Wednesday, February 19, 2025

Advances in AI can help prepare the world for the next pandemic, global group of scientists find

University of Oxford

In the next five years, integrating AI into country response systems could save more lives by anticipating the location and trajectory of disease outbreaks.

Global group of researchers call for better collaboration between academia, government and industry, to ensure safety, accountability and ethics in the use of AI in infectious disease research.

A study published in Nature today outlines for the first time how advances in AI can accelerate breakthroughs in infectious disease research and outbreak response.

The study – which is published following last week’s AI Action Summit and amidst increasing global debate on AI investment and regulation – puts particular emphasis on safety, accountability and ethics in the deployment and use of AI in infectious disease research.

Calling for a collaborative and transparent environment – both in terms of datasets and AI models – the study is a partnership between scientists from the University of Oxford and colleagues from academia, industry and policy organisations across Africa, America, Asia, Australia and Europe.

So far, medical applications of AI have predominantly focused on individual patient care, enhancing for example clinical diagnostics, precision medicine, or supporting clinical treatment decisions.

This review instead considers the use of AI in population health. The study finds that recent advances in AI methodologies are performing increasingly well even with limited data – a major bottleneck to date. Better performance on noisy and limited data is opening new areas for AI tools to improve health across both high-income and low-income countries.

Lead author Professor Moritz Kraemer from the University of Oxford’s Pandemic Sciences Institute, said: “In the next five years, AI has the potential to transform pandemic preparedness.

“It will help us better anticipate where outbreaks will start and predict their trajectory, using terabytes of routinely collected climatic and socio-economic data. It might also help predict the impact of disease outbreaks on individual patients by studying the interactions between the immune system and emerging pathogens.

“Taken together and if integrated into countries’ pandemic response systems, these advances will have the potential to save lives and ensure the world is better prepared for future pandemic threats.”

Opportunities for AI and pandemic preparedness identified in the research include:

Promising advances in improving current models of disease spread, aiming to make modelling more robust, accurate and realistic.
Progress in pinpointing areas of high-transmission potential, helping ensure limited healthcare resources can be allocated in the most efficient possible way.
Potential to improve genetic data in disease surveillance, ultimately accelerating vaccine development and the identification of new variants.
Potential to help determine the properties of new pathogens, predict their traits and identify whether cross species jumps are likely.
Predicting which new variants of already-circulating pathogens – such as SARS-CoV-2 and influenza viruses – might arise, and which treatments and vaccines are best in reducing their impact.
Possible AI-aided integration of population-level data with data from individual-level sources – including wearable technologies such as heart rate and step counts – to better detect and monitor outbreaks.
AI can create a new interface between the highly technical science and healthcare professionals with limited training, improving capacity in settings that need these tools the most.

Not all areas of pandemic preparedness and response will be equally impacted by advances in AI, however. For example, whereas protein language models hold great promise for speeding up understanding of how virus mutations can impact disease spread and severity, advances in foundational models might only provide modest improvements over existing approaches to modelling the speed at which a pathogen is spreading.

The scientists urge caution in suggesting that AI alone will solve infectious disease challenges, but that integration of human feedback into AI modelling workflows might help overcome existing limitations.

The authors are particularly concerned with the quality and representativeness of training data, the limited accessibility of AI models to the wider community, and potential risks associated with the deployment of black-box models for decision making.

Study author Professor Eric Topol, MD, founder and director of the Scripps Research Translational Institute, said: “While AI has remarkable transformative potential for pandemic mitigation, it is dependent upon extensive worldwide collaboration and from comprehensive, continuous surveillance data inputs.”

Study lead author Samir Bhatt from the University of Copenhagen and Imperial College London said: “Infectious disease outbreaks remain a constant threat, but AI offers policymakers a powerful new set of tools to guide informed decisions on when and how to intervene.”

The authors suggest rigorous benchmarks to evaluate AI models, advocating for strong collaborations between government, society, industry and academia for sustainable and practical development of models for improving human health.

