Why does polarized light shift color in reflections on the sea?

[Artlav]

Hello.

I've noticed a thing i could not understand, and hope someone can explain it to me.

Given: The sun at about 45* angle above, the calm sea flat below.
The sky is gray with thin, fog-like clouds.
There is a reflection in the sea of the sun above.
http://orbides.1gb.ru/img/pol-1.jpg

Now, let's look at it through a horizontally polarized filter:
http://orbides.1gb.ru/img/pol-a.jpg

There is an area, roughly 45* looking down, where the part of the reflection is bright and saturated blue.

Turning the filter 90 degrees, and there is nothing special:
http://orbides.1gb.ru/img/pol-b.jpg

Why is there the blue color?
As far as i understand, the light reflected from a border between two mediums with different indexes of refraction would be partially polarized perpendicularly to that surface, regardless of wavelength.

What else is at work here?

 

[Andy Resnick]

That's an interesting effect.

I'm not entirely clear, but presumably when you say 'horizontal' and 'vertical', you mean with respect to the horizon. Also, I am assuming you oriented the polarizer (w.r.t the horizon), and then tilted and tipped your camera down, changing the orientation of the polarizer with respect to the water surface.

Light reflected off a water surface is primarily polarized in the horizontal direction (Brewster's angle): the P-polarized component is poorly reflected.

However, your blue speckles are clearly reflections.

I am wondering what exactly you are using: is it a linear polarizer only? are there other filters in place (e.g. is it really a circular polarizer?) Does this effect happen only for certain "magic" angles, or is it a general effect?

I'd like to understand this as well...

 

[Dadface]

Another effect is that the atmospherically scattered wavelengths,primarily around the blue area of the spectrum and hence the reason for blue skies,are themselves partially plane polarised(completely plane polarised for scattering angles of 90 degrees).