How does an iceberg get such an intense color of blue?

July 10, 2006 9:18 PM

Blue ice. How? (gorgeous picture involved, best one you'll see today)

The photographer claims this is an un-doctored image. It was taken near Argentina. He says the 'berg was tipped upside-down by the wind. What is going on that makes ice have this intense color? anyone wanna clicky this? thanks.
http://flickr.com/photo_zoom.gne?id=184500912&size=o

posted by UnclePlayground to science & nature (44 comments total)

cl-click.
posted by icosahedral at 9:24 PM on July 10, 2006

that lighting is all wrong. It's completely obvious that it's altered.
posted by cellphone at 9:25 PM on July 10, 2006

photographer is lying
posted by unSane at 9:25 PM on July 10, 2006

unless there's some magic filter that will do that.
posted by cellphone at 9:25 PM on July 10, 2006

4th Google result for "blue ice".
posted by ikkyu2 at 9:26 PM on July 10, 2006

How fast is it going?
posted by pompomtom at 9:27 PM on July 10, 2006

Um, it's altered. Obviously, and badly.
posted by ruby.aftermath at 9:27 PM on July 10, 2006

They really are that blue.

Google Answers has tackled this one.

The site it refers to blueiceberg.com doesn't allow deep links, so you have to navigate around to find the explanation there.
posted by aneel at 9:27 PM on July 10, 2006 [1 favorite has favorites]

I am curious about whether it may be merely an optical illusion rather than being blue. As if the pattern into which it was frozen is refractory in some capacity that magnifies the blue of the ocean throughout itself?
posted by vanoakenfold at 9:32 PM on July 10, 2006

Not to say the photo wasnt messed with.

The same guy took this waterfall (taken by someone else)

http://flickr.com/photos/bbr78/44789342/in/photostream/

and turned it into this:

http://flickr.com/photos/ricardobevilaqua/180662917/in/photostream/
posted by vacapinta at 9:34 PM on July 10, 2006

looking at the other shots in his photostream make me really wonder if he didn't doctor every one. I don't have the eye to say how, exactly, but they look really bizarre and fake.
posted by docpops at 9:35 PM on July 10, 2006

That photographer may have bumped up the saturation on that shot, but some glacial ice really is very blue.

I was on a big boat when I took these, so I couldn't get that close, and the glacier I was looking at was mostly covered with newer snow, so there was a white coating over it, but where you could see fresh cracks, or deep into the ice it was a gorgeous deep blue.

that lighting is all wrong. It's completely obvious that it's altered.

If the day that photographer was shooting was anything like the day that I was in Torres del Paine, it was mostly overcast, with occasional breaks in the clouds that lit the ice with bright sun.
posted by aneel at 9:38 PM on July 10, 2006

I concede, but I think the saturation was very noticeably altered, regardless.
posted by cellphone at 9:39 PM on July 10, 2006

vanoakenfold: yes, it's an "optical" effect, not because there's a pigment absorbing all of the non-blue colors. It's similar to why the sky is blue or why a blue butterfly wing is blue.
posted by aneel at 9:45 PM on July 10, 2006

The sky is overcast, but the water is glowing brightly.
posted by lain at 9:47 PM on July 10, 2006

Actually, I'm wrong about the butterfly wings. That's a different kind of structural color. But also not due to a pigment in the material.
posted by aneel at 9:54 PM on July 10, 2006

Bulk ice is blue; unusually for a solid object in that region, this one isn't covered in snow.
posted by flabdablet at 10:26 PM on July 10, 2006

I have photos of blue ice from the top of a glacier in New Zealand. Undoctored the blue isn't nearly as striking as that iceberg, but it was most definitely blue ice. I'll see if I can dig them up to share.
posted by shoepal at 10:51 PM on July 10, 2006

That picture (or atleast the iceberg) is just a 3D rendering...totally fake.
posted by JPigford at 10:51 PM on July 10, 2006

Water has a bluish color as well.
posted by null terminated at 11:25 PM on July 10, 2006

yeah thats from terradreams or maya. straight up 3-d rendering.
posted by sophist at 11:28 PM on July 10, 2006

Ice is blue, though that photo does look enhanced.
posted by LobsterMitten at 11:57 PM on July 10, 2006

That's what happens when you take something like a glacier with a bluish tint and use photoshops "auto colors" on it. The same thing happens to the vacation pictures of Alaskan glaciers I've done it to.
posted by Justinian at 12:08 AM on July 11, 2006

Okay first of all. It is possible that the image came out of his camera looking like that, not because it "really looked like that" but because he did a long exposure.

The waves crashing around the base of the thing are blured, suggesting a long exposure. Long exposure makes things look far more colorfull then they do to your eyes.

For example this picture of mine, which is a 15 second exposure, I think. It's still not nearly as brilliant as the glacier in that shot.

the other problem is that the background and the sea are not super-saturated, so I think it was photoshoped.
posted by delmoi at 12:10 AM on July 11, 2006

this photo has obviously had the saturation jacked up, and it's during the daytime so it's obviously not a long exposure.
posted by delmoi at 12:12 AM on July 11, 2006

In this picture the clouds look really fake (not the only one with fake looking clouds). I think the guy is just a photoshopping HACK.
posted by delmoi at 12:16 AM on July 11, 2006

They might be High dynamic range photos. Examples here and here. I think these can also give the photos the appearance of a long exposure (blurry water, etc).
posted by clearlydemon at 12:51 AM on July 11, 2006

Guten tag, Her Doctored!



Here's a screen capture of the image blown up in Photoshop. There are a couple of problems noticeable right away.

