And changing they are, for the extent of the Arctic sea ice in summer has declined by 30% in the past 30 years, and that loss is accelerating.
The Arctic is also a driver of climate change, though, because the whiteness of ice means it reflects sunlight back into space, thus cooling Earth,
whereas the darkness of open water means it absorbs that light.
The less of the former that is happening, and the more of the latter, the faster global temperatures will rise.
Start, then, with the ice. At the moment this is monitored mainly by satellite. Measuring the extent of the Arctic's ice from space is easy.
Measuring its thickness is trickier. From orbit, this is done by a mixture of radar and laser beam.
Icesat 2, an American craft, provides laser-altimeter data that record the height above sea level of the top of the snow that overlies the ice.
Cryosat 2, a European one, uses radar to penetrate the snow and measure the height of the top of the ice itself.
The thickness of the ice in a particular place can then be calculated by applying Archimedes' principle of floating bodies to the mixture of ice and snow,
and subtracting the thickness of the snow. However, Julienne Stroeve of University College London, now safely returned from her leg of the mission,
believes that the data collected by these two satellites may be inaccurate, leading to an overestimation of the ice's thickness.
When all is working perfectly, the return signal for Cryosat 2 comes exactly from the boundary between the ice and any overlying snow.
Dr Stroeve thinks, though, that this is not always what happens.
Variables such as layering within the snow, along with its temperature and salinity,
might affect the returning radar signal by changing the snow's structure and density.
This could cause the signal to be reflected from inside the snow layer, rather than from the boundary where it meets the ice.
If that were happening, it would create the illusion that the ice beneath the snow is thicker than is actually the case.
To investigate this possibility Dr Stroeve took a purpose-built radar on board Polarstern.
Each week, she and a colleague mounted this 170kg instrument on a sled and dragged it to a new site, to sample different snow conditions.
As they towed it, they sent radar pulses on the frequency bands used by the satellites downwards into the snow and measured the amount of backscatter.
The deflection of the signals in this backscatter gives a picture of how particular snow conditions might be changing the way the satellite's radar is returned.
Dr Stroeve's radar died on January 31st— one of many of the expedition's machines that fell victim to the Arctic winter.
But by the time that had happened she had managed to gather a fairly good set of data.
Her conclusion is that the reflection does indeed sometimes come from the interface between snow and ice, as it is supposed to. But not always.
The discrepancy is important. Her measurements already show that the ice is "definitely thinner than the satellites suggested".
She has yet to analyse the data fully, but preliminary investigation indicates that both snow depth and temperature influence backscatter.
It therefore looks likely that the amount of Arctic sea ice around has been overestimated.