Concentric Rings: A hidden geometry of a Nested Caldera

Trial by Fire: The loss of an ancient world

• Phase 1 (Plinian Fall): A dry, gas-driven Plinian pumice fall blanketed the island's ruins up to 7 meters deep. 
• Phase 2 (Pyroclastic Surges): Wet phreatomagmatic surges and flows created a 12-meter stratified layer that completely buried the settlement. 
• Phase 3 (Caldera Collapse): Ash and massive 2-meter lithic blocks were emplaced, signifying the structural collapse of the caldera. 
• Phase 4 (Final Debris Flows): A thick sequence of ignimbrites and debris flows tens of meters deep formed the modern coastal cliffs

“Muddy Sea”: Floating Pumice and Post-Eruptive Hazards

Plato’s measurements against geodynamic reality: A striking dichotomy between spatial scales

Geological Inspiration vs. Mythical Reality 

•       Plato, Critias (c. 360 BCE),  (https://topostext.org/work/553, https://topostext.org/work/552)
•       Palaeomagnetic analysis on pottery as indicator of the pyroclastic flow deposits temperature: new data and statistical interpretation from the Minoan eruption of Santorini, Greece (E. Tema, E. Zanella, F.J. Pavon-Carrasco, D. Kondopoulou and S. Pavlides), 2015
•       Atlantis: The Truth Behind the Legend. (Galanopoulos, A. G., & Bacon, E. (1969))
•       Santorini Volcano. (Druitt, T. H., et al. (1999) )

References

Greene R, Elffers J: The 48 Laws of Power. Penguin Books, New York, New York, USA. 2020. 

Unlike Odysseus, however, the obstacles faced by international researchers rarely come in the form of mythical creatures or raging seas. Instead, they often appear as visa applications, residence permit renewals, unfamiliar healthcare systems, administrative paperwork, language barriers, and the uncertainty that comes with building a career far from home. These challenges often remain invisible within academia, yet they shape the daily lives of thousands of researchers around the world (Fazel et al., 2025). Research has consistently shown that moving abroad for study or work involves much more than geographic relocation. International researchers frequently navigate cultural adaptation, language barriers, changing social networks, and the psychological demands of building a new life while maintaining academic productivity. Rather than isolated experiences, these challenges are part of a broader process of adaptation that has been widely described in studies on international mobility and researcher well-being (Thomas, H.N. et al. 2026).

The call to leave Ithaca.

Every scientific journey begins with curiosity. For many researchers, that curiosity eventually extends beyond laboratories and classrooms, encouraging researchers to cross borders in search of new ideas, collaborations, and opportunities for growth. The decision to leave home is rarely driven by the desire to migrate itself, but rather by the pursuit of knowledge and the belief that science flourishes through the exchange of people, perspectives, and cultures.

Like Odysseus setting sail from Ithaca, we often embark on this journey with a destination clearly in mind: a PhD, a postdoctoral position, a research collaboration, or the opportunity to work alongside scientists whose work we admire. Yet, studies on international academic mobility suggest that crossing borders is never simply a geographical transition (Fazel et al., 2025). It also marks the beginning of a process of cultural adaptation, personal growth, and unexpected challenges that extend far beyond research itself.

Sailing into unknown waters. 

Scientific careers are often planned around grant deadlines, conferences, field campaigns, and publication schedules. Administrative processes, however, rarely follow the same timeline. For many international researchers, periods of uncertainty become an unavoidable part of academic life while waiting for visas, residence permits, work authorizations, or other documents required to continue moving across borders.

Unlike scientific projects, these processes are largely beyond a researcher's control. A delayed residence permit may postpone conference travel, limit international mobility, complicate fieldwork, or simply make everyday decisions unexpectedly uncertain. While experiments and manuscripts continue to move forward, administrative processes often do not.

Between Scylla and Charybdis.

In The Odyssey, one of Odysseus' greatest challenges was navigating the narrow strait between Scylla and Charybdis, where avoiding one danger meant moving closer to another. For many international researchers, academic life often feels remarkably similar. The pursuit of scientific excellence unfolds alongside a second, less visible journey: adapting to a new country while trying to meet the expectations of an increasingly demanding academic environment. 

Recent studies have highlighted that many of the pressures experienced by researchers arise from the interaction between internal academic demands, such as publishing, securing funding, career uncertainty, and external challenges, including migration, financial constraints, family responsibilities, and the process of adapting to a new society (Thomas, H.N. et al. 2026, Nicholls et al. 2022). Rather than existing separately, these pressures often reinforce one another, making the scientific journey considerably more complex than the publication record alone might suggest.

Like navigating between Scylla and Charybdis, international researchers are often required to balance competing priorities for which there is no perfect solution. Time devoted to administrative responsibilities, relocation, or personal adaptation is time that cannot be devoted to writing, publishing, or preparing grant proposals. In contrast, focusing exclusively on scientific productivity may leave little opportunity to address the practical demands of building a life in a new country.

Finding Ithaca.

Perhaps the greatest lesson of The Odyssey is that the journey changes the traveller as much as the destination. For international researchers, the same may be true. Along the way, we learn far more than new scientific techniques or analytical methods. We learn resilience, adaptability, patience, and the ability to find opportunities in unfamiliar places. Every border crossed, every collaboration formed, and every challenge overcome quietly shapes not only the scientist, but also the person we become.

Ithaca may not be a permanent position, a particular country, or even the laboratory we once dreamed of joining. Instead, it may be the moment when we realize that home is no longer defined by geography alone, but by the communities we build, the knowledge we share, and the people we meet along the way. Science may have no borders, but researchers do, and perhaps crossing those borders is what ultimately allows us to grow beyond them.

References

Fazel, P., Kovacs, K. E., Jourian, N., & Kovács, J. Overcoming borders: A systematic review of mental health issues and challenges faced by international students. Social Sciences & Humanities Open, 2025. https://doi.org/10.1016/j.ssaho.2025.102263

Nicholls, H., Nicholls, M., Tekin, S., Lamb, D., & Billings, J. The impact of working in academia on researchers’ mental health and well-being: A systematic review and qualitative meta-synthesis. PLOS ONE, 2022. https://doi.org/10.1371/journal.pone.0268890

Thomas, H. N., Norman, M. K., Vempalli, H., Hamm, M., White, G. E., Morone, N. E., Murrell, A. J., Rubio, D. M., & Gauvin, N. "Academia Is a Very Unforgiving Space": A Qualitative Study of Challenges Faced by Under-represented Scholars in Biomedical Research. Journal of General Internal Medicine, 2026. https://doi.org/10.1007/s11606-026-10366-x

The findings

Looking ahead

Edited by Marius Schlaak

Reference   

Hart-Davis M.G., Scherer D., Schwatke C., Sawyer A., Pavelsky T., Ray R.D., Matte P., Dettmering D, Seitz F. 2026. Observing the tidal pulse of rivers from wide-swath satellite altimetry. Nature. https://doi.org/10.1038/s41586-026-10287-z.

What satellites made visible

The same salt, different physics

Where salt starts to steer

More than an observing gap

Further Reading

I was very surprised when I received the email, and I read it several times to make sure it was real. After that, I felt extremely proud of my work because I could see the evaluation scores, comments, and recommendations from the judges. Their feedback confirmed that my research, the design of my poster, and the overall quality of my work met the high standards of the EGU and were valued by the scientific community

Could you briefly describe your EGU presentation, what is your core research about, and what inspired you to focus on this topic?

My presentation focused on the interhemispheric differences in Total Electron Content (TEC) during 25 geomagnetic storms. Specifically, I investigated how TEC enhancements propagate from the polar regions toward lower latitudes by analyzing data from GNSS receivers located at magnetic high-latitudes in Antarctica and North America. My interest in this topic comes from studying ionospheric processes over Antarctica and noticing that the Southern Hemisphere does not always exhibit the same behavior observed in the Northern Hemisphere. While many of our current models and understanding are based on observations from the Northern Hemisphere, several studies suggest that the Southern Hemisphere can exhibit different ionospheric behaviors. I was motivated to explore these differences and contribute to a better understanding of how geomagnetic storms affect both hemispheres.

My interest in this topic comes from studying ionospheric processes over Antarctica and noticing that the Southern Hemisphere does not always exhibit the same behavior observed in the Northern Hemisphere.

I think one of the most impactful aspects of my research was the use of observations from Antarctica, a region where ionospheric measurements are still relatively limited compared to the Northern Hemisphere. Studying similar magnetic high-latitude locations allowed me to directly compare the response of both hemispheres during geomagnetic storms. Another important aspect was the analysis of 25 geomagnetic storms occurring under different solar cycle conditions, seasons, and solar wind drivers. Rather than focusing on a single event, I was able to identify recurring patterns as well as important differences between storms. During the poster session, several researchers highlighted this comprehensive approach as one of the strengths of the study.

Looking back, the most memorable part was the week before the conference. At that time, I was simultaneously preparing my EGU poster and the final defense of my master's degree. After months of work, I completed my master's defense, and the very next day I boarded a plane from Chile to Austria to attend the EGU General Assembly for the first time. It was an exciting and unforgettable experience. My advice to future OSPP participants is to understand both the strengths and the limitations of their research. Being honest about what your study can and cannot explain is very important. The conference provides a unique opportunity to discuss your work with researchers from different backgrounds, receive valuable feedback, and gain new ideas. Those conversations can help improve your research and often open new directions for future studies.

The conference provides a unique opportunity to discuss your work with researchers from different backgrounds, receive valuable feedback, and gain new ideas.

Presenting at EGU provided me with very valuable feedback that I have since incorporated into my research. Several comments helped me refine my analysis and strengthen the conclusions of my work. In addition, I had the opportunity to meet researchers working in the polar upper atmosphere, which is my main area of interest. These interactions were very important for me, as they allowed me to better understand current research directions and methodologies in the field. I consider this type of networking an essential step for my academic development and future career.

    [caption id="attachment_4834" align="alignnone" width="217"] Carlos Castillo-Rivera presenting his poster during the EGU25[/caption]

In your view, how could your findings contribute to future research developments or real-world applications in your field?

In my view, this work contributes to the understanding of magnetosphere–ionosphere coupling, which is a highly complex system that is still not fully understood. In addition, it may have relevant implications for GNSS performance, particularly in the Antarctic region, where ionospheric conditions can significantly affect signal propagation and positioning accuracy.

Was there any unexpected or especially rewarding experience during EGU that made the participation particularly meaningful for you?

The presentation was especially meaningful for me because I received many questions and very valuable feedback. I remember that my poster session was scheduled from 4 to 6 pm, but I continued discussing my work until around 7 pm. When I finally looked around, I realized that I was the last person in the room—everyone else had already left and all the posters had been taken down. It was a very rewarding experience because it showed me that there was genuine interest in my research, and it gave me the opportunity to engage in deeper scientific discussions than I had expected.

    Carlos Castillo-Rivera is currently a PhD student at Sapienza Università di Roma, affiliated with the Istituto Nazionale di Geofisica e Vulcanologia (INGV), Italy. His research focuses on the recently discovered optical phenomenon known as Strong Thermal Emission Velocity Enhancement (STEVE), which is distinct from traditional auroral emissions. In particular, his work investigates ionospheric scintillations associated with STEVE in order to improve the understanding of its physical mechanisms. This is a multi-instrumental study that combines data from GNSS receivers, optical emission measurements, ionosondes, low-Earth-orbit satellites, radar observations, and numerical models.

Tectonic uplift before tectonic uplift

What are forearc ranges, and how they uplift in sediment-rich margins?

The Nahuelbuta Range: an unusual high in the Patagonian forearc

The unknowns of the Nahuelbuta Range

• When and why did the Nahuelbuta Range begin to uplift?
• How have uplift rates changed through time and why?

