It’s cold near the poles where the sea ice forms, which makes it difficult and dangerous to monitor the rapid environmental changes that are happening. The ice is also dark to look at during the winter polar night, so satellite-mounted cameras are of limited use. It’s also often cloudy which can block satellite-mounted lasers that actively illuminate the surface. In response to these challenges, satellite-mounted, cloud-penetrating radar systems are a popular way of monitoring the sea ice cover.
CryoSat-2 monitors Arctic sea ice thickness in the polar winter, even through clouds.
In a similar way to how the components of white light bounce differently off objects to give them color, different radar frequencies bounce off objects differently too. Sea ice is typically monitored using satellite-radars ranging in wavelength from a couple of centimeters (very similar to the energy in a microwave oven) to a few mm (the same frequency used by 5G internet and Iridium phones). Two particular radar frequencies (sometimes called bands¹) used in recent satellite missions are termed ‘Ku and Ka’ bands. The ‘K’ stands for Kurz (German for short), and the ‘a’ and ‘u’ suffixes stand for ‘above’ and ‘under’: Ka is around twice the frequency of Ku.
The same sea ice viewed at different radar frequencies (from Howell et al., 2018). Different images are obtained from viewing the same scene in different frequency bands. Some sea ice movement has occurred between the two satellite overpasses.
In winter, sea ice is covered in snow, which presents itself to radar differently to Ka and Ku-band radar waves. How differently? It’s tough to say. We can’t easily work it out from satellites because they have different orbital configurations, operating modes and antenna patterns. That is to say, we never look at *exactly* the same patch of sea ice in *exactly* the same way in both bands. It’s only by doing this that we can isolate the role of the radar frequency.
The KuKa radar instrument was capable of viewing the snow at a variety of angles (scatterometer mode). It was also capable of ‘staring’ down at the snow while being towed along by a snowmobile. Photographs from Stefan Hendricks, displayed in Stroeve et al. (2020; The Cryosphere).
That’s why a dual-frequency radar machine was deployed on the sea ice during the MOSAiC expedition last year by Julienne Stroeve (my PhD supervisor) and Vishnu Nandan. Details of the instrument and the deployment were published in The Cryosphere this week, along with some early results. Because the machine looks at the same snow and ice at both frequencies with (almost²) the same settings, any differences in what the machine sees can be attributed to the different frequencies.
The MOSAiC Expedition also provided the ideal framework for continuous monitoring of the snow and sea ice at both frequencies. A vast array of other physical data was taken alongside the radar measurements like temperature profiles, snow stratigraphy and microstructure characterisation, and laser scans of the snow surface. This means that changes in how the snow looks to the radar can be related back to its changing physical properties. As well as contextual physical data, information from several other ‘remote sensing’ instruments that were also deployed on the ice will ultimately be compared to data from the KuKa radar.
As well as better understanding the correspondence between measurements of existing Ku- and Ka-band satellite missions (e.g. CryoSat-2 and AltiKa), these measurements also help lay the groundwork for the CRISTAL mission. This is a proposed ‘next generation’ satellite mission that will carry both Ku- and Ka-band instruments, potentially launching in 2027.
¹ satellite radars don’t actually operate at one frequency, but instead over a small, continuous range of frequencies. That’s what a frequency band actually is.
² the two radar instruments have slightly different beam widths, so one footprint is slightly larger than the other. The radar footprints are also not exactly overlapping.