Read the paper in Nature: https://www.nature.com/articles/s41586-024-08564-w

DOI: 10.1038/s41586-024-08564-w

ENDS

Notes to editors:

For media enquiries and requests for interview contact: Sarah Nelson, Head of Communications, Pandemic Sciences Institute, University of Oxford: sarah.nelson@ndm.ox.ac.uk / +447812 152044
The study’s authors include scientists from the University of Oxford (Pandemic Sciences Institute; Oxford Martin School; Department of Biology; Oxford Vaccine Group; Department of Computer Science; Department of Statistics; Doctoral Training Centre; Ethox Centre); Stanford University; the University of Tokyo; University of Copenhagen; NIHR Oxford Biomedical Research Centre; Imperial College London; Northeastern University, Boston; Santa Fe Institute; World Health Organization; Stellenbosch University; African Institute for Mathematical Sciences (AIMS); Genomics England; Scripps Research, la Jolla, CA; ETH Zürich; Swiss Institute of Bioinformatics; Harvard T.H. Chan School of Public Health; The Open Data Institute, London; University of California, Los Angeles; The University of Sydney; Max Planck Institute for Software Systems; Max Planck Institute for Intelligent Systems; ELLIS Institute Tübingen, Germany; The Royal Veterinary College, London; Institut Pasteur, Paris, France.

The Pandemic Sciences Institute is an interdisciplinary research institute at the University of Oxford dedicated to confronting the challenge of epidemic and pandemic infectious diseases. We work with academia, industry and public health organisations across the world to create science-led innovations, accelerate understanding, and develop new diagnostics, treatments, vaccines and disease control tools. PSI is hosted by the University’s Nuffield Department of Medicine.

Journal

Nature

DOI

10.1038/s41586-024-08564-w

Method of Research

Computational simulation/modeling

Subject of Research

People

Article Title

Artificial intelligence for modelling infectious disease epidemics

Article Publication Date

19-Feb-2025

Melting glaciers accelerate sea level rise and put drinking water supply at risk

Last decade, the loss of ice in the more populated regions, such as Europe, increased at ever-faster rates

Delft University of Technology
Facebook
X Bluesky Message WhatsApp Email

image:
Mountains in Switserland, melting water and a small part of a glacier
view more
Credit: @TUDelft

All relatively small glaciers combined (because those of Antarctica and Greenland are not included) lost approximately 5% of their volume between 2000 and 2023. This amounts to a loss of 273 billion tonnes of ice, which is more than double that of the Antarctica Ice Sheet. In Central Europe, think of the Alps, even 39% of the ice mass was lost.
Intensifying loss
Unsurprisingly, the amount of meltwater coming down varies per year. But the pattern that is becoming visible the researchers call shocking. Between 2012 and 2023, much more ice was lost (36% more meltwater) compared to the decade before.
It is not only about rising sea levels, says Bert Wouters. “We will directly notice the melting of these glaciers. Because they are located where many people live, it will affect drinking water supplies, in particular in South America and Asia. And the risk of flooding after the melt season also poses a danger.”
Combining global melt data
Bert Wouters is Associate Professor of Geoscience and Remote Sensing at TU Delft and ensured that the data from many studies together produced a solid estimate. “From the accessible glaciers, we have lots of field measurements. From all those other glaciers, we have data from satellites. The methods and thus the meltwater estimates were often varying. It was a big challenge to make it scientifically unified.”
The team of researchers, part of the Glacier Mass Balance Intercomparison Exercise (Glambie), succeeded. This resulted in an annual time series of glacier mass changes for all glacier regions globally from 2000 to 2023.

Journal

Nature

DOI

10.1038/s41586-024-08545-z

Method of Research

Data/statistical analysis

Article Title

Community estimate of global glacier mass changes from 2000 to 2023

Article Publication Date

20-Feb-2025

Global retreat of glaciers has strongly accelerated

International researchers with the participation of Graz University of Technology present a global assessment of ice loss since the beginning of the millennium. In a global comparison, the glaciers in the Alps and Pyrenees are melting the fastest.

