Warm winds and melting ice

We’ve long known that the Antarctic and Greenland Ice Sheets contain more than 99 percent of the freshwater ice on Earth. But recently, a study by researchers at the University of California, Irvine (UCI) and Utrecht University revealed that these two ice sheets have begun trending in different directions: Greenland’s ice surface has been melting faster in recent years while the Antarctic sheet is experiencing a contrasting trend.

Why did you decide to look at the role of downslope winds in ice sheet melt?

The idea was sparked by colleagues at the University of Utrecht who discovered that a lot of melt was occurring during polar nights around their weather stations on an ice shelf in Antarctica. At one station, 23 per cent of the total melt every year was occurring after the sun went down. As a cryospheric climatologist, that surprised me because I wouldn’t expect melt when the sun wasn’t up. These colleagues made the case that wind was causing this melt.

Everyone thinks of surface melt as driven by sunlight—and we know it is the number one contributor to sea-level rise from new freshwater entering the ocean. But I don’t think anyone appreciated how much of sea-level rise might be due to the wind component of melt. So, we set ourselves the task of finding out.

How did you go about answering that question?

We had already looked at all the weather stations on the Antarctic peninsula and the relative role of wind versus sunlight in ice melt. But to answer the question on a more regional scale, we needed to go to areas where observations hadn’t been made yet.

We trained a machine learning algorithm to infer what the melt would be in between these weather stations by considering things like wind speed, sunlight and distance from the coast. The algorithm was able to emulate what the weather stations were measuring. This enabled us to estimate the melt produced in between the observations and look at the entire Antarctic peninsula.

When did you start looking at how these winds were affecting melt in Greenland?

Antarctica has ice shelves. Understanding surface melt is super important here because the more liquid there is, the more tendency there is for ice shelves to hydro-fracture and disappear. Hydro-fracturing is when ice melts so fast on the surface that it creates cracks or fractures in the ice shelf, accelerating its overall thaw. Once we established that wind-driven melt was contributing to the break-up of ice shelves, we wanted to answer the question on a global scale: What’s the relative role of wind versus sunlight heating in causing surface melt generally?

No one had looked at this very carefully in Greenland because there isn’t as much reason to do it there. Greenland lacks significant ice shelves, but it has extensive katabatic winds—another type of downslope wind. These winds gather momentum from the coldest region at the top. They flow downhill and they’re persistent. So, we knew the winds would have an impact on overall melt in Greenland—we just didn’t know how much.

We also knew Greenland was generally much warmer than Antarctica. What we did not yet realize was that Greenland doesn’t need wind as much as Antarctica does to tip the energy balance from frozen to melting.

When you compared Greenland to what you had been finding in Antarctica, what was the difference?

There were two big differences: Antarctica has a long-term trend toward decreasing surface melt, which surprised us. We think that’s partly due to the recovery of the ozone: the more ozone there is, the more sunlight gets absorbed in the upper atmosphere rather than hitting the ice surface and contributing to melt. We were surprised because with global warming, we assume everything is heating up at the same rate. But in fact, it’s regionally very different. Greenland experienced a transition from a very gradual increase in surface melt from the sixties to the nineties to a doubling of the rate starting around 1990, which is when we think the ice sheet really started to be influenced by global warming.

The second difference between the two is that in Antarctica, the wind-contributed portion of the melt is not changing very much—whereas in Greenland, wind as a contributor to surface melt is getting less and less important as the ice sheet reaches the thawing point. In Greenland, you don’t need wind to observe melting because the sun warms the ablation zone up to zero degrees all on its own.

What can we learn about climate change and the cryosphere from these findings?

Antarctica is, in many ways, the last refuge from climate change. It’s so isolated and cold that the warming it is experiencing overall is still a ways off from triggering the rapid melt we are seeing in Greenland. Greenland is giving us a glimpse of what the future of the Antarctic ice sheet might look like.

The other thing is that, although the overall melt rate for Antarctica is decreasing, during certain seasons in the warmest parts of Antarctica, there are ice shelves at risk of being fractured by the weight of melted water. When those ice shelves disappear, they will release the buttressing force that was holding back the ice streams on the continent itself. When the ice shelf disappears, that land-based glacier is going to speed up and go straight into the ocean because there’s no ice shelf there anymore.

So, understanding the processes that lead to the surface melt can help us make the first predictions of ice shelf break-up in Antarctica. In other words, which ice shelf will be next, and when can we expect it to disintegrate due to hydro-fracture? These results will also improve our estimates of sea-level rise because we’ll better understand the balance of wind- and solar-driven melt.