Finding a marker of uplift onset

• Kaizuka's isolated surfaces are actually part of a continuous paleolandscape preserved across the top of the range.
• Even better, part of this paleolandscape had already been dated by Rehak (2008) between 2.3 Myr and 4.2 Myr.

What rivers can reveal about mountain building

• Uplift started ~2.5 Myr ago, matching the age of the relic landscape.
• Even cooler, instead of rising steadily, the range grew through at least three distinct pulses.

What probably didn't drive the uplift pulses

A tectonic-climatic interplay?

Does basal accretion actually make sense?

• Pulse 1: Starting ~2.5 Myr ago, coinciding with abandonment of the paleolandscape. It lasted ~0.8 Myr and affected a region about 70 km wide.
• Pulse 2: Starting ~1.7 Myr ago. ~1 Myr of duration and had a similar wavelength of ~70 km.
• Pulse 3a: Starting ~0.6 Myr ago, coinciding with the emersion of the Arauco Peninsula. It lasted ~0.3 Myr years and had a wavelength of ~40 km.
• Pulse 3b: Starting ~0.2 Myr ago. ~0.2 Myr of duration and also had a wavelength of ~40 km.

The existence of our basal accretion bodies is supported by seismic data

Implications for subduction dynamics

• Does this mechanism operate elsewhere along the margin, and if so, are these uplift pulses synchronized across Patagonia or do they vary from place to place?
• Are uplift pulses synchronized across Patagonia or do they vary from place to place?
• More broadly, could climate play a larger role in the long-term evolution of forearc systems than previously thought?

References:
[1] Bohm, M., Lüth, S., Echtler, H., Asch, G., Bataille, K., Bruhn, C., Rietbrock, A., and Wigger, P., 2002. The Southern Andes between 36° and 40°S latitude: seismicity and average seismic velocities. Tectonophysics, 356, 275–289. https://doi.org/10.1016/S0040-1951(02)00399-2

[2] Contreras-Reyes, E., Flueh, E. R., and Grevemeyer, I., 2010. Tectonic control on sediment accretion and subduction off south central Chile: Implications for coseismic rupture processes of the 1960 and 2010 megathrust earthquakes. Tectonics, 29. https://doi.org/10.1029/2010TC002734

[3] Encinas, A., Sagripanti, L., Rodríguez, M. P., Orts, D., Anavalón, A., Giroux, P., Otero, J., Echaurren, A., Zambrano, P., and Valencia, V., 2021. Tectonosedimentary evolution of the Coastal Cordillera and Central Depression of south-Central Chile (36°30′-42°S). Earth. Sci. Rev., 213, 103465. https://doi.org/10.1016/j.earscirev.2020.103465

[4] Folguera, A. and Ramos, V. A., 2009. Collision of the Mocha fracture zone and a <4 Ma old wave of orogenic uplift in the Andes (36°–38°S). Lithosphere, 1, 364–369. https://doi.org/10.1130/L66.1

[5] Haberland, C., Rietbrock, A., Lange, D., Bataille, K., and Hofmann, S., 2006. Interaction between forearc and oceanic plate at the south-central Chilean margin as seen in local seismic data. Geophys. Res. Lett., 33. https://doi.org/10.1029/2006GL028189

[6] Hoth, S., Hoffmann-Rothe, A., and Kukowski, N., 2007. Frontal accretion: An internal clock for bivergent wedge deformation and surface uplift. J. Geophys. Res. Solid Earth, 112. https://doi.org/10.1029/2006JB004357

[7] von Huene, R. and Scholl, D. W., 1991. Observations at convergent margins concerning sediment subduction, subduction erosion, and the growth of continental crust. Reviews of Geophysics, 29, 279–316. https://doi.org/10.1029/91RG00969

[8] Kaizuka, S., Matsuda, T., Nogami, M., and Yonekura, N., 1973. Quaternary tectonic and recent seismic crustal movements in the Arauco Peninsula and its environs, central Chile. Geographical Reports of Tokyo Metropolitan University, 1–49.

[9] Krawczyk, C. M., Mechie, J., Lüth, S., Tašárová, Z., Wigger, P., Stiller, M., Brasse, H., Echtler, H. P., Araneda, M., and Bataille, K., 2006. Geophysical Signatures and Active Tectonics at the South-Central Chilean Margin, in: The Andes: Active Subduction Orogeny. Edited by: Oncken, O., Chong, G., Franz, G., Giese, P., Götze, H.-J., Ramos, V. A., Strecker, M. R., and Wigger, P.. Springer Berlin Heidelberg, Berlin, Heidelberg, 171–192.

[10] Lohrmann, J., Kukowski, N., Krawczyk, C. M., Oncken, O., Sick, C., Sobiesiak, M., and Rietbro, A., 2006. Subduction Channel Evolution in Brittle Fore-Arc Wedges — a Combined Study with Scaled Sandbox Experiments, Seismological and Reflection Seismic Data and Geological Field Evidence, in: The Andes: Active Subduction Orogeny. Edited by: Oncken, O., Chong, G., Franz, G., Giese, P., Götze, H.-J., Ramos, V. A., Strecker, M. R., and Wigger, P.. Springer Berlin Heidelberg, Berlin, Heidelberg, 237–262.

[11] Melnick, D., Bookhagen, B., Echtler, H. P., and Strecker, M. R., 2006. Coastal deformation and great subduction earthquakes, Isla Santa María, Chile (37°S). GSA Bulletin, 118, 1463–1480. https://doi.org/10.1130/B25865.1

[12] Melnick, D., Bookhagen, B., Strecker, M. R., and Echtler, H. P., 2009. Segmentation of megathrust rupture zones from fore-arc deformation patterns over hundreds to millions of years, Arauco peninsula, Chile. J. Geophys. Res. Solid Earth, 114, 1407. https://doi.org/10.1029/2008JB005788

[13] Menant, A., Angiboust, S., Gerya, T., Lacassin, R., Simoes, M., and Grandin, R., 2020. Transient stripping of subducting slabs controls periodic forearc uplift. Nature Communications, 11:1, 11, 1–10. https://doi.org/10.1038/s41467-020-15580-7

[14] Racano, S., Schildgen, T. F., Cosentino, D., and Miller, S. R., 2021. Temporal and Spatial Variations in Rock Uplift From River-Profile Inversions at the Central Anatolian Plateau Southern Margin. J. Geophys. Res. Earth Surf., 126, e2020JF006027. https://doi.org/10.1029/2020JF006027

[15] Rehak, K.: Pliocene-Pleistocene Landscape Evolution in South-Central Chile [Ph. D. thesis], 2008.

[16] Vega-Ruiz, A., Delgado, V., Racano, S., Clementucci, R., Veliz-Borel, V., Espinoza, M., Encinas, A., Melnick, D., Larregla, R., Asenjo, C., and Zambrano, P., 2026. Tectonic and climatic controls on uplift transients of the Nahuelbuta Forearc Range, Northern Patagonian Andes. Earth Planet. Sci. Lett., 690, 120098. https://doi.org/10.1016/j.epsl.2026.120098

[17] Veliz-Borel, V., Mouslopoulou, V., Glodny, J., Begg, J., Metzger, S., Sakellariou, D., and Oncken, O., 2024. Exploring Uplift Mechanisms Across the Forearc of a Subduction System: Karpathos Island as a Natural Transect Across the Eastern Hellenic Margin. Tectonics, 43, e2023TC008156. https://doi.org/10.1029/2023TC008156

[18] Völker, D., Geersen, J., Contreras-Reyes, E., and Reichert, C., 2013. Sedimentary fill of the Chile Trench (32-46°S): Volumetric distribution and causal factors. J. Geol. Soc. London., 170, 723–736. https://doi.org/10.1144/jgs2012-119

Glacial lakes: a Himalayan hazard

You may or may not be familiar with glacial lakes. Around the world, as the cryosphere changes and glaciers thin and retreat, meltwater collects in depressions on glacier surfaces, at their margins and in recently deglaciated terrain, forming lakes that can expand and evolve rapidly. From the Himalayas and Andes to the Arctic and the Greenland Ice Sheet, these lakes are becoming one of the most visible signs of rapid cryospheric change.

Across High-Mountain Asia (HMA), also known as the ‘Third Pole’ as it contains the largest concentration of snow and ice outside the polar regions, glacial lakes are rapidly evolving in response to a warming climate. The region is experiencing some of the highest rates of glacial lake growth globally (Shugar et al., 2020; Zhang et al., 2024) and whilst they play an important role in storing and regulating meltwater runoff, they can also pose serious hazards to downstream communities and infrastructure.

As existing lakes become larger and new lakes form, concerns grow over their potential to generate glacial lake outburst floods (GLOFs). While many lakes are stable, some are held back by weak natural dams of sediment, rock debris or even remnants of glacial ice. If these dams fail, large volumes of water can be rapidly released, travelling tens of kilometers downstream and causing significant environmental, economic and societal impacts.

This is where the Glacial Lake Observatory (GLO) project comes in. The goal is to improve understanding of glacial lakes across HMA, how they’re evolving and the flood hazards they may pose. By combining satellite observations with drone surveys and lake depth measurements collected in the field, we can generate high-resolution maps and 3D models of lakes and their surrounding landscapes. These datasets allow us to better characterise lake geometry, provide field-based estimates of lake depth and volume, map moraine dams and provide detailed topographic information (that is currently lacking) for flood modelling and simulations.

Whilst satellite imagery and the processing of these datasets happen from the comfort of the office in Leeds (UK), the collection of high-resolution datasets to produce them begins in the field. In May-June 2025, this took the GLO team and I to Sagarmāthā National Park in Nepal, known to many as the Everest region. Famous for its towering mountains including Mount Everest, high-altitude trekking routes and spectacular scenery, it became the backdrop to four weeks of unforgettable fieldwork.

From satellites to reality

Before travelling to Nepal, I had already spent months looking at glacial lakes in the region from satellites. From my desk, I traced lake outlines, planned drone surveys and examined the valleys we would eventually visit. However, from a computer screen it is difficult to appreciate the scale of the landscape from hundreds of kilometers above the Earth surface.

A few months later, I found myself stepping off the small plane at Lukla airport, famous for its short (527 m), sloping (11%) runway and dramatic mountain setting. It was the gateway to the Everest region. Even before the trek began, I was blown away by the sheer size of the mountains (sounds cliché I know!).

From there, it was all on foot. Over the next four weeks, our route would take us through many of the major valleys and villages of the Khumbu region, from Thame and Gokyo in the west to Lobuche and Chukhung in the east, crossing two high mountain passes and, of course, surveying glaciers and lakes along the way (Figure 2).

Our first major destination was the village of Thame. Less than a year earlier in August 2024, a glacial lake outburst flood (GLOF) had swept through the village, destroying homes, the local school and damaging hydropower infrastructure. Approaching the village, the impacts were clearly visible with wide landscape scars, huge boulders strewn across the valley floor and deep cracks in the ground (Figure 3a). The impacts became even more striking when we visited what had been the local school, with classrooms filled with mud, debris and scattered textbooks (Figure 3b). Luckily no casualties occurred here, however it is a stark reminder of the human impacts of such events. The Upper and Lower Ngole glacial lakes, the source of the cascading GLOF, were 10 km upstream (Figure 3c) and the first of the lakes to be surveyed on this Everest region round trip, and my first experience of trekking to nearly 5000 m (above sea level, or a.s.l).

After a few days of hiking and rest, we reached the Ngole lakes (Figure 3d). Five clustered turquoise blue lakes, two of which the cause of the 2024 GLOF – the sheer scale of the landscape very much hard to comprehend from my computer screen in the office. Surveying began!

