Production co-ordinator: Josie Hardy

(Image: 3D render of a Topographic Map of Western Europe with the clouds from 27 January, 2025. Credit: Frank Ramspott/Getty Images)

Last week, the website covid.gov looked very different, containing information on coping with covid and US research. This week it leads you to a White House webpage outlining the lab-leak hypothesis – that the pandemic was the result of dodgy lab work at the Wuhan Institute of Virology. The scientific consensus however continues to suggest a zoonotic crossover event. And a preprint recently published by Zach Hensel of ITQB NOVA (NOVA University Lisbon) and Florence Débarre of the Sorbonne, adds new early sequences to the evidence base. As they suggest, it doesn’t seal the debate, but it certainly doesn’t surprise the scientists.

As the Trump administration continues to rattle swords with US science, we speak with Lisa Fazio of Vanderbuilt University who found out abruptly her funding from the National Science Foundation is ending, and Don Ingber, founding director of Harvard University’s Wyss Institute about the impacts they are feeling from the federal belt tightening.

Finally we learn this week of an eerie species of Hawai’ian caterpillar moth. Not only is it carnivorous, but it ties small pieces of the indigestible remains of insects and bugs found in spider webs to its case, like a gruesome, camouflaged suit of armour. Why? As Dan Rubinoff of Hawai’i describes, it protects it from the hungry host spider, making it appear as the leftovers of yesterday’s dinner, or perhaps the shed exoskeleton of the spider itself.

Presenter: Roland Pease Producer: Alex Mansfield Production Coordinator: Josie Hardy

(Image: The view of Huanan seafood market on February 9, 2021 in Wuhan, Hubei Province, China. Credit: Getty Images)

After 60 years of doubling computer complexity every two years, can Moore’s law still predict the future power of the devices we use?

In 1965, electronics pioneer Gordon Moore was asked to predict the next ten years of progress with the then new-fangled silicon integrated circuits. He estimated, based on physics and manufacturing technologies then available what seemed remarkable: that every two years they would double in complexity, and halve in price, until 1975.

60 years on, perhaps the even more remarkable thing is that they just kept doubling.

Can Moore’s law hold into future decades? What are the next technological innovations that might keep it running?

Sri Samavedam is the vice president for silicon technologies at imec in Belgium, whose job it is to think about the practicalities of manufacturing the next generations of chips years before they become real.

Scott Aaronsen of the University of Texas is a thinker in the field of Quantum Computing – could quantum computing keep the rate of growth going? Or does it need to be thought of differently?

One of the limitations on chip miniaturisation is the dissipation of heat from conventional electronic flow. Nick Harris of Lightmatter is looking at using photons rather than electrons to carry info and logic around a circuit with lower power losses.

Stan Williams has spent much of his career thinking about new devices that could be fabricated into integrated circuits to give it all a push forward. And he tells Roland how the memristor could effectively bring the power of analogue computing to bear as we reach some of the limits of the digital age we have been living in.

Presenter: Roland Pease Producer: Alex Mansfield and Gareth Nelson-Davies