1. The most obvious problem is the smudge right in the middle of this capture. The edge is clearly defined on either side of it. It almost looks like he tried blending the berg into the grey background, then thought twice about it but never went back and cleaned it up.

2. Upper-right looks like a bad cut-n-paste (some parts of the iceberg are missing if you follow the edge). It also appears a few grey pixels of the background sky are "bleeding through" a hole in the edge. Probably used the magic lasso to cut the object instead of taking his time. Tsk, tsk.

3. High contrast between berg and background. So instead of a smooth transition from "close object" to "far sky", you'll get this: bright white followed by a 1 pixel line of dark blue grey followed immediately by a lighter blue-grey.

4. A noticeable absence of any meta data. I mean, like, none. Perhaps that's just flickr... but usually you expect to see something in the header info.

My seat-of-the-pants guess is that this is a real iceberg, but the "artist" cut the berg from the original, tweaked the colors, then overlayed it on the original and tried blending it in. Badly.

My other evidence is, well, the entire photographer's gallery. Exhibit A. B. C. D.
posted by Civil_Disobedient at 12:51 AM on July 11, 2006 [1 favorite has favorites]

4. A noticeable absence of any meta data. I mean, like, none. Perhaps that's just flickr... but usually you expect to see something in the header info.

Flickr actually keeps the metadata example.
posted by delmoi at 2:22 AM on July 11, 2006

This picture is obviously faked. But glacial ice is blue (just not that blue!). From www.carleton.edu/departments/GEOL/Links/AlumContributions/blueice/blue.html

[Ice] only appears blue when it is sufficiently consolidated that bubbles do not interfere with the passage of light. Without the scattering effect of air bubbles, light can penetrate ice undisturbed. In ice, the absorption of light at the red end of the spectrum is six times greater than at the blue end. Thus the deeper light energy travels, the more photons from the red end of the spectrum it loses along the way. Two meters into the glacier, most of the reds are dead. A lack of reflected red wavelengths produces the color blue in the human eye.
posted by maxpower at 4:54 AM on July 11, 2006

If it helps, here's an (un-doctored) photo of a blue iceberg from my trip to Tracey Fjord, Alaska. Also, what maxpower said.
posted by tlong at 5:05 AM on July 11, 2006

I agree with Civil_Disobedient. The basic composition of the photo may not have been altered (much) but it looks to me like the photographer drastically lowered the saturation of the background while increasing the saturation of the iceberg. This is not at all hard to do in Photoshop, particularly if you are using a version that supports layer masks.
posted by TedW at 5:36 AM on July 11, 2006

All the blue ice chunks and icebergs I've seen in Alaska and Canada have been a blue shade like the ones in aneel's pictures. The picture in question in the opening post also looks false to me.
posted by Savannah at 7:29 AM on July 11, 2006

About the metadata: If you choose save for web in photoshop cs2, it appears to strip out a lot (all?) of the metadata. It won't, for example, leave in anything about the camera the photo was taken in, or any of the settings of the camera.

So now I use save, not save for web.
posted by cactus at 7:47 AM on July 11, 2006

It's pretty easy to determine whether the color's been played with: pull the image into Photoshop (which I don't have handy right now), and look at the 4-color histograms.

If any of them are non-continuous, then *something* has been done to the picture (where by non-continuous, I mean "patterned gaps in the curve").
posted by baylink at 8:22 AM on July 11, 2006

That image is doctored. C_D has it. He put that iceberg in front of that landscape by manipulating two different photos. All of this is besides the point of whether icebergs can be a brilliant blue: they can, although his exposure is extremely over-saturated via photo manipulation.

Here's the histogram. See that midtone clipping across all three colors? That's not very easy to do without some hanky-panky. (baylink, merging two different exposures will often give you a compressed, rather than expanded, histogram.)


posted by deadfather at 9:30 AM on July 11, 2006

Yeah, but I was more looking for histograms with big comb-teeth gaps in them. Those are less indicative than i was hoping for, though thanks for pulling them out, anyway. :-)
posted by baylink at 9:57 AM on July 11, 2006

baylink, my point was that this type of doctoring wouldn't produce that comb-and-teeth style expansion, which is indicative of a single image being fudged around. The big midtone clip in all three colors is highly suggestive of two images being combined into one.
posted by deadfather at 10:05 AM on July 11, 2006

Well, you know, this *could* indeed have been a 30-60 second exposure by moonlight, which would account for almost everything...

Whichever, it makes a pretty desktop background. :-)
posted by baylink at 10:29 AM on July 11, 2006

Here's some blueish ice from atop Fox Glacier in NZ. (undoctored and not cleaned up in any way.)
posted by shoepal at 10:30 AM on July 11, 2006

I seriously doubt that wind can tip over an iceberg, since something like 85% of the ice is under water. So I think his explanation of how that photo came about is... rather suspect.
posted by oneirodynia at 11:02 AM on July 11, 2006

Blue icebergs do exist, but this guy over-saturates many of his color photos.
posted by redteam at 12:51 PM on July 11, 2006

If you choose save for web in photoshop cs2, it appears to strip out a lot (all?) of the metadata.

There's all kinds of meta data in your standard JPG file. Not just the EXIF data. Open a file and save it in Photoshop, and you can clearly read that the file was created in Photoshop.

The only way an image would be completely devoid of meta data is if someone went out of their way to clean out the data, or (as already proffered), Flickr does that for you. So at least that part is explained.
posted by Civil_Disobedient at 5:49 PM on July 11, 2006

Um...in Flickr you just chose whether to share EXIF data or not. He has chosen not to share it.
posted by jb at 7:12 PM on July 11, 2006