The Ngole lakes, however, were just the beginning…

Over the following weeks, our route took us across much of the Khumbu region, surveying lakes at the margins of glaciers dammed by loose sediment and rock, while others spanned the surface of glaciers itself. No two lakes were the same, each coming with its own logistical challenges, whether that be the distance travelled to reach them or navigating the maze of rocks, ice cliffs and meltwater on debris-covered glacier surfaces.

But the field season was about far more than just surveying glacial lakes. It was also connecting with and appreciating the landscape, the culture and the local people who call these mountains home. Between survey sites, we spent days trekking with our local guides through spectacular (and sometimes sketchy) terrain marked with colourful prayer flags, crossing suspension bridges high above roaring glacial rivers, climbing high mountain passes and passing yak and donkeys transporting supplies between remote settlements. The tinkling of yak bells became our trekking soundscape. Evenings were spent in family-run teahouses, with welcoming hosts fuelling us with Dhal Bhat (a traditional Nepali meal of lentil soup, rice and vegetables) and countless cups of Chiya (a sweel Nepali milk tea) whilst warming up around the communal stove often fuelled by dried yak dung.

Surveying the landscapes and waiting for blue skies

After hours (sometimes days) of trekking to reach a survey site, it wasn’t as easy as just turning up, unpacking and collecting data. The weather was in charge. Whilst trekking determined where we could survey, and altitude determined how fast we got there, the weather dictated when. Mornings began much in the same way: unzip the sleeping bag, open the teahouse curtains and look outside. A clear morning, yay! Time to get moving.

Drone surveying started early to get ahead of the afternoon cloud that regularly drifted into the valleys. Sometimes I spent hours waiting for a clear window. Then it was GO GO GO. Batteries unpacked, flight plans checked and the drone launched (Figure 4). The drone was flown back and forth in a grid-like formation across the glacier, lake(s) and valley, capturing hundreds of overlapping photographs. Cold temperatures limited flights to around 20 minutes before batteries needed changing and recharging from a portable power bank. To help maintain line of sight, I often walked several kilometres along the moraine as it followed its survey route.

Of course, no field campaign is complete without unexpected equipment issues. One of our larger drones decided it had no interest in flying in the Khumbu (likely due to geofencing restrictions at the time). Thankfully, we had a backup drone with us, which saved the day and allowed the surveys to continue (Figure 4). A challenge I had also not fully appreciated beforehand was helicopters. Even outside of the main tourist season, helicopters would zip up and down valley, particularly to and from Everest base camp. This meant constantly scanning the sky and listening out for the hum of rotor blades, and either lowering the drone or landing back at the moraine until the aircraft passed.

What about below the lake surface?

Whilst I focused on collecting aerial imagery and high-precision GPS (GNSS) measurements of the lakes, moraines and surrounding valleys, GLO Principal Investigator Scott Watson carried out depth surveys, known as bathymetry surveys, from a small inflatable kayak (Figure 5). Using a sonar system mounted on the kayak, he mapped the depth and shape of the lake bed while paddling along parallel survey lines. Combined with the drone imagery, these datasets allow us to build detailed 3D models and better estimate lake water volume. From the moraine ridges, I could often see the kayak moving as a tiny speck in the distance.

Slowly, slowly

An important aspect of the field trekking was shaped by the need to acclimatise to the increasing altitude. The route was carefully planned with rest days and gradual ascents to allow our bodies to adjust to the elevation. After all, with survey sites regularly hitting altitudes of 4500+ m, and the mountain passes of Renjo La and Cho La reaching upwards of 5400 m a.s.l, this was no easy task – it was both physically and mentally demanding.

A few days into fieldwork, I learned one of the most valuable Nepali phrases (actually, more the mantra), from my wonderful guide Mahesh: “slowly, slowly”. Alongside trekking and working in this remote, beautiful landscape, the altitude was a whole different ball game. At first, I thought the phrase was simply encouraging me to take my time. However, it didn't take long to realise that "slowly, slowly" was essential. Despite the acclimatisation schedule, there were days when my brain and body struggled. Headaches occasionally crept in, appetite reduced, and even relatively gentle uphill sections left me out of breath, with progress often becoming a case of stop, start. Mentally, I found this particularly frustrating, having to accept that the mountains would dictate my pace, not me.

Nowhere was that lesson clearer than on the two high mountain passes of Renjo La (5,360 m a.s.l.) and Cho La (5,420 m a.s.l.). On a map, they looked like little more than lines connecting neighbouring valleys. On foot, they were anything but. Renjo La rewarded hours of climbing with spectacular views across Ngozumpa Glacier and the turquoise Gokyo lakes (Figure 6). A few days later, Cho La brought another demanding crossing on the 24 km trek to the village of Lobuche. Here, overlooking Khumbu Glacier from the moraine ridge, the remnants of Everest Base Camp could be seen upvalley (Figure 7). By then, Mahesh's mantra of "slowly, slowly" had become second nature.

Looking back (and forwards…)

By the end of the field campaign, we had crossed much of the Khumbu (Everest) region. We surveyed 30+ lakes, collected thousands of drone images, and covered more than 200 km through some of the world's most spectacular mountain landscapes. Data on these lakes are now being transformed into detailed maps and 3D models, which alongside satellite observations, will help improve our understanding of glacial lake evolution, better estimate lake volumes and improve flood hazard modelling in the Himalayas.

As a whole, I learned a tremendous amount about this remarkable landscape and country and was fortunate to meet some truly incredible Nepalis along the way, including our wonderful trekking team, without whom none of the fieldwork would have been possible. I also learned a great deal about myself. From navigating the challenges of altitude and unpredictable weather to trekking across glaciers and high mountain passes, the experience pushed me well beyond my comfort zone and left me with memories that will stay with me forever.

Fortunately, this is not the end of the fieldwork story. In May and June 2026, we, the GLO team, returned to Nepal, this time to the Annapurna-Manaslu region, to continue surveying glacial lakes and downstream valleys (Figure 8). Another set of lakes, plenty more data collected and another chapter in understanding how this mountain landscape is changing.

Perhaps a story for another day...

You can watch a short outreach video of our fieldwork so far here!

With thanks to: Scott Watson (GLO PI); Rajendra Shrestha (NDRI); Mahesh Magar (Guide); Kaji Lama (Guide); Nepali chefs and porters; Himalayan Research Expeditions (HRE); Civil Aviation Authority Nepal; Department of National Parks and Wildlife Conservation.

References and further reading

• ICIMOD (2019) Glacial lake outburst flood video https://www.youtube.com/watch?v=awLGxl5JmjY

• Khadka, N., Liu, W., Shrestha, M., Watson, C.S., Acharya, S., Chen, X. and Gouli, M.R., 2025. Multidisciplinary perspectives in understanding Himalayan glacial lakes in a climate challenged world. Information Geography, 1(1), p.100002.

• Shugar, D. H., Burr, A., Haritashya, U. K., Kargel, J. S., Watson, C. S., Kennedy, M. C., Bevington, A. R., Betts, R. A., Harrison, S., and Strattman, K.: Rapid worldwide growth of glacial lakes since 1990, Nat Clim Change, 10, 939–945, https://doi.org/10.1038/s41558-020-0855-4, 2020

• Zhang, T., Wang, W., and An, B.: Heterogeneous changes in global glacial lakes under coupled climate warming and glacier thinning, Commun. Earth Environ., 5, 374, https://doi.org/10.1038/s43247-024-01544-y, 2024a

Edited by Monojit Saha and Leah Sophie Muhle 

Introduction

From industrial hydrogen to hydrogen made by the Earth

Why subduction zones matter

Looking for an invisible gas: a multi-proxy approach

From faults to fluids

• Structural characterization and strain-field analysis to identify faults and fracture networks that may act as pathways or traps for hydrogen-bearing fluids, enabling the assessment of their geometry, connectivity, and kinematic history, and thereby constraining fluid migration pathways, potential accumulation zones, and the temporal evolution of permeability within the system.
• Seismological imaging, using Local Earthquake Tomography (LET) to resolve fine-scale velocity variations related to fluids and altered rocks, and Ambient Noise Tomography (ANT) to image broader crustal structures independent of earthquake distribution.
• Geophysical methods, including Magnetotellurics (MT) and Transient Electromagnetics (TEM), to detect electrically conductive zones associated with fluids, alteration, or geothermal reservoirs.
• Geochemical and isotopic analyses of gas samples from soils, springs, and fumaroles, including hydrogen concentrations and noble gas isotopes (e.g. ³He/⁴He) to trace mantle versus crustal sources.

Why three study sites?

• T1 (Irruputuncu) is dominated by major strike-slip faults that provide direct pathways for deep, mantle-derived fluids, as evidenced by high ³He/⁴He ratios
• T2 (Láscar) features intersecting strike-slip and thrust faults, promoting fluid storage at depth and more complex circulation patterns, with stronger crustal influence.
• T3 (Cerro Pabellón) is controlled by normal faulting linked to crustal extension, where fluids circulate through highly differentiated volcanic rocks.

Why this matters for society

1. What is the geological origin of natural hydrogen in the Chilean Altiplano, and what concentrations can be expected?
2. How are natural hydrogen systems distributed, and what role do faults, volcanic structures, and regional stress fields play in their formation and storage?
3. How can integrated geochemical, structural, and geophysical approaches improve future hydrogen exploration strategies?

• Zgonnik, V. (2020). The occurrence and geoscience of natural hydrogen: A comprehensive review. Earth-Science Reviews, 203, 103140. https://doi.org/10.1016/j.earscirev.2020.103140
• Jackson, O., Lawrense, S.R., Hutchinson, I.P., Stocks, A.E., Barnicoat, A.C., Powney, M. (2024). Natural hydrogen: sources, systems and exploration play. Geoenergy2024-002, vol. 2. https//doi.org/10.1144/geoenergy2024.002
• Lévy, D., Roche, V., Pasquet, G., Combaudon, V., Geymond, U., Loiseau, K., & Moretti, I. (2023). Natural H2 exploration: tools and workflows to characterize a play. Science and Technology for Energy Transition, 78, 27. http://doi.org/10.2516/stet/2023021
• Moretti, I., Baby, P., Alvarez Zapata, P., & Mendoza, R. V. (2023). Subduction and hydrogen release: The case of Bolivian Altiplano. Geosciences, 13(4), 109. https://doi.org/10.3390/geosciences13040109
• Veloso, E.E., Tardani, D., Elizalde, D., Godoy, B.E., Sánchez-Alfaro, P.A., Aron, F., Reich, M., & Morata, D. (2020). A review of the geodynamic constraints on the development and evolution of geothermal systems in the Central Andean Volcanic Zone (18–28 Lat. S). International Geology Review, 62(10), 1294-1318. https://doi.org/10.1080/00206814.2019.1644678

About the author
Sergio León-Ríos is Associate Researcher at the Advanced Mining Technology Center (AMTC) at the University of Chile, working on seismology and its applications to mineral exploration. He holds a PhD in Natural Sciences from the Karlsruhe Institute of Technology (KIT), Germany. His research focuses on understanding the physical behavior of active margins, including megathrust earthquakes, crustal fault systems, magmatic fields, and their relationship with the emplacement of natural resources.

Shrishail Raut

Why Your Positioning Works: The Hidden Role of the IVS Combination Centre.

The TUDelft Global GRACE(-FO) Level-3 Equivalent Water Height Uncertainty Product.

Once a machine learning model performs well, what exactly has it learned? 

