News detail view

Prologue: The Arctic, Fram Strait – a fateful recon flight

The weather on Tuesday, 14 June 2021 offered ideal conditions for the Polarstern’s onboard helicopter. The sun was shining on the pack ice of the western Fram Strait, with only the occasional cirrostratus cloud, as opposed to the foggy skies often encountered there. “Given the weather conditions, I was in high spirits. I assumed the helicopter crew would shortly return from their recon flight and tell us they’d found a large floe that we could later use to gather snow, ice and water samples for my project on the spread of microplastic in the Arctic,” recalls Dr Melanie Bergmann, a polar researcher at the Alfred Wegener Institute, Helmholtz Centre for Polar and Marine Research (AWI) and part of the Arctic expedition on board the German icebreaker Polarstern in June 2021.

At the time, the ship was cruising at 78.6 degrees north latitude off the east coast of Greenland, just a few nautical miles from Station Ostgrönland 4, part of the AWI deep-sea observatory “Hausgarten”. In this network, which consists of 21 long-term monitoring stations, Melanie Bergmann and the AWI Deep-Sea Group have been observing life in the Arctic depths for more than two decades. At selected stations, they deploy e.g. the deep-sea camera system OFOS, which can capture photos and video of everything it sees at the ocean’s floor, 2,500 metres beneath the surface. For this particular expedition, however, Melanie Bergmann had put the topic of deep-sea fauna on the back burner; her focus was on microplastic in the Arctic.

What she couldn’t know back then: with just one sentence, the helicopter crew was about to change all her plans and set a new research project in motion, one that would occupy Melanie Bergmann and all other participants for the next five years.

1^st act, Polarstern expedition: A strange dark spot amidst the ice

“We found an island in the pack ice,” said one of the helicopter pilots jokingly once back on board. This immediately caught Melanie Bergmann’s attention; she asked what they meant by that. Really an island? Well, maybe not a real island, but a large patch of ice covered with a layer of black rock. “I knew right away that we had to take a closer look,” the marine biologist explains.

Rock and stones are of critical importance for life in the deep sea. Especially on the seafloor, sessile organisms like sponges, soft corals and anemones need them as a stable basis; otherwise they can’t settle there. This also means that, when stones sink down to the otherwise soft and mostly silty floor of Fram Strait, it can have a significant impact on biodiversity.

“After looking through the footage from previous years, we had found the first indications that the number and density of stones on the floor of Fram Strait were on the rise. But since we had no explanation for the phenomenon, we didn’t initially investigate further,” she recalls. But large sheets of ice filled to the brim with rock might explain how the stones found their way to the deep sea.

2^nd act, Polarstern expedition: An iceberg full of stones

Eight eyes can see more than two. As such, Melanie Bergmann didn’t fly to the stony island on her own: three other expedition members, including her American colleague Dr Kirstin Meyer-Kaiser, joined her in the helicopter. The biologist from Woods Hole Oceanographic Institution had previously investigated the biodiversity on stones in the Arctic deep sea for her dissertation. Further, she was one of a handful of experts worldwide on these “dropstones” – stones released by melting icebergs that then sink to the seafloor.

“When we landed on the ice, it felt like we were on another planet. Wherever we looked, we saw stone-covered hillocks. So we couldn’t tell for sure whether we were standing on a large floe or an iceberg. The expanse was much larger than a football field. But we couldn’t gauge its size more precisely, since we could hardly tell its edges from the white pack ice,” Kirstin Meyer-Kaiser relates.

While the co-pilot took position at the highest point on the ice to keep watch for polar bears, Kirstin Meyer-Kaiser began collecting rock samples, taking readings, and photographing individual stones. Some were black, others more dun or russet. Their sizes varied as well: some were hardly larger than pebbles, while others fit the category “boulder”. They were composed of slate and quartz, as subsequent analyses would reveal.