Graz University of Technology

image:
The Sulzenauferner in the Stubai Alps (summer 2024).
view more
Credit: Hanna Oberkofler

There are currently around 275,000 glaciers worldwide, in which huge quantities of fresh water are stored. But this reservoir is increasingly shrinking. Since the turn of the millennium, glaciers around the world – i.e. ice masses on land excluding the Greenland and Antarctic ice sheets – have lost around 273 billion tonnes of ice per year. This corresponds to about five and a half times the volume of Lake Constance. Overall, the world’s glaciers have lost around five per cent of their total volume since the year 2000. This is the conclusion reached by an international research team of which Tobias Bolch from the Institute of Geodesy at Graz University of Technology (TU Graz) is a member. The team published the corresponding, comprehensive study today in the scientific journal Nature. It is striking that ice loss has accelerated significantly in recent years. In the second half of the period under investigation (2012 to 2023), it was 36 per cent higher than in the period from 2000 to 2011.

For their study, the researchers collected, homogenised and evaluated glacier data from different sources, including field measurements directly on glaciers as well as radar, laser and gravimetric data from numerous satellite missions. “We compiled 233 estimates of regional glacier mass changes from about 450 data contributors organised in 35 research teams,” explains Michael Zemp, who led the study. Tobias Bolch adds: “The data from ESA Earth observation satellites, as well as from other international space organisations, is particularly important for our research. By analysing this data – measurements of elevation changes are particularly valuable here – we were able to determine the condition of glaciers worldwide.” The result is a unique time series of annual glacier mass changes in the years from 2000 to 2023 for all glacier regions of the world. Due to the large amount of precise data, this study is much more reliable than previous studies of global glacier changes, which were based on less accurate or incomplete data.

18 millimetres sea level rise

The loss of ice from the glaciers since 2000 has led to a rise in sea level of 18 millimetres. This makes the melting of glaciers the second strongest driver of sea level rise after ocean warming, well ahead of the mass loss of the Greenland and Antarctic ice sheets.

Strong regional differences

However, not all glacier regions are affected to the same extent. While the glaciers of the Antarctic and sub-Antarctic islands have only lost 1.5 per cent of their mass, they have shrunk the most in the Alps and the Pyrenees, at around 39 per cent. “Due to their low altitude, they are particularly affected by the higher temperatures,” explains Tobias Bolch. “Additionally, the Alpine and Pyrenean glaciers are comparatively small, which is also a disadvantage. Glaciers generally have a cooling effect on the microclimate of their surroundings. However, this effect is only weakl for small glaciers, which is another reason why the glaciers in the Alps and Pyrenees are shrinking the most.”

Declining meltwater supply in Alpine streams

Valuable freshwater supplies are being lost with the ice from the glaciers. Paradoxically, this is not yet noticeable in many of the world’s glacier-fed rivers; the water volumes from glacier melt have actually increased in most cases. However, these outflows will peak in the future and then decline steadily. “In the European Alps, we have already exceeded this peak discharge. Hence our glaciers will supply the rivers with less and less water,” says Tobias Bolch. “This is becoming a problem especially during longer dry periods. Glacier tributaries are then particularly important as continuous water suppliers. This stabilising effect is increasingly being lost.”

The study on the development of glaciers was carried out as part of the ESA-supported research initiative “Glacier Mass Balance Intercomparison Exercise (GlaMBIE)”. GlaMBIE is coordinated by the World Glacier Monitoring Service (WGMS) hosted at the University of Zurich in collaboration with the University of Edinburgh and the company Earthwave.