Opening the black box: a feature incorporation framework

Static versus dynamic predictors  

Main takeaway  

The journey to my first EGU General Assembly (GA) 2026 (EGU26) was memorable even before the conference began. What should have been a straightforward train ride to Vienna turned into an unexpected adventure when we were asked to get off at a small station and wait several hours for alternative transport. Surrounded by travellers, luggage, and a fair amount of confusion, it felt as though the EGU26 adventure had already begun before we had even reached Vienna. Fortunately, we eventually made it to Vienna, ready for the week ahead.

My first impression was simply how huge the conference was. Walking through the conference centre for the first time, I was astonished by the scale of the event and the diversity of researchers attending. Scientists from across the geosciences had come together to share ideas, discuss science, and learn from one another, creating an atmosphere unlike anything I had experienced before. One thing that stayed with me throughout the week was seeing researchers whose work I had only known through journal articles. It was exciting to see authors whose work I had been reading discuss their research firsthand. I still remember sitting through some talks thinking, I know this name from a paper I read a few weeks ago.

My first impression was simply how huge the conference was.

They were standing at posters, giving talks, answering questions, and discussing science over coffee. Seeing this transformed my view of research from something that existed on paper into a real, active, and collaborative community.

Discussing my poster at EGU 2026 

I attended many talks on waves and space plasma physics, which are closely related to my research interests. Listening to both observational and simulation studies helped me better understand how different researchers approach similar scientific questions. Several presentations sparked new ideas and gave me a broader perspective on the kinds of questions I might explore in the future.

One of the most important parts of the week was presenting my own poster on a statistical study of wave activity in the magnetosheath and foreshock regions. As someone who had only recently started working in this field, I found it a valuable opportunity to share my research with the wider community. I was particularly interested in hearing feedback, exchanging ideas, and learning how others approached similar scientific questions.

I was particularly interested in hearing feedback, exchanging ideas, and learning how others approached similar scientific questions.

After the first few discussions, I became more comfortable talking about my work and exchanging ideas with others. Scientists stopped by my poster, asked thoughtful questions, and engaged in interesting discussions about the results and their implications. I particularly appreciated how genuinely curious people were and how open they were to exchanging ideas. Some suggested possible directions for future work, while others offered insights to strengthen the project. What made these discussions especially meaningful was realising how interested people were in understanding my work and sharing ideas that could help move it forward. By the end of the session, I had gained valuable feedback, several new ideas to explore, and a clearer understanding of how I could further develop my research.

Beyond the scientific sessions, I attended several Solar-Terrestrial Sciences (ST) Division Early-Career Scientist (ECS) activities, including networking events for first-time attendees and ECS. Hearing about the experiences of fellow PhD students, ECS, and senior members of the community gave me a broader view of the different paths within our field. Some of the most memorable conversations I had during the week happened at these events, where discussions naturally moved from research projects to career experiences, plans, and the challenges of pursuing a PhD. Listening to these experiences gave me a greater appreciation of the variety of opportunities within the field and the supportive nature of the scientific community.

Listening to these experiences gave me a greater appreciation of the variety of opportunities within the field and the supportive nature of the scientific community.

The ST ECS Picnic/Meet the experts activity was another highlight of the week. I enjoyed the opportunity to step away from the conference halls and interact with community members in a more relaxed setting. Conversations often moved beyond specific research topics and into broader discussions about ideas, experiences, and future projects. It was interesting to see how many creative discussions and potential collaborations could emerge in such an informal environment. Unfortunately, I was unable to attend the ST ECS Dinner activity this year. Having heard so many positive things about it from colleagues, I definitely look forward to experiencing it at a future EGU.

At the ST-ECS Picnic

A particularly special moment during the week was witnessing the recognition of our professors and mentors, Prof. Jana Šafránková and Prof. Zdeněk Němeček, who received the 2026 Hannes Alfvén Medal. As a student from their group, it was a proud moment to see their contributions recognised by the wider community. Seeing their work celebrated was inspiring and made me appreciate even more the opportunity to be part of a group that has contributed so significantly to our understanding of space plasma physics.

Our research group, with Prof. Jana Šafránková and Prof. Zdeněk Němeček, at the Hannes Alfvén Medal award talk 

Looking back, what impressed me most was not only the scientific interactions but also the collaborative nature of our field. Events like EGU GA bring together people with different expertise, backgrounds, and ideas, creating opportunities for feedback, discussion, and future collaborations. These experiences showed me how scientific ideas develop through curiosity, questions, and the exchange of knowledge.

As a first-year PhD student, EGU26 was far more than my first conference. It was my first real introduction to the international scientific community and an opportunity to experience firsthand how research extends beyond papers and presentations. Throughout the week, I learned from talks, discussions, and the experiences shared by others, and I had the chance to present my own work and receive valuable feedback.

EGU26 was far more than my first conference. It was my first real introduction to the international scientific community and an opportunity to experience firsthand how research extends beyond papers and presentations.

I left EGU with new ideas for my research, a broader understanding of the field, and a deeper appreciation for the community that drives it forward. For a newcomer to the field, attending EGU26 was not only a memorable experience but also an inspiring start to my PhD journey.

Pragya Balot is a PhD student in Space Physics at Charles University. Her research focuses on understanding how solar wind structures interact with Earth's magnetosphere. In particular, she investigates the transfer of energy and mass across magnetospheric boundaries using multi-satellite observations combined with machine-learning techniques.

Why do coral reefs flourish along some continental margins, while elsewhere they form scattered platforms and isolated atolls? In this new blog post, Dr. Zhibin Lei explores this question through the lens of biogeodynamics, showing how tectonics can shape the physical foundations of coral reef habitability. From continental rifting to passive-margin subsidence and surface processes, geodynamic evolution helps create the shallow marine environments where coral reefs can establish, grow, and persist over geological timescales.

Coral reefs are among the most biodiverse ecosystems on Earth, supporting roughly a quarter of all marine species despite covering less than 1% of the ocean floor. Yet the geological conditions that allow these ecosystems to exist in the first place — a firm, shallow seafloor bathed in warm, clear water — are far from random. Walk the coastlines of the central and northern Red Sea, and you will find reefs packed into a narrow fringe clinging to the shore. Venture into the South China Sea and the picture is entirely different: hundreds of isolated atolls and submarine banks scattered across a vast oceanic plateau. Why does the same geological setting — a rifted passive continental margin — produce such radically different reef landscapes?

This is the question at the heart of our recently published study in Earth and Planetary Science Letters, which argues that the answer lies deep in Earth's geological past, in the style and pace of continental rifting itself.

                                 More than 75% of the world's coral reefs sit on rifted passive margins

Passive continental margins, the drowned edges of continents formed when plates rift apart, host the vast majority of the world's shallow coral reefs (Fig. 1). This is not a coincidence. The subsidence and morphological legacy of continental break-up creates exactly the kind of shallow, laterally extensive, hard-substrate terrain that corals need to settle and grow. But the specific geometry of that terrain, its extent, connectivity, and topographic relief, depends critically on how the rift evolved over geological time.

Coupling geodynamics with surface processes

We modelled the full rift-to-drift cycle, from initial continental extension through to seafloor spreading, using a coupled framework that links the ASPECT geodynamic code, which simulates lithospheric deformation and thermal evolution, with the FastScape surface process model, which tracks erosion, sediment transport, and deposition. The models run forward in time across tens of millions of years and track the area of seafloor that falls within the 0-30 m water depth window: the Habitable Basement Zone (HBZ) for shallow coral reefs.

We then benchmarked the model output mainly against two natural case studies spanning the range of global passive margins: the Red Sea and the South China Sea, while also briefly extending our findings to the Great Barrier Reef and mature passive margins on both sides of the Atlantic Ocean.

A four-quadrant framework for reef basement classification

Our results show that two factors, acting largely independently, control the geometry and area of the HBZ throughout the rift lifecycle (Fig. 2).

Rift type (wide vs. narrow) determines the structural template. Wide rifts, like those that formed the South China Sea, involve distributed faulting across a broad zone of extended crust. This produces numerous isolated fault blocks, rotated horsts, and submarine highs spread across a wide area, generating an archipelago-style reef landscape with HBZ areas more than twice as large as those produced by narrow-rift systems. Narrow rifts, like the Red Sea, concentrate deformation along a single narrow axial zone, yielding a linear margin with limited lateral variability.

Surface process intensity, meaning the efficiency of erosion and sediment redistribution, acts as a powerful amplifier. Among all the variables we tested, sediment transport is by far the most sensitive control on HBZ area. Efficient sediment delivery fills and smooths the rough rift topography, dramatically expanding shallow-platform area, in some cases increasing HBZ size by an order of magnitude relative to a tectonically equivalent margin with minimal sediment input. This finding helps explain why the southern Red Sea coast, which receives more fluvial sediment, hosts wider fringing reefs than its northern counterpart, and why the northern margin of the South China Sea, proximal to the South China mainland and its rivers, is morphologically distinct from the carbonate-dominated Nansha Islands to the south.

Geology lays the table; biology comes to dinner

One of the implications of our framework that we find most striking is that the gross architecture of today's coral reef biogeography, the difference between compact coastal reefs and dispersed oceanic archipelagos, was essentially written into Earth's geology tens of millions of years ago, long before any coral polyp ever settled on these substrates.

This does not mean biology and oceanography are irrelevant; of course they are critical for understanding which reefs grow, survive, and bleach under modern conditions. But the physical stage on which all of that plays out, the distribution, extent, and connectivity of shallow seafloor habitat, reflects a tectonic inheritance that operates on timescales of 10-100 million years.

A comparison of modern reef morphology against our numerical rift stratigraphy (Fig. 3) shows that the present-day Red Sea and South China Sea reef geometries are broadly consistent with the topographic predictions of narrow- and wide-rift evolution, respectively.

Towards a quantitative biogeodynamics

We think of this as a small contribution to an emerging way of working, call it biogeodynamics, that asks how deep-Earth processes shape the physical conditions for life over geological time. The four-quadrant framework is intentionally general: it applies not just to coral reefs but to any shallow marine ecosystem whose habitat depends on the geometry of tectonically sculpted seafloor.

Looking forward, the natural next steps are to couple this framework with dynamic coral growth and carbonate accumulation models, and to extend the analysis to other global margins, particularly those now experiencing rapid ecological change under climate warming. Understanding the geological baseline of reef habitat is a prerequisite for predicting how much room there is for reef ecosystems to migrate or persist as sea levels and temperatures continue to change.

We also hope this work demonstrates the value of cross-disciplinary approaches that connect mantle dynamics, crustal deformation, and surface processes to ecological outcomes, bridges that are still relatively rare in the geodynamics community but that we believe hold significant scientific promise.

Lei, Z., Xia, S., Brune, S., Pons, M., Zhang, C., Gou, T., Zhang, C., and Fan, C. (2026). Geodynamic controls on the habitable basement of shallow coral reefs at rifted continental margins. Earth and Planetary Science Letters, 687, 120073.

Making decades of observations comparable

What is new in version 3?

Open data still depend on people

Further reading

• Explore the GLODAPv3 website and learn more about the observations included in the new release.
• Access and download the GLODAPv3 merged and adjusted data product.
• Read the GLODAPv3 Data Use Statement for guidance on citing the dataset and recognising the scientists who collected the original observations.
• Learn more about the new quality-control methodology used in GLODAPv3.

References

Lange, N., Lauvset, S. K., Carter, B. R., Humphreys, M. P., Woosley, R. J., Olsen, A., Bittig, H. C., Kozyr, A., Álvarez, M., Azetsu-Scott, K., Becker, S., Brown, P. J., Cotrim da Cunha, L., Hoppema, M., Ishii, M., Jeansson, E., Murata, A., Müller, J. D., Pérez, F. F., Schirnick, C., Steinfeldt, R., Ulfsbo, A., Velo, A., and Tanhua, T. (2026). The Global Ocean Data Analysis Project version 3 (GLODAPv3) – an internally consistent biogeochemical data product for the World Ocean (NCEI Accession 0315582). NOAA National Centers for Environmental Information. Dataset. https://doi.org/10.25921/m6tp-mj50. Accessed 16 July 2026.
Lange, N., Lauvset, S. K., Carter, B. R., Humphreys, M. P., Woosley, R. J., Olsen, A., Bittig, H. C., Kozyr, A., Álvarez, M., Azetsu-Scott, K., Becker, S., Brown, P. J., da Cunha, L. C., Dias, L., Hoppema, M., Ishii, M., Jeansson, E., Murata, A., Müller, J. D., Pérez, F. F., Schirnik, C., Steinfeldt, R., Ulfsbo, A., Velo, A., and Tanhua, T. (submitted). The Global Ocean Data Analysis Project version 3 (GLODAPv3) – an internally consistent biogeochemical data product for the World Ocean.
Humphreys, M. P., Lauvset, S. K., Lange, N., Bittig, H. C., Carter, B. R., Hoppema, M., Murata, A., Olsen, A., Tanhua, T., Ulfsbo, A., Velo, A., Woosley, R. J., Azetsu-Scott, K., Müller, J. D., and Pérez, F. F. (2026). Furthest-first inversion for internal consistency adjustments in the biogeochemical data product GLODAPv3. EGUsphere [preprint]. https://doi.org/10.5194/egusphere-2026-3063.

]]> https://blogs.egu.eu/divisions/os/2026/07/20/glodapv3-release/feed/ 0 https://blogs.egu.eu/divisions/cl/2026/07/20/soot_warming_cooling_effects/ https://blogs.egu.eu/divisions/cl/2026/07/20/soot_warming_cooling_effects/#comments Mon, 20 Jul 2026 08:50:18 +0000 how much are they actually warming the planet? Estimates in the scientific literature vary widely. That is a troubling level of uncertainty when trying to understand - and ultimately manage - our climate future. Our new study, recently posted at Atmospheric Chemistry and Physics, attempts to explore a key reason for this confusion: the same amount of black carbon can behave in fundamentally different ways depending on where it is and what it looks like at the microscopic level. And those differences matter enormously for regional cooling or warming.

From the Sky to the Particle

When sunlight hits a black carbon particle floating in the atmosphere, two things can happen: the particle can absorb the light or scatter. Which effect dominates depends not just on how much black carbon is present, but on its shape, size, and what it is coated with. Fresh soot emerges from a combustion source as a relatively bare particle. But as it ages in the atmosphere over hours and days, it accumulates coatings of other materials (sulfates, nitrates, or organic compounds). This coating acts as a lens, concentrating more sunlight onto the absorbing core. The thicker the coating relative to the core, the more this "lensing effect" amplifies heat absorption. At the same time, heavily coated particles scatter more light. The net radiative outcome (warming versus cooling) depends on the delicate balance between these two competing processes. Current climate models largely sidestep this complexity. Many assume that black carbon particles are simple, externally mixed spheres with fixed optical properties, or that the relationship between the amount of black carbon in the atmosphere and its radiative effect is linear. Our study challenges both of these assumptions.

A Tale of Two Cities

We investigated black carbon aerosols over two very different urban environments: Xuzhou, China (a major industrial hub on the North China Plain), and Dhaka, Bangladesh; a rapidly expanding megacity in a tropical river delta, fed by traffic, brick kilns, and small-scale industries. To characterize the black carbon particles in these cities, we combined observations from multiple satellites (including the TROPOMI instrument and MODIS) with ground-based sensors (e.g., sun-sky photometer networks such as AERONET and SONET). These observations constrain a physical model that treats each black carbon particle as a sphere with an absorbing carbon core surrounded by a scattering shell. The contrast between the two cities was striking. In Xuzhou, black carbon particles tended to have smaller cores but thicker coatings, consistent with rapid secondary aerosol formation driven by diverse industrial co-emissions. In Dhaka, the cores were larger but more thinly coated, reflecting less efficient combustion and a less oxidative atmosphere. Even when the total amount of black carbon in the air column (as measured by its ability to block light) was comparable between the two cities, their microscopic differences produced different absorption-scattering balances, and therefore different climate effects. [caption id="attachment_5737" align="alignnone" width="1024"] Figure 1. Conceptual workflow of the Core-Shell Mie Model Optimization (COSMO) framework for quantifying regional Black Carbon (BC) radiative forcing. The flow diagram details the retrieval of key regional column products (BC size distribution, mixing state, columnmass, and number concentration) by integrating multi-waveband satellite observations. Particle light-scattering and light-absorption properties and radiation interaction of pollutant particles (asymmetry parameter (ASY), and single scattering albedo (SSA)) are iteratively computed using a core-shell configuration optimization at 388, 470, and 550 nm. Following quality assurance (discarding physically inconsistent retreivals), the final optimized optical properties are ingested into the radiative transfer model (SBDART) to calculate the multi-level (TOA, ATM, BOA) BC radiative forcing.[/caption]

A Machine Learning approach that keeps the Physics

Running detailed and complex radiative transfer simulations to calculate how light interacts with the atmosphere over large regions and long time periods is incredibly slow and expensive (computationally speaking). Because of this, many climate studies use simplified "forcing efficiency" formulas to estimate . We demonstrate that this approach misses critical information. Instead, we trained a machine learning surrogate model (a model that mimics the behaviour of complex and expensive simulations), specifically, a random forest ensemble (a model prediction based in multiple decision trees), on more than 400,000 individual radiative transfer simulations, using microphysical properties retrieved from our observational constraint framework as inputs. This model learns the full non-linear relationship between particle properties, column loading, and radiative forcing, without needing to re-run expensive physics calculations each time. The model achieved very high accuracy (explaining more than 95% of the variance in our physically derived forcing values) and, crucially, provides interpretable insights into what drives the forcing in each region. Using a technique called SHAP analysis (borrowed from cooperative game theory), we quantified how much each input variable contributed to top of the atmosphere forcing estimate. The results reveal a world that simple parameterisations cannot capture: the same amount of black carbon loading in the atmosphere can contribute either to cooling or to warming, depending on how the particles are mixed and distributed. There is no universal threshold that separates the two regimes. This non-linearity is not a modelling artefact; it reflects real-world physics.   [caption id="attachment_5738" align="alignnone" width="594"] Figure 2. Performance evaluation and feature importance analysis of predictive models for Black Carbon (BC) Top-of-Atmosphere (TOA) Direct Radiative Forcing (DRF). Panels (a–c) show the performance for Xuzhou and panels (d–f) show the performance for Dhaka. Density scatter plots compare the physics-based SBDART model (x-axis) against three predictive approaches (y-axis): (a, d) a standard Linear Model, (b, e) a Multiple Linear Regression (MLR) model, and (c, f) a Machine Learning (ML) model. The dashed red line represents the perfect 1:1 agreement, while the solid black line shows the actual regression fit. Statistical metrics (R2, RMSE, MAE, MBE, MAPE) are provided in the inset boxes, with warmer colors indicating a higher density of data points. The bottom rows display the interpretability analysis for both regions using SHAP (SHapley Additive exPlanations). The horizontal bar charts (left) rank the overall importance of input features (e.g., BCAOD₅₅₀, N_f, Mixing state, BC size) based on their average impact on the model's output magnitude. The corresponding beeswarm plots (right) illustrate how high (red) or low (blue) values of these features specifically contribute to increasing or decreasing the predicted DRF.[/caption]

Crossing Borders: Does the Model Transfer?

A critical test for any scientific tool is whether it works somewhere it was not explicitly designed for. We applied our combined machine learning model (ML^**) trained only on data from Xuzhou and Dhaka to two completely new regions: Delhi, India, during the autumn crop residue burning season, and Mongu, Zambia, during the peak savanna fire season. The transfer to Delhi worked reasonably well. This is perhaps not surprising: Delhi's black carbon, while heavily influenced by agricultural burning, shares enough microphysical similarity with the urban environments the model was already trained. Mongu, however, was a different story. Here the model systematically underestimated the radiative forcing by around 4 W m^-2 on average, a large error. The root cause? Savanna fire black carbon is physically distinct from urban black carbon. The particles emitted by these remote fires tend to have smaller cores and thinner coatings, and occur in very high column abundance, all outside the range the model had previously seen. This is not just a technical finding. It suggests that the current practice in global climate models, which often group all "biomass burning aerosols" into one generic category, may be fundamentally inadequate. The particles produced by savanna fires in Zambia behave differently from those produced by agricultural fires in India, which in turn behave differently from industrial urban soot in China. Once we re-trained the model with data from Mongu and Delhi, its accuracy drastically improved. The result was a large improvement at the biomass burning sites, reducing the error at Mongu by 68%, while maintaining accuracy at the original urban sites. The lesson is that machine learning tools for regional black carbon forcing need to be built on a training dataset that adequately represents the diversity of aerosol types they will encounter.   [caption id="attachment_5740" align="alignnone" width="732"] Figure 3. Statistical performance and transferability metrics of the machine learning surrogates across distinct regional aerosol regimes. The table contrasts the predictive capabilities of the baseline two-region combined model (ML**, trained only on Xuzhou and Dhaka data) against the expanded four-region model (ML***, incorporating Delhi and Mongu into the training domain). Performance is evaluated using, Adjusted Coefficient of Determination (Adj. R2), Root Mean Square Error (RMSE W m-2), and Mean Bias Error (MBE W m-2). Red values denote the baseline ML** framework, while blue values highlight the optimized ML*** framework. The comparative metrics demonstrate that expanding the training feature space dramatically mitigates the systematic underprediction bias observed over Mongu's savanna-fire regime while maintaining or refining robust predictive accuracy within the native East and South Asian domains.[/caption]

What This Means

Black carbon is a short-lived pollutant. Unlike carbon dioxide, which persists in the atmosphere for centuries, black carbon typically stays aloft for only one to two weeks. This means that reducing emissions can have immediate climate benefits, but only if we accurately know where the warming is happening and how large it is. Our study provides a framework that is both physically rigorous and computationally practical: it uses satellite observations to constrain the microphysical state of particles, feeds these through detailed radiative calculations, and distils the results into an efficient and interpretable machine learning tool. The broader message is one of irreducible regional specificity. The climate impact of black carbon cannot be read from a single number describing how much is in the air. It depends on what the particles look like, what they are coated with, how many of them there are, and where they sit in the atmospheric column. Understanding and ultimately mitigating  black carbon's role in climate change requires coming to terms with this complexity, not smoothing over it. Read the full open-access study in Atmospheric, Chemistry and Physics here.

This post has been edited by the editorial board

References:
Tiwari, P., Cohen, J. B., Gao, H., Lu, L., Wang, J., Dubovik, O., and Qin, K.: Microphysical evolution and column loading drive nonlinear regional contrast in black carbon top-of-atmosphere forcing, Atmos. Chem. Phys., 26, 9149–9180, https://doi.org/10.5194/acp-26-9149-2026, 2026.

 ]]> https://blogs.egu.eu/divisions/cl/2026/07/20/soot_warming_cooling_effects/feed/ 0 https://blogs.egu.eu/divisions/gm/2026/07/16/when-mud-flows-behave-like-glaciers-discovering-the-secrets-of-azerbaijans-mud-volcanoes/ https://blogs.egu.eu/divisions/gm/2026/07/16/when-mud-flows-behave-like-glaciers-discovering-the-secrets-of-azerbaijans-mud-volcanoes/#comments Thu, 16 Jul 2026 09:00:16 +0000 Introduction When people hear the word „volcano“, they usually think of a magmatic volcano with impressive pyroclastic eruptions or lava flows. However, mud volcanoes are different. Mud volcanoes erupt what we call mud breccias, a mixture of gas, water, and fine sediments. The eruptions generally occur due to deep generation of hydrocarbons and gravitational instability of rapidly buried, buoyant sediments. In Azerbaijan, where more than 400 kilometre-sized mud volcanoes exist, they literally shape the landscape. Mud volcanoes can, like their magmatic cousins, have massive eruptions and kilometre-long mud flows. For a long time, geoscientists assumed that these extensive mud flows formed during rare, catastrophic eruptions that expelled enormous volumes of material in a short time. It seemed like the only reasonable explanation. After all, how else could a kilometre-long flow appear? There are even records for extensive mud flows during eruptions, the latest being on September 23^rd, 2018 on the Otman Bozdag mud volcano. However, in 2019, this certitude started to crumble as our work at the Lokbatan mud volcano opened a new perspective. A Puzzle at Lokbatan and Otman Bozdag While working on my master thesis on mud volcano dynamics, I realised how many questions remain unanswered about the behaviour of large mud volcanoes, and the 2018 eruption of Otman Bozdag offered a rare window into these processes. In parallel, during an Azerbaijani summer school which took place at the same time, my future colleagues Petr Brož and Adriano Mazzini were investigating the Lokbatan mud volcano, one of the most active mud volcanoes on Earth. Historical records document numerous eruptions, and its western flank hosts a mud flow longer than a kilometre. Standing on top of the structure, the flow resembles a frozen river of grey sediment winding across the landscape. [caption id="attachment_2944" align="alignright" width="300"] Field photo of the Lokbatan mud volcano flow crossing a road. PC: Petr Brož.[/caption] So, what began as a routine field investigation for my colleagues evolved into a surprising discovery. They found very recent fractures encircling the presumed paleo-flow, which suggests the largest mud flows may not be built during single eruptions at all! Instead, it seems like mud volcano flows can slowly creep downslope over many years. This realisation led us on a scientific journey from muddy field sites to satellite imagery and ultimately changed how we think about mud volcano dynamics. Following eruptions in 2022, 2024, and 2025, our team visited the Lokbatan mud volcano, and what we found confirmed the first discovery. The eruptions themselves were relatively small. The volumes of newly erupted mud were nowhere near large enough to explain the existence of the enormous mud flow extending downslope. Yet, each time, fresh deformation had appeared across the surface. Walking along the margins of the paleo-flow, we encountered long cracks running parallel to the direction of movement, as well as roads that were cut and pipes that were bent. Near the crater, newly erupted material had piled up against older deposits, producing structures that looked as if the flow had been pushed from behind. The landscape was showing us something that did not fit the traditional model. Watching a Mud Flow Move Field observations provided important clues, but we needed to see what was happening on a bigger scale and find evidence that the entire flow was actually moving. Back in the office, we turned to satellite imagery. Google Earth's historical archive allowed us to compare images collected over the years. We selected Lokbatan first. As we inspected the historical images of the flow, distinctive surface features began to reveal something remarkable: they indeed had changed position. And not only small mud breccia, but blocks that are more than 10 m tall moved across the years. The movement was not dramatically fast. No sudden landslide or debris flow had occurred. Instead, the flow had gradually shifted downslope. To verify this behaviour, we combined our observations with InSAR data, a remote-sensing technique detecting subtle ground deformation from space. Together, the datasets painted a consistent picture. The kilometre-scale mud flow was creeping downslope, year after year. Just as temperate glaciers move through a combination of internal deformation and sliding over a lubricated base, the Lokbatan mud flow appeared to be advancing through a similar process. Newly erupted material accumulated near the crater, increasing the load on older deposits. Beneath the flow, a liquid-rich layer reduced friction and allowed the mass above to slowly slide downhill. [caption id="attachment_2942" align="alignleft" width="271"] Satellite image with mud volcano locations and eruption timelines based on published data from Fenske et al., 2026 (https://doi.org/10.5194/esurf-14-433-2026).[/caption] Suddenly, the kilometre-long flow no longer required a gigantic eruption to explain its existence. It could grow incrementally through repeated episodes of slow movement. However, scientific discoveries often raise more questions than they answer: Was Lokbatan unique? Or had we overlooked a common process operating across Azerbaijan's marvellous mud-volcano province? To find out, we expanded our satellite investigation to dozens of mud volcanoes distributed throughout the Caspian Basin. It became an exercise in landscape detective work. We compared image after image across decades, searching for signs of movement. The results did not disappoint. Evidence of creeping mud flows was identified at 19 mud volcanoes, out of 47 randomly selected ones. Some moved only a few metres over a decade, while others displayed much more impressive displacement. At Goturdag mud volcano, for example, movement could be traced along the entire length of the flow for years. Field visits revealed sharp boundaries between active moving deposits and stable slopes, making the creeping process visible directly in the landscape. The creep displaced tens of metres of mud across a decade, burying valleys at the bottom of its slope. However, not every mud volcano behaved the same way. Some appeared to move only after eruptions, while others showed nearly continuous deformation. Yet the overall pattern was clear: creeping mud flows are not an isolated curiosity. They represent an important geomorphic process shaping many Azerbaijani mud volcanoes. Reading the Landscape Differently One very rewarding aspect of geomorphology is learning to see familiar landscapes in new ways. Before this work, kilometre-scale mud flows were interpreted as products of rare, massive eruptions. Today, we understand that many may instead represent the cumulative result of small events acting over decades. This shift in perspective also carries practical implications. Slow-moving mud flows can deform pipelines, roads, and other infrastructure. In some locations, they even threaten nearby settlements. A landscape that appears stable may still be moving centimetre by centimetre beneath our feet. For us, the most exciting aspect of this discovery was how it emerged from combining field observations with modern monitoring techniques. A few unexpected fractures in the mud led to satellite investigations, geophysical surveys, and ultimately a new conceptual model for how these landforms evolve. Closing Highlight Geomorphology often advances when landscapes reveal behaviours we did not expect. At Azerbaijan's mud volcanoes, a process hidden in plain sight turned out to be shaping entire landforms. By following clues from the field and combining them with satellite observations, we discovered that mud flows can behave much like glaciers, slowly creeping downslope long after eruptions have ended... It is a reminder that even familiar landscapes can still surprise us when we take the time to watch them carefully. This work has been published recently at Esurf and Geology, so go take a look if you want to know more! Thanks to my colleagues Petr Brož and Adriano Mazzini to make this collaboration and work happen!   References: Mazzini, A., Brož, P., Lupi, M., Břežný, M., Jodry, C., Sfalcin, J., Fenske, C., Morelli, G., Fishanger, F., Huseynova, A., Huseynov, A. : Mud volcanism and creepy mud flows: A new model. Geology 2026; doi: https://doi.org/10.1130/G54583.1 Fenske, C., Brož, P., and Mazzini, A.: Mud volcano dynamics in Azerbaijan: the overlooked role of creeping mud flows in landscape evolution, Earth Surf. Dynam., 14, 433–442, 2026, https://doi.org/10.5194/esurf-14-433-2026]]> https://blogs.egu.eu/divisions/gm/2026/07/16/when-mud-flows-behave-like-glaciers-discovering-the-secrets-of-azerbaijans-mud-volcanoes/feed/ 0 https://blogs.egu.eu/divisions/gmpv/2026/07/16/geology-in-bjorks-soundscapes-from-tectonic-metaphors-to-emotional-terrains/ https://blogs.egu.eu/divisions/gmpv/2026/07/16/geology-in-bjorks-soundscapes-from-tectonic-metaphors-to-emotional-terrains/#comments Thu, 16 Jul 2026 09:10:59 +0000

Iceland is pure geology. When I travelled there in 2015, it felt like stepping into another world: surreal, raw, and unforgettable. Geology wasn’t just something to observe or study: in Iceland, it’s something you inhabit. The landscape doesn’t just sit there, it feels alive. It hums, it shifts, it speaks. And whether you’re a geologist or not, you can’t help but listen. To make the experience even more immersive, my friends and I put together the perfect playlist: Icelandic artists like Of Monsters and Men, Sigur Rós, Kaleo, and, of course, the one and only Björk.

Biophilia (2011)

Vulnicura (2015)

I have a confession to make before we start: I cannot stand Odysseus. Yes, he is dramatic and clever and the Odyssey is a masterpiece, but he is also constantly crying. “Oh, you crossed the sea, sacked cities, killed thousands of people and now you cannot get home? How tragic.” The worst part is when he is finally sailing back to Ithaca and keeps stopping to tell strangers how much he has suffered—only to break down again and postpone the actual story until after someone has given him a whole cow, a couple of litres of wine, a mountain of fruit and a very comfortable bed. Maybe tomorrow, after another good cry, he will be able to explain in more detail how he killed those “people” in the country next door. He is the hero of the poem, but he is also an excellent example of a drama queen, world's greatest overthinker.

And yes, I know it is “only” a epic poem (I did read it, I promise), but that does not change the fact that Odysseus is one of the most exhausting protagonists I have ever met. So now that I have been at least as dramatic as he is, we can get to the point: we are here for the geology (I guess).

Before we start moving faults around, a tiny bit of geography. We are in the Ionian Sea, off western Greece. Picture a little cluster of islands: big Kefalonia in the middle, long and mountainous; to its east, smaller Ithaki, steep and rugged; to its west, a lumpy peninsula called Paliki, sticking out into the open sea like Kefalonia’s tectonically confused arm. Paliki is connected to the rest of Kefalonia by a narrow, 6‑kilometre‑long strip of land called the Thinia isthmus. On modern maps, “Ithaca” is the island called Ithaki. On Homer’s map, as we will see, things are not that simple.

Over the last two decades, this literary annoyance has turned into a full‑blown geoscience project. The “Odysseus Unbound” hypothesis proposes that Paliki was once a separate island—the real Ithaca—cut off from Kefalonia by a narrow marine channel described by the ancient geographer Strabo as “so low‑lying that it is often submerged from sea to sea”. Testing that idea has dragged an impressive arsenal of methods—gravity, resistivity, seismic refraction, helicopter‑borne electromagnetics, cores and coastal stratigraphy—into what is, at heart, an argument about a poem. In this post, we will follow that investigation across the Thinia valley and see how much geodynamics can (and cannot) say about where Odysseus (looooser) called home.

Land ho!

Remember (or let me clarify) that The Odyssey is an epic poem, not a modern novel like The Lord of the Rings. It leans heavily on repetition, formulaic phrases, extended metaphors and extravagant, evocative names to describe people, places and events—which makes it both harder and more fun to interpret geologically. When Homer describes an island, he is not writing a GPS manual; he is writing poetry. Great for literature, terrible for georeferencing (try to write a paper like that and we will see).

When Odysseus finally introduces himself to the Phaeacians in Book 9 (after a good cry, obviously), he gives one of the most argued‑over geographical descriptions in ancient literature. In most translations (I read it in Catalan, but the idea is the same), Ithaca is “bright” or “clear‑seen”, with a forested mountain (Neriton) visible from afar, surrounded by other islands—Dulichium, Same and wooded Zacynthus—while Ithaca itself “lies low” and is “the furthest towards the west, toward dusk; the others lie apart toward dawn and the sun”. It is a rugged land and a good “nurse of young men”, not a flat sandbank, but the combination of “low‑lying” and “furthest west” has become the key to every Ithaca treasure map since antiquity. Read it by yourself here:

εἴμ’ Ὀδυσεὺς Λαερτιάδης, ὃς πᾶσι δόλοισιν

ἀνθρώποισι μέλω, καί μευ κλέος οὐρανὸν ἵκει.