Melanie Bergmann took samples of the snow, seawater and ice algae at the edges of the ice, while sea-ice physicist Mario Hoppmann collected ice cores and photographed the area from virtually every angle. He also suggested trying to capture the entire field of ice and stone from the air. “We made small piles of stones at five-metre intervals so that, when we looked at the aerial photography later and identified the piles, we’d have an indication of scale that would allow us to calculate the size of the stone-covered areas,” Kirstin Meyer-Kaiser explains. The plan worked, yielding a calculated area of roughly 1,200 square metres.

3^rd act, video analyses: Stones that rain down from icebergs

Once back on board the Polarstern, the American immediately got to work analysing the deep-sea photos from the 2021 Hausgarten expedition, taken just days earlier. According to Kirstin Meyer-Kaiser: “I was so excited that I immediately measured the size and density of the stones in the footage.” The images showed a host of stones that appeared to have just recently sunk to the seafloor. Their surfaces weren’t covered by a thin layer of deposited microparticles, nor had any worms, sponges of anemones settled on them.

Unfortunately, the images from 2021 couldn’t be compared with those from previous years. As she explains: “The camera system is lowered into the water alongside the Polarstern’s hull and towed in that position for roughly 2,500 metres. Since the ice conditions vary from year to year, we can’t repeat exactly the same route, which makes direct comparisons of the data gathered more challenging.”

Nevertheless, several weeks later her luck changed: she found that the OFOS images from the years 2015 and 2017 had been taken along precisely the same route over the seafloor. “We could clearly see that many more stones reached the seafloor in 2017 than just two years earlier. That means, in the span of two years the ice transported substantially more stones from land to Fram Strait, which verifiably changed the living conditions in the deep sea.”

But neither she nor Melanie Bergmann had an answer to the most important question: where had the ice and stones come from?

4^th act, ice observations on the Polarstern: A slumbering treasure trove of data

Luckily, Melanie Bergmann knew someone who might be able to provide the answer. Back on the research icebreaker, she promptly emailed her AWI colleague Dr Thomas Krumpen in Bremerhaven. The sea-ice physicist had developed a method for using satellite data to track pack-ice drift through the Arctic Ocean. Her question: could he use the same method on the “island” of ice and rock?

“We could soon see from the photos that the supposed island was actually an iceberg. Deonie Castle, our colleague from New Zealand, was able to confirm that thesis by measuring the salinity of multiple ice samples. The iceberg’s salinity was much lower than that of the surrounding sea ice. So the presumed island was definitely an Arctic iceberg,” Thomas Krumpen explains.

Unfortunately, this finding made drift tracking more complicated. “We can only use satellite data to retrace the movements of pack ice. But the icebergs in Fram Strait only drift like the pack ice when they become trapped in and freeze to thick sea ice. But when the icebergs move separately from the sea ice, it’s impossible to retrace their route, since we can’t distinguish them from the sea ice in the satellite images,” says Thomas Krumpen.

Disappointed by the hardly satisfying outcome of his first attempt, the 47-year-old started searching for data on iceberg sightings in Fram Strait. He found what he was after in the AWI data archive; after all, iceberg sightings are one of the standard components of weather observations on board the research icebreaker Polarstern. “Ever since 1982, when the icebreaker was commissioned, each expedition’s weather observers have recorded whether or not they sighted any icebergs during their regular sweeps. And that meant there also had to be this type of data on Fram Strait, reaching back 40 years.”

His hunch paid off: he found a complete and digitalised record of the weather observations, starting with the first sweep at 6 a.m. on 28 December 1982. But all the observations were in code. To make sense of them, he would need the code tables used by the World Meteorological Organization (WMO). And there was exactly one person at the Alfred Wegener Institute with the tables in their desk drawer: Dr Holger Schmithüsen, scientific coordinator of meteorological observations on the research icebreaker Polarstern and at the Antarctic research station Neumayer III.

