Canada’s Northwest Passage is widely known as one of the last great frontiers. 120 years since Amundsen finally cracked its icy narrows, it remains a challenge even today.

Amy Swiggs, a PhD student at Northumbria University and the UK Centre for Polar Observation and Modelling, took on the challenge by measuring the sea ice of this complex Arctic region using ESA’s CryoSat and NASA’s Landsat 8.

The precarious Northwest passage

Many books and documentaries have been produced that detail the terror of getting stuck in the Northwest passage, where many ships met a frostbitten demise in search of a way through.

Whilst shipping routes may benefit from better access through the Northwest passage, as many a European explorer from Cabot to Amundsen set out to discover, its shape and location mean it’s still hard to measure sea ice, never mind predict it.

Even though ice is melting rapidly in the Arctic, and the opening of shipping routes is an increasingly likely opportunity, it is not all plain sailing. The Canadian Arctic is still a tricky place to navigate and is not continually navigable even to the sternest ships.

It is a region that is also missing from many Arctic sea ice models.

“There’s a lot of shipping interest in the Canadian Arctic Archipelago, but there isn’t a consensus as to what’s happening there in terms of sea ice,” explains Amy.

“You would think that because the Arctic is warming so quickly, the Northwest passage would become more navigable. But what’s been found is that, as the Arctic is warming, it’s pushing thick and complex ice through the Northwest passage, and that ice is potentially more hazardous.”

It is also a hard region to measure with satellite altimetry.

The power and potential of CryoSat

ESA’s ice mission, CryoSat, has the longest ever satellite altimeter record of global ice in existence.

In over 14 years it has markedly improved our understanding of not just polar sea ice and the ice sheets covering Greenland and Antarctica, but also sea level, mountain glaciers, and rivers, lakes and reservoirs worldwide.

“CryoSat has an amazing record, which we use to produce sea ice thickness estimates,” says Amy.
“But there's data that we can get even prior to that along the processing chain. I wondered if we could use some of that data to have a look at sea ice a bit more closely, to map fine scale ice dynamics in the Canadian Arctic.”

However, when measuring sea ice at fine scale CryoSat does have some challenges.
It works by sending radar microwaves towards Earth’s surface and measuring the differences in the signal that echoes back off different surfaces, such as ice or water.

Sea ice can either be tethered to the land (fast ice) or floating in ice floes. Between the ice floes there are leads, which are open stretches of water. CryoSat essentially measures the elevation of sea ice by using these stretches of open water as a reference.

However, due to the way it detects radar echoes, and the way the data is processed to remove noisy signals, it can sometimes overestimate the presence of leads and miss some of the floes.

It can also be difficult to pick up very rough sea ice surfaces, which can sometimes look (to a satellite’s processing algorithms) like a choppy sea.

A more accurate way of identifying sea ice leads and floes from space is to use a satellite like NASA’s Landsat 8, which takes optical and thermal images. The sea is a lot warmer than the ice, so you can tell them apart quite easily. Landsat 8 also has a resolution of between 30 to 100 metres compared to 300 metres for CryoSat.

However, Landsat 8 images are harder to use in the long, dark Canadian winter nights when they are often completely black. For that reason, it is difficult to get year-round images of sea ice in the Northwest Passage.

A particularly useful feature of CryoSat is that it can measure all year round, including during the dark winter months. Its radar can also penetrate through clouds, which is an obstacle for optical satellites.

Improving CryoSat measurements

To get a more consistent look at sea ice in the Northwest Passage, Amy therefore compared CryoSat data with Landsat 8 images and used the latter to tweak CryoSat’s data to better fit the reality.

Initially, she calculated percentage of leads and floes for all the areas that overlap between the two satellites, finding that CryoSat detected lead density 14% higher than Landsat 8 and floe density 45% lower.

To improve the CryoSat dataset, she worked to bring it more in line with the Landsat 8 data.

“We made the assumption that the Landsat 8 images were correct,” explains Amy. “Then, for every month we had data, we created an average and modelled the CryoSat data to be more in line with Landsat 8.”

This bias adjustment brought the root mean square difference between the satellites’ measurements down from 20% to 5% for leads and 47% to 6% for floes. A substantial improvement.

“It’s wonderful to see that in 2024, 14 years after CryoSat was launched, we are still finding new and important uses for the measurements it has collected” says CPOM Director Andrew Shepherd of Northumbria University.

“Amy has done a wonderful job showing how it is possible to detect the abundance of sea ice floes, and this paves the way for applying the technique across the entire Arctic region.”

The next step will be to see whether this modelling approach works not just in the rest of the Canadian Arctic, but the Arctic as a whole.

“This is yet another exciting application of CryoSat research that shows ESA’s well-loved ice mission has much to offer as it approaches its fifteenth anniversary in space,” says CryoSat Mission Manager, Tommaso Parrinello.

“Along with our ongoing Cryo2ice collaboration with NASA’s ICESat-2 satellite, our ability to map ice from space continues in leaps and bounds. I can’t wait to see what else we can achieve before we welcome the future CRISTAL mission into orbit to take on the baton from CryoSat.”

The study, “Detecting sea ice leads and floes in the Northwest Passage using CryoSat-2” was published in the IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing.