Recreating Iceberg Flips in a Lab
Icebergs can sometimes lose their stability and suddenly flip over. The exact cause of these capsize events has been unclear. A new experiment recreating iceberg flips in the lab suggests that changes in shape due to melting determine if and how an iceberg tips over [1]. The results could help improve climate models by providing a better picture of ice dynamics in the oceans.
Iceberg flips are rare but dramatic occurrences, which can sometimes generate destructive waves that impact ships and coastal communities. Scientists have long suspected that uneven melting on the top and bottom of an iceberg leads to an unstable shape and an eventual flip, but some studies have suggested that melting could instead be a stabilizing force. The problem is complicated by the fact that melting changes iceberg geometry, which, in turn, impacts the melting rate. Measuring shape changes in detail on a real iceberg is challenging. Moreover, waves, currents, and wind can destabilize icebergs, and it is hard to separate out all the different factors, says Leif Ristroph from New York University. So he and his colleagues set up an experiment to investigate how the shape of ice evolves as it melts.
The researchers created cylindrical ice slabs that were 8 cm wide, 18 cm long, and free of any bubbles or contaminants, which ensured smooth melting. These slabs were placed in a water tank the size of a large suitcase, where they could freely rotate and where their cross sections could be filmed as they melted. “Considerations in our experiment were more about control, reproducibility, and measurability than realism,” Ristroph says.
The team observed that an ice slab would stay in a position for a while before it abruptly rolled over into another position. During each stationary phase, the ice melted faster from the submerged bottom portion than from the top portion peeking out of the water. This differential melting caused two pinched corners to develop on the sides of the slab at the waterline. Gradually, the regions around the corners became unstable, causing the ice to flip. On each transition, one of the two corners returned to the waterline. Over time, multiple corners developed along the perimeter of the slab, molding its cross section into a pentagon-like shape that both carried a record of past flips and primed the ice for future flips.
The lab icebergs flipped 10 to 15 times before they completely melted away in a half hour. The time between the flips was highly variable and sensitive to the precise geometry of the icebergs. “Extremely subtle differences in shape determine if a piece of ice, whether it’s our miniature icebergs or real icebergs, capsizes or not,” Ristroph says.
The researchers modeled the ice-melting dynamics with numerical simulations. Despite simplifications, such as assuming a constant melting rate across the submerged part of the iceberg, the simulations yielded results that were qualitatively similar to the experiments. Specifically, the simulated icebergs flipped at roughly the same rate as the laboratory icebergs while also developing pentagon-like features. No external forces, as would be akin to other climatic factors, were applied to the icebergs in either experiments or simulations, suggesting that melting alone can drive iceberg flipping.
In future work, the researchers plan to investigate other factors that might impact the link between melting and iceberg shape. Temperature could be critical, for example. The team used room-temperature water in their tank, but most of the melting of ice in the ocean occurs in much colder waters, Ristroph explains.
The new work could help improve global climate models used in predicting weather and in tracking climate-change patterns. These models rely on smaller-scale submodels of various climate inputs, such as sea-ice cover. Ice reflects sunlight more than water, so it’s important for climate models to accurately account for ice melting and for atmospheric changes over ice, Ristroph says.
However, the shape of the lab icebergs could limit the broad utility of the experiments, according to Michaela Fendrock, a geologist at Alfred University in New York. “It could be hard to apply [the insights from this experiment] to real-life icebergs, which tend to be long, flat, and rectangular in shape,” she says. But Alexis Kaminski, an expert in geophysical fluid dynamics at the University of California, Berkeley, thinks the new work could be generalizable. The simulations showed that the melting behavior could be reproduced with simplifying assumptions about melting along the ice surface, says Kaminski. “I think [this parameter insensitivity] says their theory will work well for a real iceberg.”
–Sachin Rawat
Sachin Rawat is a freelance science writer based in Bangalore, India.
References
- B. Johnson et al., “Shape evolution and capsize dynamics of melting ice,” Phys. Rev. Fluids 10, 093801 (2025).