“Encoding the observations makes sense, as the WMO has to ensure that they are carried out and recorded the same way all over the globe. For trained experts, the codes are sensible and straightforward,” the meteorologist explains.

The technical guide to decoding the weather observations is over a thousand pages long. Table 0439 lists how iceberg sightings are to be recorded. Each sighting is assigned a number ranging from 0 to 9, where 0 means no glaciers were seen, while 9 means the observers, after walking the ship’s perimeter, counted more than 20 icebergs of varying sizes.

“We’ve been gathering data on iceberg sightings during Polarstern expeditions for more than 40 years now. Thomas Krumpen was the first researcher to ever request access to them. That’s why it took a bit of time for me to decode the datasets from nearly 100 expeditions and get them back to him,” says the AWI’s chief meteorologist.

A short time later, Thomas Krumpen and Kirstin Meyer-Kaiser could hardly believe their eyes: “The Polarstern iceberg data proved to be a treasure trove in two different ways. Firstly, it showed that, from the year 2000, the number of iceberg sightings in Fram Strait soared – from an average of 20 sightings per year to more than 80 in some cases. That told us there was a substantial increase in the number of icebergs in Fram Strait,” says the AWI sea-ice physicist.

Secondly, the team identified 106 icebergs that were trapped in the pack ice, and not just when they were sighted but also for weeks before. “We were then able to individually trace the drift of these icebergs, or more precisely, the drift of each ensemble of icebergs and pack ice.”

The drift calculations indicated that the icebergs chiefly stemmed from two different regions. “Icebergs sighted west of the 5^th meridian west mostly come from northeast Greenland, where the massive 79 N Glacier and its neighbour, the glacier Zachariae Isstrøm, empty into the sea. We’ve been able to link icebergs drifting east of the 5^th meridian west to glaciers on the Russian Arctic islands of Franz Josef Land and Severnaya Zemlya, which are between 1,000 and 2,000 kilometres from the Hausgarten observatory in Fram Strait,” explains Kirstin Meyer-Kaiser.

Determining the icebergs’ areas of origin was a major breakthrough. But could these glaciers really be the source of the icebergs? And had their calving fronts really been producing more icebergs since 2000?

5^th act, facts from glaciology: Greenland’s glaciers are sliding faster and faster into the sea

To answer these follow-up questions, Thomas Krumpen and Kirstin Meyer-Kaiser were going to need some help. They contacted Dr Abbas Khan from the Technical University of Denmark, a widely respected expert on the glaciers of northeast Greenland. The 53-year-old geodesist and his team had been monitoring the flow speed and ice loss for the 79 N Glacier and Zachariae Isstrøm for roughly 20 years, using satellite data and a network of GPS sensors to do so. In addition, the expert on changes in Earth’s surface could draw on observational data from the 1980s and 1990s, which he had retrieved from the Technical University’s basement archives and digitalised.

“The rising flow speeds and ice loss in the Zachariae glacier since the year 2003 precisely match the growing number of icebergs and stones that we can see in the Polarstern data and in deep-sea video footage alike,” says Abbas Khan. Further, in his view another observation indicated that the stony iceberg mostly came from the Zachariae glacier.

“We know two types of iceberg from Greenland. Glaciers with narrow tongues between five and ten kilometres wide tend to produce icebergs that are up to one kilometre long. When they calve, they topple into the ocean headfirst. In other words, they twist in the process and often hit the seafloor with sufficient force to produce earthquakes that can be detected as far away as Australia. With that much force, the icebergs lose any material they might have had on their surface,” Abbas Khan explains.

The stony iceberg was spared this trauma. As such, it had to be the second type of iceberg – a so-called tabular iceberg – as are especially seen in the Antarctic. “When it comes to glaciers with broad tongues like Zachariae Isstrøm and the 79 N Glacier, what we see time and again is that such enormous blocks of ice calve from them that the blocks can rightly be called tabular icebergs. These icebergs are several kilometres long and often up to 900 metres thick. When they break off the glacier tongue, it’s hardly spectacular. They simply drift away – like a ship pulling away from a harbour. Any rock or stones on the surface simply stay where they are,” says Abbas Khan before adding: “The stone on the ice comes from landslides. The glaciers move through deep valleys. As they do, they frequently trigger landslides on the mountainsides left and right of the tongue, leaving behind material on the surface of the glacier.”