ναιετάω δ’ Ἰθάκην ἐυδείελον: ἐν δ’ ὄρος αὐτῇ

Νήριτον εἰνοσίφυλλον, ἀριπρεπές: ἀμφὶ δὲ νῆσοι

πολλαὶ ναιετάουσι μάλα σχεδὸν ἀλλήλῃσι,

Δουλίχιόν τε Σάμη τε καὶ ὑλήεσσα Ζάκυνθος.

αὐτὴ δὲ χθαμαλὴ πανυπερτάτη εἰν ἁλὶ κεῖται

πρὸς ζόφον, αἱ δέ τ’ ἄνευθε πρὸς ἠῶ τ’ ἠέλιόν τε,

 

I am Odysseus, Laertes’ son, world-famed

For stratagems: my name has reached the heavens.

Bright Ithaca is my home: it has a mountain,

Leaf-quivering Neriton, far visible.

Around are many islands, close to each other,

Doulichion and Same and wooded Zacynthos.

Ithaca itself lies low, furthest to sea. Towards dusk; the rest, apart, face dawn and sun. Od 9.19-26 (text from http://www.odysseus-unbound.org)

 

Now put that next to the real map again. Modern Ithaki, the island that has carried the name since classical times, is steep and mountainous, with high peaks that do not immediately scream “low‑lying”. In the Ionian group, it also sits to the east of Kefalonia, not at the extreme western edge. If you were an ancient sailor heading into the sunset, the land that feels “furthest towards the west” is not Ithaki at all, but Paliki on the western side of Kefalonia—except that Paliki is not an island today, but a peninsula tied on by the Thinia isthmus.

So we have a problem: either Homer was very bad at counting islands (possible, but not very satisfying), or the coastline has changed since the Late Bronze Age (more interesting), or our modern habit of slapping ancient names onto present‑day shapes is misleading us. For a long time, many scholars picked the first option and treated these contradictions as proof that Homer’s geography was essentially fictional, or at least hopelessly distorted by centuries of oral performance. That attitude began to soften after the late nineteenth‑century excavations at Hisarlık, widely accepted as the site of Troy, showed that places from the Iliad were rooted in real Bronze Age cities rather than pure invention. If Troy could be dug up with a spade (and partly destroyed in the process), then perhaps Ithaca, too, deserved the indignity of a gravity survey.

Strabo’s clue and the Odysseus Unbound hypothesis

Centuries after Homer, the geographer Strabo added a bit more mystery. Describing Kefalonia in his Geography, he mentions that “where the island is narrowest it forms an isthmus so low‑lying that it is often submerged from sea to sea”, and even locates it between the territories of Pale (on Paliki) and Cranii (on eastern Kefalonia), exactly where the Thinia valley now sits as the only land bridge between peninsula and main island. In other words, Strabo seems to remember a place on Kefalonia that behaved less like a stable ridge and more like a half‑drowned sill.

This is the sentence that launched Odysseus Unbound Foundation to work on "The Search for Homer’s Ithaca". In his 2005 book, Robert Bittlestone, working with philologist James Diggle and geologist John Underhill, proposed that during the Late Bronze Age and classical period Paliki was not a peninsula at all, but a separate island divided from Kefalonia by a narrow marine strait running through Thinia. In their reconstruction, repeated large earthquakes along nearby faults triggered catastrophic landslides and rockfalls from the steep limestone walls, gradually choking “Strabo’s Channel” with debris and leaving the former sea passage literally high and dry. Paliki would then have been the low‑lying, westernmost island of Homer’s description; Thinia would be the scar where a geodynamic process quietly rewrote the map while nobody was watching.

It is a wonderfully audacious idea: a classicist, annoyed by a few lines of Greek, accidentally invents a geomorphology problem and then drags half of geophysics into solving it (and me writting this).

How do you X‑ray a mythic island?

Once you decide to take Odysseus seriously—his descriptions, not necessarily his personality—Thinia turns into a list of very blurry questions:

• Was the Thinia saddle ever close to sea level in the last few thousand years, or has it been high ground for much longer?
• Is there a buried channel—a genuine “Strabo’s Channel”—hidden beneath the present valley floor?
• If there was once a seaway, how much rock and sediment would you need to move to lift it 180-200 metres above sea level, and can earthquakes and landslides realistically do that job?

Answering those questions without digging a trench across an entire Greek island requires a lot of indirect vision. The Thinia project (geological and classical research initiative focused on the Greek island of Kefalonia, primarily seeking to locate the true historical home of Homer’s Odysseus) therefore turned into a kind of geophysical forensics lab: instead of fingerprints and DNA, the team used tiny variations in gravity, electrical conductivity and seismic wave speed to sketch the shape and history of the subsurface.

Gravity and airborne electromagnetic (EM) surveys provided the first wide‑angle X‑ray. Loose valley fill—landslide debris, soils, colluvium—tends to be less dense than solid limestone bedrock, and often more conductive if it is clay‑rich or water‑saturated. By flying EM instruments over northern Paliki and measuring tiny changes in the Earth’s gravitational and electromagnetic fields, the team could estimate how thick the low‑density, conductive material beneath Thinia is and where denser bedrock rises closer to the surface. A deep, continuous trough of low‑density material might signal a former channel; a patchy, shallow infill hints at a more modest valley.

Seismic methods offered a sharper, more local picture. On land, seismic refraction profiles used controlled sources and lines of geophones to track how fast waves travelled through the subsurface: slow through loose sediments, faster through compacted marls and fastest through hard limestone. Offshore, shallow seismic reflection in the adjacent bays imaged buried erosional surfaces and palaeo‑valleys cut into older units, including a deeply incised drainage system beneath the modern Gulf of Argostoli that records how water flowed when sea level was much lower during the last glacial maximum. Together, these data show where former valleys and potential channel routes would have been, even if later sediments have partly filled them.

Finally, resistivity surveys and boreholes/cores did the unglamorous but essential “ground truthing”. Electrical resistivity measurements helped distinguish more resistive bedrock from wetter, finer‑grained sediments, while 17 shallow cores across key sites in Thinia and nearby marshes yielded actual material to look at under the microscope. Marine microfossils, algae and sedimentary structures in these cores allowed biostratigraphers to say “this was a shallow sea” or “this was a lake or wetland”, and to assign ages to those environments using established microfossil ranges. That is how you turn a poetic “isthmus often submerged from sea to sea” into a timeline of when, exactly, salt water last occupied the valley.

Put together, it really does feel like CSI: Ithaca (yeaaaaaahhhh, hope you read this like CSI opening): each method is one more piece of forensic evidence brought in to interrogate a few lines of Homer and a throwaway remark from Strabo. The surprise is that the victim on the table is not just a missing island, but also our assumptions about how fast landscapes can change on human timescales.

Building (and questioning) Strabo’s Channel

The first wave of work in Thinia seemed to play nicely with Strabo. Early geophysical soundings and boreholes found marine sediments beneath the present valley floor, including fine‑grained deposits with marine microfossils that clearly formed below sea level. In the adjacent coastal embayments, shallow seismic lines and geomorphic mapping revealed buried drainage features and palaeo‑valleys aligned with the Thinia saddle, as if a former waterway had once connected the Gulf of Argostoli in the south with the bay to the north. On the valley walls themselves, scarps and hummocky topography testified to large landslides and rockfalls cascading off the steep limestone slopes—a ready‑made mechanism for dumping huge volumes of debris into any channel that might have existed.

Zooming out, the tectonic context makes such a fragile strait seem almost over‑determined. Kefalonia sits in the outer Hellenide belt, where the African plate is colliding with Eurasia and where the Pre‑Apulian units are being shortened and uplifted along major structures like the Aenos Thrust. Just offshore, the dextral Kefalonia Transform Fault accommodates part of the relative motion between the Aegean and Adria microplates, generating frequent moderate to strong earthquakes, including the infamous 1953 sequence that produced shaking equivalent to roughly magnitude 7 and caused extensive damage across the island. This combination of active thrusting and strike‑slip faulting creates exactly what you see around Thinia: steep relief, oversteepened valley sides and a landscape primed for slope failure.

The conceptual model that emerged from these pieces is appealing in its simplicity. Start with a low‑lying marine strait at Thinia—a narrow sill just above or at sea level, perhaps already cut into relatively young marine marls. Shake it repeatedly for a few thousand years with earthquakes on the Kefalonia Transform and related faults; each major event triggers rockfalls and landslides from the canyon walls, sending blocks and debris tumbling into the channel. At the same time, long‑term regional uplift on the outer arc slowly raises the whole area, turning yesterday’s sea passage into today’s 180-200 metre high valley floor. In this picture, Strabo’s “isthmus so low‑lying that it is often submerged from sea to sea” is not poetic exaggeration, but a snapshot of a strait in the process of being choked and lifted by the combined effects of tectonics and gravity.

Poseidon's fury: when did the sea really leave?

This is where the story stops behaving like a simple detective novel. A three‑year investigation of Thinia threw almost every available geophysical tool at the problem—helicopter‑borne EM, detailed gravity and resistivity surveys, seismic refraction on land, shallow‑marine reflection offshore, plus 17 shallow cores across the valley and nearby coastal sites. When all of these data were stitched together, the subsurface did not look like a Holocene channel casually infilled since Homer’s time. Instead, the valley fill turned out to be dominated by steeply dipping, tectonised marine sediments of Early Miocene to Early Pleistocene age, deformed and uplifted, with only relatively thin veneers of younger colluvium and landslide material on top.

In several places, bedrock cropped out close to the surface, and neither the onshore profiles or the offshore lines managed to trace the clear sides and bottom of a young, sea‑to‑sea channel at anything like present sea level. Biostratigraphic work on the cores pushed the last unequivocal marine conditions in the Thinia area back to around 1.8 million years ago, at the Gelasian-Early Pleistocene boundary, and failed to recover any younger, obviously marine sediments that would indicate a Late Quaternary strait. Taken together, these results suggest that if a through‑going channel once connected the two coasts at Thinia, it probably disappeared hundreds of thousands of years before the Late Bronze Age, not a few millennia ago.

In other words, the big plot twist is that the rocks insist on a much older, messier history than the neat picture of a Bronze Age seaway rapidly filled by historical landslides. Strabo’s low‑lying isthmus may still record some genuine geomorphic quirk, but the deeper geodynamics point to long‑term uplift and Miocene-Pleistocene tectonics, not a channel that closed just in time to annoy Odysseus (he's annoying enought).

What geodynamics can (and cannot) do for Homer

At this point it is tempting to ask geodynamics to settle the argument once and for all: did Paliki match Homer’s Ithaca or not? The Thinia studies show very clearly what the rocks are willing to tell us, and where they fall stubbornly silent. On the “yes, we can” side, geology and geophysics can constrain when different parts of Kefalonia were above or below sea level, how quickly uplift has raised marine sediments to hundreds of metres elevation, and how much material could realistically have been moved by earthquakes, landslides and long‑term slope failure. They can rule out a Holocene sea‑to‑sea channel through Thinia with a straight face, and they can sketch a plausible tectonic history involving thrusts, normal faults and gravity‑driven deformation linking structures like the Atheras Thrust and Agia Ioanni Fault.

What they cannot do is fill in the human details. No amount of biostratigraphy can tell us whether Homer (or the poets behind “Homer”) personally set foot on Paliki, or whether “low‑lying” referred to absolute elevation, gentle relief compared to neighbouring peaks, or perhaps even to a harbour rather than the whole island. Different datasets, collected at different times and scales, also leave room for interpretation: early boreholes and seismic profiles that suggested thick landslide fill and young marine fossils beneath Thinia supported a recently infilled channel, while later, broader surveys and additional cores pointed towards an older, more tectonically controlled story. That kind of tension is normal when geology, geophysics and literary history collide; it is part of what makes this case so instructive for students on both sides of the Aegean.

