The false color image below shows synthetic aperture radar imaging from ESA's Sentinel-1 satellite. The map from the 20-Jan-2017 BBC article Larsen ice crack continues to open up illustrates the same general area.

The Larson C Ice Shelf is the large blue-ish region in the middle of the false-color radar image, and is labeled as such. The crack is the thin lighter line just above the red arrow.

Moving diagonally upwards to the right, the gray region is labeled "Sea Ice". Is this just much thinner but mostly solid ice, or water with pieces of ice in it? While the ice shelf seems relatively uniform in tone, the gray sea ice area has a mottled pattern.

Also the boundary between the two regions has what looks to me like some kind of 3D effect - the black, ribbon-like variable width region separating the two. Is this an effect of shadowing from oblique incidence? Is this actually the sidewall height of the ice shelf?

There is also some discussion of radar imaging of sea ice in this answer.

enter image description here

above: Cropped and enlarged highlight from Sentinel-1 synthetic aperture radar image of the Larson C ice shelf, beginning of January 2017, found here. Image credit: ESA/EU/Copernicus

enter image description here

above: Sentinel-1 synthetic aperture radar image of the Larson C ice shelf, beginning of January 2017, found here. Open in separate viewer or tab for full size. The width of the image is roughly 300km, south is towards the lower right. Image credit: ESA/EU/Copernicus

enter image description here

above: Map of the region around the Larson C ice shelf, beginning of January 2017, from BBC here. Image credit: MIDAS Project, A. Luckman, Swansea University. Because it's difficult to read, I'll point out that Wales not shown in it's correct location - it's there for scale only.


1 Answer 1


natural-color I found a natural-color image of the same area: 68.28 South, 60.47 East, North is up, resolution is 130m/px, which makes the whole image about 100 km x 100 km. Looking at the image i'd say that:

  • sea ice is water with pieces of ice in it,
  • the boundary between the two regions is clear water, probably caused by winds and currents moving the floating ice away from the edge of the ice shelf.

I got the image at Sentinel Playground using 100% cloud filter, 0.3 gain and date selection of 2016-11-21.

EDIT: Added more info about the position, scale and orientation of the image.

  • $\begingroup$ The location is 68.28 South, 60.47 East, and the scale is about 130m/px. North is up in the image. $\endgroup$
    – mkadunc
    Jan 24, 2017 at 14:09
  • $\begingroup$ OK great! The image makes much more sense to me now. Thanks! One quick question - is your image from visible light, IR, radar, or something else? $\endgroup$
    – uhoh
    Jan 24, 2017 at 14:39
  • 1
    $\begingroup$ Side note: There's a distinction between sea ice and the ice of an ice shelf. The ice shelf is the extruded end of a glacier, snow and ice mounded on land over thousands of years and pushed out to sea; sea ice is the ice that has formed on top of the ocean. Sea ice is usually 1-2 meters; this region of Larsen C is about 1,100 meters thick. So there's a significant height difference for casting a shadow. $\endgroup$ Jan 24, 2017 at 16:30
  • $\begingroup$ @jeffronicus I'm sure you are saying something illuminating but I'm not getting it. Every time I look at the image in the question, my brain tells me it's three dimensional, and the dark ribbon in the middle is a shadow. But it isn't, is it? Is the ice shelf also ~10% above the water, ~90% below? $\endgroup$
    – uhoh
    Jan 25, 2017 at 17:33
  • $\begingroup$ I'm not an expert on radar sensing, but i think for a cast shadow to be visible in an image you'd need the light source direction to be different from viewing angle. Radar emits its own 'light', so unless the image has been orthorectified to change its apparent biewing angle, you won't see cast shadows on radar imagery. Also, for a couple of km long shadow on a cliff that's ~ 100m high, the illumination angle would need to be very low, definitely not consistent with sentinel-1, which i think is pointed more or less directly downwards. $\endgroup$
    – mkadunc
    Jan 25, 2017 at 17:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.