The glaciers on Russia’s Arctic islands have also been flowing out to sea faster since the year 2000, as confirmed by various studies. The thickness of their ice tongues has also declined. But current data on their flow speeds, glacier lengths and ice-mass loss is practically non-existent. Thomas Krumpen’s bewildered conclusion, following an extensive literature review: “As such, it’s difficult to trace back the rising iceberg numbers in the eastern Fram Strait to a specific glacier.”

But the team wasn’t ready to throw in the towel, as there was still one option they hadn’t tried yet. What if they could use the Alfred Wegener Institute’s sea ice/ocean computer model to show that the icebergs that calved in e.g. Franz Josef Land actually drifted to Fram Strait and melted there?

6^th act, AWI Bremerhaven: Teaching a computer model how to move icebergs

The right woman to answer these questions: Dr Claudia Wekerle, a 43-year-old ocean modeller from the Physical Oceanography section of the Alfred Wegener Institute in Bremerhaven, who knew the AWI sea ice/ocean computer model FESOM2 inside and out. She promptly agreed to investigate the iceberg question, even if it meant doing the impossible.

“Our computer model consisted of a global ocean model and a sea-ice model. As such, it could simulate sea-ice flows and ocean currents – for the Arctic, even at a higher resolution of 4.5-kilometre grid cells. What it hadn’t been able to do so far was to simulate the calving, movements and melting of individual icebergs in the Arctic so that we could retrace their drift routes across the Arctic Ocean,” the ocean modeller explains.

To deliver good results, computer models need the best possible initial data. In this case, that meant precise information on how many icebergs the glaciers of northeast Greenland and in the Russian Arctic had released into the Arctic Ocean over the past 50 years and their sizes – and not just for a given year, but for every year. “We have detailed and extensive iceberg analyses for Greenland. For some glaciers, we know the exact number and size distribution of icebergs. But we don’t have this annually compiled data for the Russian islands. After a lengthy search, the best we could find was an approximate mean value that we could use as an annual constant. Accordingly, we ultimately used constant calving rates for the glaciers of northeast Greenland and Russia’s Arctic islands alike,” she says.

In the next step, the ice constants had to be allocated to individual icebergs. “In our modelled world, icebergs are represented as simple cubes of ice, some larger, some smaller. For each year, we set more than 100 of them in motion – in both northeast Greenland and the Russian Arctic. And for each of the more than 5,000 icebergs, we had to tell the model where it calved and how large it was,” Claudia Wekerle explains.

Since entering the data manually would have taken months, Claudia Wekerle turned to her colleague Dr Lars Ackermann for help. The 36-year-old AWI climate modeller knew how to write small formula scripts that automatically generated the icebergs’ calving data for the model. As a paleoclimate expert, Lars Ackermann normally worked on model-based simulations covering thousands to tens of thousands of years. Without scripts for the initial data, his work would have been impossible.

Once all the iceberg data had been fed into the model, the researchers hit the start button. It took the AWI supercomputer ALBEDO ten days to calculate all the ocean, sea-ice and iceberg data for 50 years. The result: a multi-Terabyte mass of data, from which Claudia Wekerle and Lars Ackermann had to extract the calculated positioning data for more than 5,000 icebergs. “At first, there were small errors in the model. For instance, our modelled icebergs trapped in very thick pack ice didn’t drift along with the other ice. Lars then revised the model’s code, which got rid of these problems,” says Claudia Wekerle.

After several test runs, how the sea ice and icebergs behaved in the model matched fundamental principles regarding the natural interplay of the ocean and ice. Now the individual icebergs’ positioning data could be extracted and their drift routes reconstructed. “Much to our amazement, we could also see in the model that there were more icebergs drifting through Fram Strait starting in the year 2000. As such, our model-based data matched the real-world observations very well,” enthuses Claudia Wekerle.

“We then asked ourselves why we were seeing more icebergs in Fram Strait and conducted a closer analysis of the ocean and sea-ice data in the model. Our modelling results not only showed that the projected iceberg routes were correct. They also confirmed that the sea ice made a major contribution to iceberg density. The increasingly thin and young sea ice is drifting faster, accelerating the movement of those icebergs trapped in it at the same time – a trend that Thomas Krumpen and his colleagues have also observed in the field,” adds Lars Ackermann.

The grand finale: A seamless chain of evidence

In modern climate research, it’s a rarity for modelling and observational data to match so precisely. The two climate modellers promptly informed the rest of the team. Armed with the iceberg routes from the ocean model, they could finally combine the wealth of individual facts into a sound chain of evidence. By pooling their resources, the researchers had not only solved the riddle of the stony iceberg in Fram Strait, but also uncovered a previously unknown impact of climate change in the Arctic.

“We can demonstrate that, due to global warming, the glaciers of northeast Greenland and Russia’s Arctic islands are now flowing faster and producing more icebergs, including many that transport stones. Due to accelerated sea-ice drift, icebergs from the Russian Arctic in particular no longer remain as long in the Central Arctic. Instead, they’re drifting faster and in greater numbers toward Fram Strait, where they gradually melt and release any stones they contain into the ocean,” says Thomas Krumpen. “The stones sink to the ocean floor where, in the span of several decades, they are settled on by deep-sea organisms like worms, sponges, anemones and soft corals, species that have previously found little suitable habitat in Fram Strait. In other words, the biodiversity of the deep sea is changing,” Kirstin Meyer-Kaiser adds.

Despite being thrilled about their seamless chain of evidence, the researchers are also troubled by their findings. “What we’re seeing is a climatic domino effect, from the atmosphere to the glaciers and sea ice, to the Arctic deep sea. In other words, anthropogenic climate change is producing far-reaching impacts, including lasting changes to interactions between the atmosphere, ice and ocean,” says Melanie Bergmann.

Nevertheless, all participating experts are proud of their achievement. “The new study is one of the best I’ve ever been part of,” says Abbas Khan. “It shows the kind of insights and breakthroughs that are possible when experts from glaciology, biology, sea-ice physics, meteorological observation and climate modelling collaborate, with each contributing their own know-how. Personally, I not only learned so much, but also now have a much better understanding of how my glacier research could help peers from other disciplines and how I can profit from their findings. A clear win-win situation.”

The results of the five-year research project have now been released as a feature article in the scientific journal Nature:

Thomas Krumpen, Kirstin S. Meyer-Kaiser, Claudia Wekerle, Lars Ackermann, Deonie Castle, Melanie Bergmann, Mario Hoppmann, Shfaqat A. Khan, Autun Purser, Holger Schmithüsen, Amplified Arctic iceberg traffic reshpares benthic biodiversity, Nature (2026). DOI: 10.1038/s41586-026-10630-4.

Contact

Dr. Thomas Krumpen (AWI)

Dr. Klaus Grosfeld (AWI)

Dr. Renate Treffeisen (AWI)

Dr. Kirstin Meyer-Kaiser (WHOI)

Prof. Abbas Khan (DTU)

Dr. Melanie Bergmann (AWI)

Dr. Autun Purser (AWI)

Dr. Mario Hoppmann (AWI)

Dr. Claudia Wekerle (AWI)

Dr. Lars Ackermann (AWI)

Dr. Holger Schmithüsen (AWI)

Deonie Castle

Questions?

Write us an 

email or use the 

contact form.

Author

Sina Löschke (Science Writer)

www.schneehohl.net