The only firm conclusion, for now, is that the rocks have certainly moved. Whether Homer’s Ithaca moved with them is still up for peer review.

Better if Odysseus had never reached Ithaca

Odysseus, famously, needs ten years and a lot of divine dysfunction to get home (looooser). Geoscientists have now spent several decades trying to pin that home down on a faulted, uplifted carbonate platform in the outer Hellenides—and the end of the journey is no less ambiguous. Ithaca has become a moving target in two senses: a landscape actively reshaped by plate convergence, earthquakes and sea‑level change, and a literary place whose meaning shifts as new data arrive from boreholes and seismic profiles.

This matters for more than classical trivia. The same tools used to chase Odysseus—gravity surveys, reflection seismology, coastal cores—are the ones we use to reconstruct palaeogeography and relative sea‑level change in active margins, to assess tsunami and landslide hazards, and to understand how fast coasts can rise, fall and rearrange themselves on human timescales. The Ithaca story is a neat reminder that geoscience can meaningfully test ideas from archaeology and literature, sometimes supporting them, sometimes complicating them, and often replacing a simple answer with a better question.

And maybe the ultimate test of all these ideas would be a field trip: someone please invite me either to the premiere of a Christopher Nolan's The Odyssey movie adaptation or, even better, to a scientific cruise around these Greek islands so we can validate everything on outcrop. I promise I will not cry as much as Odysseus. Or at least I will try, I just need as good food as him.

Bibliography:

Homer. (2011). Odissea (J. F. Mira, Trans.; J. Cornudella, Intro.). Barcelona, Spain: Proa.

Bittlestone, R., Diggle, J., & Underhill, J. R. (2005). Odysseus unbound: The search for Homer’s Ithaca. Cambridge, UK: Cambridge University Press.

Hodges, G., Kilcoyne, D., Eddies, R., & Underhill, J. R. (2009, September). Geophysics and the search for Homer’s Ithaca. In Proceedings of SAGEEP 2009 / EAGE Near Surface Meeting, Dublin, Ireland. Retrieved June 9, 2026, from https://www.ags.org.uk/2009/12/geophysics-and-the-search-for-homers-ithaca/

Hunter, K. L. (2013). Evaluating the geological, geomorphic and geophysical evidence for the re-location of Odysseus’ homeland, “Ancient Ithaca” (Doctoral dissertation, The University of Edinburgh). Edinburgh, UK: The University of Edinburgh. http://hdl.handle.net/1842/8002

Underhill, J. R. (2009). Relocating Odysseus’ homeland. Nature Geoscience, 2(7), 455–458. https://doi.org/10.1038/ngeo562

Strabo. (n.d.). Geography (H. L. Jones, Trans.), Book 10, Chapter 2. In LacusCurtius: Into the Roman World. Retrieved June 9, 2026, from https://penelope.uchicago.edu/Thayer/E/Roman/Texts/Strabo/10B*.html

Herod, M. (2014, July 21). The search for Ithaca. GeoSphere – EGU Blogs. https://blogs.egu.eu/network/geosphere/2014/07/21/the-search-for-ithaca/

Odyssey’s end?: The search for ancient Ithaca. (2013, November 16). Smithsonian Magazine. https://www.smithsonianmag.com/history/odysseys-end-the-search-for-ancient-ithaca-112739669/

How archaeologists found the lost city of Troy. (2018, November 13). National Geographic History Magazine. https://www.nationalgeographic.com/history/history-magazine/article/the-lost-city-of-troy

Odysseus Unbound Foundation. (n.d.). Odysseus Unbound: The search for Homer’s Ithaca[Project website]. Retrieved June 9, 2026, from http://www.odysseus-unbound.org

This post has been edited by the editorial board

References:
1. Hamal, K., & Pfahl, S. (2025). Physical processes leading to extreme day-to-day temperature change – Part 1: Present-day climate. Weather Clim. Dynam., 6(3), 879-899. https://doi.org/10.5194/wcd-6-879-2025
2. Hamal, K., & Pfahl, S. (2026). Physical processes leading to extreme day-to-day temperature change – Part 2: Future climate change. Weather Clim. Dynam., 7(2), 1009-1032. https://doi.org/10.5194/wcd-7-1009-2026 
3. IPCC. (2023). Climate Change 2022 – Impacts, Adaptation and Vulnerability: Working Group II Contribution to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press. https://doi.org/10.1017/9781009325844
4. Kotz, M., Wenz, L., Stechemesser, A., Kalkuhl, M., & Levermann, A. (2021). Day-to-day temperature variability reduces economic growth. Nature Climate Change, 11(4), 319-325. https://doi.org/10.1038/s41558-020-00985-5
5. Screen, J. A. (2014). Arctic amplification decreases temperature variance in northern mid- to high-latitudes. Nature Climate Change, 4(7), 577-582. https://doi.org/10.1038/nclimate2268
6. Zou, Z., Li, C., Wu, X., Meng, Z., & Cheng, C. (2024). The effect of day-to-day temperature variability on agricultural productivity. Environmental Research Letters, 19(12), 124046. https://doi.org/10.1088/1748-9326/ad8ede

What will happen to the Greenland ice sheet in the future?

From coarse climate models to high-resolution ice sheet modelling

Global climate models provide projections of how the climate may change in the future, but these models and their output usually have a relatively coarse spatial resolution, typically around 50–100 km. However, ice-sheet modelling needs much higher resolutions of climate information. Specifically, we need the surface temperature and mass balance at surface from the climate model on a resolution of around 1–16 km to obtain reliable estimates of how much ice will melt. This finer resolution is necessary to resolve the small-scale spatial variability that substantially influences the ice sheet, particularly along the margins, where the majority of melt occurs.

One way to generate such high-resolution fields is to take the low-resolution climate model output and feed it into a specialized regional climate model. These models simulate the climate over a smaller domain, such as Greenland, and produce high-resolution output. This output can then be used by ice-sheet models to estimate the future evolution of the ice sheet. However, this approach has several limitations. One problem is that the model chain is often “offline”: the surface mass balance is calculated independently of changes in the ice sheet itself. This matters because the geometry of the ice sheet can strongly influence the surface mass balance. As the ice sheet thins and its elevation decreases, surface temperatures rise locally, leading to increased melting, a process known as the melt-elevation feedback. A second problem is that regional climate models are computationally expensive. They simulate the climate at high spatial resolution and often use sub-daily time steps, which requires substantial computational resources.

How can machine learning help us? 

In our new paper, we propose an alternative approach to tackle the second problem: generative-modelling-based downscaling, which is inspired by AI-based image generation methods that have emerged in recent years. The basic idea behind generative-modelling-based downscaling is to add artificial noise to the coarse climate model output (input for our model) and let the generative model remove the noise (denoise) again while filling in fine spatial details, that it learned beforehand from high-resolution regional climate models (Figure 1).

But why do we add noise at all? The reason is that downscaling is not a one-to-one translation problem. The same coarse climate pattern can correspond to many different fine-scale outcomes. A model that is asked to predict only one fine-resolution map often learns an average of all these possibilities, which can look unrealistically smooth. By adding noise and training the model to remove it again, we give it a way to generate one realistic fine-scale realization among many possible ones. The denoising process guides the model to keep the large-scale climate information from the coarse model while filling in small-scale details that resemble those learned from high-resolution regional climate simulations.

Specifically, we train a so-called consistency model that directly learns a mapping from a noised image to a clean image. In theory, only one evaluation of the model is needed and it is therefore extremely fast. This offers an advantage over other approaches such as diffusion models, which often need several hundred or thousand evaluations to generate realistic output. By deciding how much noise is added to the coarse input, it is possible to control how much pairing or similarity between in- and output there is. In other words, the noise is a knob that controls how much information we want to retain from the coarse fields. When we add minimal noise, the output stays very close to the original coarse climate data, keeping those large patterns intact but missing the small-scale details we care about. Adding a lot of noise gives minimal pairing and the generative model basically does not retain any information from the coarse input. By testing different noise levels, it is possible to determine an 'optimal' noising strength that fills in fine details while still having reasonable pairing with the input fields. This balance is important because we do not want the model to simply create realistic-looking fields. We want it to generate fields that are both realistic and physically consistent with the large-scale climate signal from the original climate model.

Once trained, our method can generate high-resolution climate fields much faster than a regional climate model. This makes it possible to explore more climate scenarios, more model combinations, or larger ensembles of projections and helps us to better estimate uncertainties related to the sea-level rise contribution of the Greenland ice sheet.

The future of ice sheet modelling

More and more data-driven and machine-learning methods are explored in the context of ice sheet modelling but process-based models will likely remain the backbone of the field in the near future. It is also important to note, that most machine learning based methods need some training data to learn from, and these are usually coming from process-based models. Machine learning, therefore, will not make classical climate modelling obsolete but rather is a complementary tool.

Read the paper 

Bochow, N., Hess, P., and Robinson, A.: Physics-constrained generative machine learning-based high-resolution downscaling of Greenland's surface mass balance and surface temperature, The Cryosphere, 20, 1841–1866, https://doi.org/10.5194/tc-20-1841-2026, 2026.

References 

• Fox-Kemper et al., 2021. Ocean, Cryosphere and Sea Level Change. In Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change.
• Morlighem et al., 2017. BedMachine v3: Complete Bed Topography and Ocean Bathymetry Mapping of Greenland From Multibeam Echo Sounding Combined With Mass Conservation.
• Goelzer et al., 2020. The future sea-level contribution of the Greenland ice sheet: a multi-model ensemble study of ISMIP6.
• Feenstra et al., 2025. Role of elevation feedbacks and ice sheet–climate interactions on future Greenland ice sheet melt.
• Hess et al., 2025. Fast, scale-adaptive and uncertainty-aware downscaling of Earth system model fields with generative machine learning.

Edited by Christina Draeger and Leah Sophie Muhle 

How does water contribute to landslides?

How do hydrology, ecology and geomorphology together contribute to triggering a slope failure or preventing it?

How does water chemistry play a part?

Is it possible to predict landslides? And how challenging is it to predict?  

How is weather radar used for land slide early warning systems?

Is rainfall location enough and how accurate are predictions?

What are the biggest gaps in reducing major losses due to landslide disasters?

How will climate change affect landslides?

What are nature based solutions and some challenges in using those solutions?

What are the main objectives of the HELPING Hydrological Decade?

What are your achievements since you've become the chair in 2025?

What is the biggest breakthrough in hydrological research in the last 10 year? And what do you think are the trends for the next 10 years?

Could you share the best and worst piece of career advice that you've ever received?

References:

1. Lahiri, S. P., & Pant, V. (2026). Role of the northern Arabian Sea cold SST bias in delaying monsoon onset and weakening Indian summer monsoon circulation. Environmental Research Letters (2026). https://doi.org/10.1088/1748-9326/ae8460
2. Roxy, M., Ritika, K., Terray, P. et al. Drying of Indian subcontinent by rapid Indian Ocean warming and a weakening land-sea thermal gradient. Nat Commun 6, 7423 (2015). https://doi.org/10.1038/ncomms8423
3. Sharmila, S., Joseph, S., Sahai, A. K., Abhilash, S., & Chattopadhyay, R. (2015). Future projection of Indian summer monsoon variability under climate change scenario: An assessment from CMIP5 climate models. Global and Planetary Change, 124, 62-78.
4. Sulochana Gadgil. 2003. The Indian Monsoon and Its Variability. Annual Review Earth and Planetary Sciences. 31:429-467. https://doi.org/10.1146/annurev.earth.31.100901.141251.