Publication:
Community estimate of global glacier mass changes from 2000 to 2023

Authors: The GlaMBIE Team (Michael Zemp, Livia Jakob, Inés Dussaillant, Samuel U. Nussbaumer, Noel Gourmelen, Sophie Dubber, Geruo A, Sahra Abdullah, Liss Marie Andreassen, Etienne Berthier, Atanu Bhattacharya, Alejandro Blazquez,, Laura F. Boehm Vock, Tobias Bolch, Jason Box, Matthias H. Braun, Fanny Brun, Eric Cicero, William Colgan, Nicolas Eckert, Daniel Farinotti, Caitlyn Florentine, Dana Floricioiu, Alex Gardner, Christopher Harig, Javed Hassan, Romain Hugonnet, Matthias Huss, Tómas Jóhannesson, Chia-Chun Angela Liang, Chang-Qing Ke, Shfaqat Abbas Khan, Owen King, Marin Kneib, Lukas Krieger, Fabien Maussion, Enrico Mattea, Robert McNabb, Brian Menounos, Evan Miles, Geir Moholdt, Johan Nilsson, Finnur Pálsson, Julia Pfeffer, Livia Piermattei, Stephen Plummer, Andreas Richter, Ingo Sasgen, Lilian Schuster, Thorsten Seehaus, Xiaoyi Shen, Christian Sommer, Tyler Sutterley, Désirée Treichler, Isabella Velicogna, Bert Wouters, Harry Zekollari, Whyjay Zheng)

In: Nature, 2025

DOI: https://doi.org/10.1038/s41586-024-08545-z

Video of the Glacier Mass Balance Intercomparison Exercise (GlaMBIE)

The animation illustrates the different observation methods at the example of Vatnajökull in Iceland and shows the combined results of glacier mass changes from 2000 to 2023 for all glacier regions worldwide. (Animation by Planetary Visions)
https://www.esa.int/ESA_Multimedia/Videos/2025/02/Revealed_glacier_ice_loss_over_two_decades/

Journal

Nature

DOI

10.1038/s41586-024-08545-z

Method of Research

Data/statistical analysis

Subject of Research

Not applicable

Article Title

Community estimate of global glacier mass changes from 2000 to 2023

Article Publication Date

19-Feb-2025

Melting glaciers increase loss of freshwater resources and rise global sea levels

Glacier mass balance intercomparison exercise

University of Zurich

Glaciers in the Chugach Mountains of Alaska — image:
This image, recorded by the Sentinel-2 satellite on 6 October 2017, shows the melting Scott (left), Sheridan (middle), and Childs (right) glaciers feeding lakes and rivers in their forefields.
view more
Credit: Copernicus Sentinel data 2017

Separate from the continental ice sheets in Greenland and Antarctica, glaciers covered an area of 705,221 km2 and contained 121,728 billion tons of ice globally in 2000. Since then, glaciers have lost about 5% of their ice globally, and regionally between 2% on the Antarctic and Subantarctic Islands and 39% in Central Europe. Annually, glaciers lost 273 billion tons – 273,000,000,000,000 kg – of ice, with an increase of 36% from the first (2000−2011) to the second (2012−2023) half of the period. Glacier mass loss is about 18% larger than the loss from the Greenland Ice Sheet and more than twice that from the Antarctic Ice Sheet.

Worldwide research community effort

For the new study, an international research team under the coordination of the World Glacier Monitoring Service (WGMS), hosted at the University of Zurich (UZH) in Switzerland, carried out the so-called Glacier Mass Balance Intercomparison Exercise (GlaMBIE). The research community collected, homogenized, combined, and analyzed glacier mass changes from different field and satellite observation methods. The team then compared and combined the results from the different methods into an annual time series of glacier mass changes for all glacier regions in the world from 2000 to 2023.

The researchers compiled 233 estimates of regional glacier mass changes from about 450 data contributors organized in 35 research teams. “By combining the advantages of the different observation methods, GlaMBIE provides not only new insights into regional trends and year-to-year variability. We could also identify differences among observation methods, which is an opportunity to better understand and improve future estimates,” says Michael Zemp, UZH professor at the Department of Geography, who led the study.

Sinking regional freshwater resources, rising global sea levels

From 2000 to 2023, the global glacier mass loss totals 6,542 billion tons. This loss contributed 18 mm to global sea-level rise at an annual rate of 273 billion tons or 0.75 millimeters yearly. With this, glaciers are currently the second-largest contributor to global sea-level rise, after the warming of the ocean and before the contributions from the Greenland Ice Sheet, changes in land water storage, and the Antarctic Ice Sheet.

In addition, glacier melt results in the loss of regional freshwater resources. “To put this in perspective, the 273 billion tonnes of ice lost in one single year amounts to what the entire global population consumes in 30 years, assuming three litres per person and day,” states Zemp.

“Glaciers are vital freshwater resources, especially for local communities in Central Asia and Central Andes, where glaciers dominate runoff during warm and dry seasons”, says UZH glaciologist Inés Dussaillant, who was involved in the GlaMBIE analyses. “But when it comes to sea-level rise, the Arctic and Antarctic regions with their much larger glacier areas are the key players. Almost one quarter of the glacier contribution to sea-level rise originates from Alaska,” she adds.

Limiting negative effects through climate protection

The present study marks an important milestone for the International Year of Glaciers’ Preservation in 2025 and the Decade of Action for Cryospheric Sciences (2025−2034) declared by the United Nations. GlaMBIE provides a new observational baseline for future studies, allowing improved projections of freshwater resources and sea-level rise.

“Our observations and recent modelling studies indicate that glacier mass loss will continue and possibly accelerate until the end of this century”, says UZH glaciologist and GlaMBIE project manager Samuel Nussbaumer. “This underpins Intergovernmental Panel on Climate Change’s call for urgent and concrete actions to reduce greenhouse gas emissions and associated warming to limit the impact of glacier wastage on local geohazards, regional freshwater availability, and global sea-level rise”, he concludes.

A fleet of satellites has been used to monitor glaciers worldwide using optical, radar, laser and gravity measurements. From top: CryoSat, Terra, ICESat, and the twin GRACE spacecraft, above a map of elevation change for the Vatnajökull ice cap on Iceland.

Credit

ESA, NASA, and Planetary Visions

This image, recorded by the Sentinel-2 satellite on 12 September 2017, shows the blue glaciers on the reddish-brown Franz Josef Land archipelago north of the 80th parallel in the Arctic Ocean (black). The glaciers (blue) are covered with little or no snow (white), indicating a significant mass loss.

Credit

Copernicus Sentinel data 2017

Journal

Nature

DOI

10.1038/s41586-024-08545-z

Method of Research

Experimental study

Subject of Research

Not applicable

Article Title

Community estimate of global glacier mass changes from 2000 to 2023

Article Publication Date

19-Feb-2025

Natural hydrogen: a sustainable energy source in mountain ranges

GFZ researchers identify where to expect natural and exploitable hydrogen resources using state-of-the-art simulations of plate tectonic processes

GFZ Helmholtz-Zentrum für Geoforschung

Mountain range — image:
Panorama view of the Swiss Alps - Canton of Grisons in Eastern Switzerland, a potential natural H2 exploration area.
view more
Credit: Frank Zwaan, GFZ

The successful development of sustainable georesources for the energy transition is a key challenge for humankind in the 21st century. Hydrogen gas (H2) has great potential to replace current fossil fuels while simultaneously eliminating the associated emission of CO2 and other pollutants. However, a major obstacle is that H2 must be produced first. Current synthetic hydrogen production is at best based on renewable energies but it can also be polluting if fossil energy is used.

The solution may be found in nature, since various geological processes can generate hydrogen. Yet, until now it has remained unclear where we should be looking for potentially large-scale natural H2 accumulations.

A team of researchers led by Dr Frank Zwaan, a scientist in the Geodynamic Modelling section at GFZ Helmholtz Centre for Geosciences, present an answer to this question: using plate tectonic modelling, they found that mountain ranges in which originally deep mantle rocks are found near the surface represent potential natural hydrogen hotspots. Such mountain ranges may not only be ideal geological environments for large-scale natural H2 generation, but also for forming large-scale H2 accumulations that can be drilled for H2 production. The results of this research have now been published in the journal Science Advances. Also part of the team are Prof. Sascha Brune and Dr Anne Glerum of GFZ’s Geodynamic Modelling section. The other team members are based at Tufts University (Dr Dylan Vasey) and New Mexico Tech (Dr John Naliboff) in the USA, as well as at the University of Strasbourg (Prof. Gianreto Manatschal) and Lavoisier H2 Geoconsult (Dr Eric. C. Gaucher) in France.

