The color of the ice observed in glaciers, icebergs and crevasses is often blue. However, ice cubes and industrial ice blocks are perfectly transparent or white if not.

So. Why does glacier ice look blue?

Some examples of what I mean:

Ice blocks enter image description here

Glacier ice enter image description here

PS: As a glaciologist, this is a question I've received a lot, and after Googling for it, the answers I found are not fully satisfactory, so I decided to share mine here in a Q&A format.

  • 4
    $\begingroup$ A great example for us scientists of what to do when it gets a bit quiet around here on Stack Exchange. A well thought out, instructional, non-self-serving Q and self A. Great idea whenever you notice that no online explanations give complete enough answers. We have a lot to teach still :-) $\endgroup$ Commented Jan 31, 2018 at 16:24
  • $\begingroup$ A sort of related story about blue lake ice. $\endgroup$ Commented Feb 28, 2018 at 1:15

1 Answer 1


The short answer is: BECAUSE THE ICE IS BLUE.
Now we have to explain why it seems perfectly transparent on ice cubes and industrial ice blocks.

It has to do with the fact that most transparent materials are not perfectly transparent, and instead absorb (and/or scatter) part of the light that hits them. And when the transparency is better for one specific color (wavelength), the material will look that color, much like how stained glass gives a tint to the white light going through it.

Now the important fact is that this "tint" effect is stronger the thicker the material is. Let's take a standard glass as a example: When looking through the glass it seems perfectly transparent, because the thickness is too small to make its color noticeable. However, if you look at the edge of the glass, it looks green, which shows that glass is indeed green, but that only became noticeable when looking through a section thick enough.


(Picture taken from peakglass.ca)

In the same way, glacier ice is blue, but that can only be appreciated if the piece of ice you are looking at is thick enough. We can see this effect in the following picture of an iceberg, where sections of different thicknesses displays a color gradient ranging from transparent to blue. ice colors

The remaining questions is: Why are glaciers not always blue? As we know, many icebergs (and snow) are white.

The key here is "bubbles". The reason is that for each ice/air interface, part of the light gets reflected and only a fraction goes through. Therefore the greater the number of bubbles a piece of ice has, the less light will be able to proceed deeper into the ice. As such, when you view light from more bubbly ice, most of it will have traveled through less ice, and thus have had less distance to obtain the blue tint. I've made this figure to help understanding the idea: light in ice with bubbles diagram

On the left, a piece of ice with very few bubbles allows a longer travel path through the ice, inducing a marked blue tint in the outgoing light. On the right, in contrast, if there are many bubbles the light will go bounce back right away with only a very subtle tint, or no tint at all.

Snow is the extreme of this, because it is mostly air (bubbles), and only some small pieces of ice, that's why it looks so white.

An interesting ramification of this: those deep blue pieces of ice started as snow. But as the snow piles up, the weight of the snow on top compacts it, squeezing air out changing it from snow to firn until you finally reach the density of being a block of ice (with bubbles). In a big glacier, hundreds or thousands of meters of ice can exert such a pressure that the bubble spaces get compacted smaller and smaller until they disappear and the air is dissolved into the ice itself.
Therefore this means that the more blue (and bubble-less) a piece of ice is, the more it has come from deeper within the glacier... meaning it also has traveled from further up in the glacier basin and it is, therefore, older. So, now you can identify deep, old ice just by its color.

To give a more familiar example of how adding bubbles can turn something transparent with a tint into something white, take the case of egg whites, which are transparent with a yellowish tint. But as you whip it, adding in bubbles, it begins to turn into meringue which is perfectly white. making meringue

However, there is a trick that can even make the blue color of seemingly-white snow noticeable. If you let a ray of light bounce on it many times, each time it will travel through a little bit of ice, getting a bit of blue tint each time, and after enough bounces it will look unquestionably blue. That's what you see when looking at the entrance of a snow cave or a crevasse in a glacier. The light that comes from deeper in the cave has bounced many more times around the walls and becomes bluer: enter image description here (Picture taken from outdoorlifenorway.com)

  • 2
    $\begingroup$ Great answer. I always just foolishly allude to "different ice crystal shapes". Which certainly is valid in some things (why we have such a different array of snow/graupel/sleet/etc). But your answer explains it much more thoroughly and clearly than I've seen :-) $\endgroup$ Commented Jan 31, 2018 at 7:42
  • 5
    $\begingroup$ Also, to those who learned from reading this: note that that the color of (liquid) water itself works the same way as the glass thickness idea. Oceans are not blue due to the sky being blue (nor vice versa). $\endgroup$ Commented Jan 31, 2018 at 7:47
  • 1
    $\begingroup$ Great answer! Did you take all the photos/images yourself? If some are taken by others, could you provide a source attribution? $\endgroup$
    – gerrit
    Commented Jan 31, 2018 at 12:19
  • 1
    $\begingroup$ @gerrit Thanks. Not all pictures are mine, I've added references to where I took them from. I'll try to do that more often. Cheers $\endgroup$ Commented Jan 31, 2018 at 18:12
  • 1
    $\begingroup$ @gansub Glacier ice have a range of densities. Young ice is by definition as dense as the oldest firn, that's about 840 km/m3. But in old ice, that have been subject to great overburden pressures, the bubbles basically disappear and it can reach the density of regular bubble-less ice, which a is about 917 kg/m3 $\endgroup$ Commented Feb 4, 2019 at 1:51

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.