Mysterious lakes deep in the Antarctic snow

After almost three weeks in the field, we returned to the Neumayer research station on December 25. We had enjoyed excellent conditions out there, with good weather and efficient working days, but it was still a pleasure to come back to the welcoming German research station. A hot shower, fresh vegetables and fruit, real beds – and, not least, the feeling that we had actually accomplished everything we had planned.

However, the recovery was short-lived. Before long, curiosity took over: what had we actually brought home in the form of data? Out in the field, we had limited access to computers and specialised software, so the analyses had been cursory. Now we finally had the opportunity to take a closer look at our data.

We were particularly curious about the radar data. From previous studies, we know that shallow lakes can form on the ice-shelf surface during the Antarctic summer. They occur where the sun’s rays are strongest, often in darker areas where the wind has blown away the snow, leaving a more heat-absorbing surface. When the Antarctic autumn arrives, these lakes usually freeze again.

But something unexpected appeared in our radar data. Using the radar, which can image structures in snow and ice down to several tens of metres, we saw clear signals that resembled liquid water. To be sure, we drilled at the site and confirmed that there was indeed a water-filled cavity below the surface, about 30 metres down.

This was surprising. Previously, we had measured temperatures of around −15 °C in the snow – conditions in which one would not normally expect liquid water. However, the explanation may lie in what happens when meltwater penetrates the cold firn. When water freezes, large amounts of energy are released as latent heat. Over several seasons, this heat input can gradually warm the firn to 0 °C, allowing liquid water to remain at depth.

Back at the station, we began comparing our observations with satellite data. It turned out that something was happening not only below the surface. Satellite measurements of snow-surface height showed that the ice surface in this area had been rising and falling over time. In the vicinity of the Riiser-Larsen Ice Shelf, we also identified structures indicating that underground lakes can sometimes drain very quickly, which in turn can cause the ice above to collapse.

This suggests that our current picture of the Riiser-Larsen Ice Shelf may be incomplete. Surface lakes may not always freeze completely, leaving water at depth that persists for several seasons. Similar phenomena have been observed in other parts of Antarctica and, in some cases, have been linked to the weakening or even collapse of ice shelves. However, for Riiser-Larsen, we see no signs that this would be the case in the near future.

At the same time, our observations raise several important questions. Will warmer summers lead to more such structures? What does this mean for the strength and long-term stability of the ice shelf? And how widespread is the phenomenon, viewed in a broader perspective across Dronning Maud Land and Antarctica as a whole?

When we return home to Sweden, a lot of analysis, many discussions, and hopefully new insights into how these icy environments work – and how they may change in the future – await us.

Thomas Frank (PhD student, Department of Earth Sciences, Uppsala University) och Ola Fredin (NTNU)

The text is a republication of a blog post on the Swedish Polar Research Secretariat's website that was published on 1 January 2026.