Natural H2 potential in tectonic environments

Natural hydrogen can be generated in several ways, for instance by bacterial transformation of organic material or splitting of water molecules driven by decay of radioactive elements in the Earth’s continental crust. As a result, the occurrence of natural H2 is reported in many places worldwide. The general viability of natural hydrogen as an energy source has already been proven in Mali, where limited volumes of H2 originating from iron-rich sedimentary layers are produced through boreholes in the subsurface.

However, the most promising mechanism for large-scale natural hydrogen generation is a geological process in which mantle rocks react with water. The minerals in the mantle rocks change their composition and form new minerals of the so-called serpentine group, as well as H2 gas. This process is called serpentinization. Mantle rocks are normally situated at great depth, below the Earth’s crust. In order for these rocks to come in contact with water and serpentinize, they must be tectonically exhumed, i.e. being brought near the Earth’s surface. There are two main plate tectonic environments in which mantle rocks are exhumed and serpentinized over the course of millions of years: (1) ocean basins that open as continents break apart during rifting, allowing the mantle to rise as the overlying continental crust is thinned and eventually split (for example in the Atlantic Ocean), and (2) subsequent basin closure and mountain building as continents move back together and collide, allowing mantle rocks to be pushed up towards the surface (for example in the Pyrenees and Alps).

Numerical modelling helps constraining regions with natural H2 resources

A thorough understanding of how such tectonic environments evolve is key to properly assess their natural hydrogen potential. Using a state-of-the-art numerical plate tectonic modelling approach, calibrated with data from natural examples, the GFZ-led research team simulated the full plate tectonic evolution from initial rifting to continental break-up, followed by basin closure and mountain building. In these simulations, the researchers were able to determine for the first time where, when, and how much mantle rocks are exhumed in mountains, and when these rocks may be in contact with water at favorable temperatures, to allow for efficient serpentinization and natural hydrogen generation.

It turns out that conditions for serpentinization and thus natural H2 generation are considerably better in mountain ranges than in rift basins. Due to the comparably colder environment in mountain ranges, larger volumes of exhumed mantle rocks are found at favorable serpentinization temperatures of 200-350°C, and at the same time plenty of water circulation along large faults within the mountains can allow for their serpentinization potential to be realized. As a result, the annual hydrogen generation capacity in mountain ranges can be up to 20 times greater than in rift environments. In addition, suitable reservoir rocks (for example sandstones) required for the accumulation of economically viable natural H2 volumes are readily available in mountain ranges, but are likely absent during serpentinization and hydrogen generation in the deeper parts of rift basins.

Natural hydrogen exploration (and more) in mountain ranges

The results of this now published research provide a strong impulse to intensify the exploration for natural H2 in mountain ranges. In fact, various exploration efforts are already underway in places such as the Pyrenees, European Alps, and Balkans, where researchers have previously found indications of ongoing natural hydrogen generation.

“Crucial to the success of these efforts will be the development of novel concepts and exploration strategies. Of particular importance is how the formation of economic natural H2 accumulations is controlled by the tectonic history of a given exploration site. In particular, we will need to determine the timing of the key geological processes involved, because if H2 reservoirs are to form during mountain building, there must have been rifting, i.e. stretching, beforehand. So insights gained from plate tectonic simulations such as those performed in this study will be of great value”, says Frank Zwaan, lead author of the study.

Sascha Brune, head of the Geodynamic Modelling Section at GFZ, continues: “This new research advances our understanding of suitable environments for natural hydrogen generation. Given the economic opportunities associated with natural H2, now is the time to go further and also investigate migration pathways of hydrogen and deep, hydrogen-consuming microbial ecosystems to better understand where potential H2 reservoirs can actually form.”

Zwaan adds: “Overall, we may be at a turning point for natural H2 exploration. As such, we could be witnessing the birth of a new natural hydrogen industry.”

Sketch of mountain building with exhumed mantle.
Credit
CC BY-NC-SA 3.0 USGS / ESEU edited by Frank Zwaan, GFZ

Journal

Science Advances

DOI

10.1126/sciadv.adr3418

Method of Research

Computational simulation/modeling

Subject of Research

Not applicable

Article Title

Rift-inversion orogens are potential hotspots for natural H2 generation.

Article Publication Date

19-Feb-2025