The colour of the ocean

C. V. Raman's famous paper on his observation of the colour of the sea whilst travelling by ship from England to India in 1921 is available from the Royal Society ( rspa.royalsocietypublishing.org/content/101/708/64 ). It is a wonderful illustration of Raman's intense curiosity and analytical combination of theory and experiment. It marks the beginning of his interest in molecular scattering which, seven years later, resulted in his discovery with Krishnan of the inelastic molecular scattering that later took his name and won him the Nobel prize for physics.

There is much confusion in the literature about the causes of the colour of the sea - especially about the role of Raman scattering (of which Raman was of course unaware at this date). It is now realised, in the age of remote sensing of the ocean by satellite, that Raman scattering does play a relatively minor role in the red part of the spectrum. However the predominant effect is from molecular density fluctuations first described by Einstein and Smoluchowski in 1905 and 1906 respectively.

It is interesting to see if we really understand the colour of the ocean under a clear sky as seen from space: The Blue Planet, or the "Blue Marble" as NASA calls it.

See: visibleearth.nasa.gov/view.php?id=57723
and, more specifically, we can look at: eoimages.gsfc.nasa.gov/images/imagerecords/57000/57723/gl...

We should consider looking vertically downwards (the Gulf of Mexico in this case) with a high Sun - but not vertical (behind us) since we want to avoid the 'glint' (seen in this image round about the Baja peninsula). Assume the air is perfectly clear, i.e. no aerosols and no water vapour. Below the atmosphere is a calm sea of infinite depth with no suspended particulate matter: pure water.

Looking at the western globe image (actually a 'true colour' composite), we see that the Gulf water surface appears to be a very dark blue, a colour that is a combination of the light coming from the atmosphere and from the body and surface of the underlying water. Looking to the west (over the Pacific), the blue becomes lighter due predominantly to the increasing airmass as we move away from the vertical: more 'blue sky'.

The images shown above are the separate blue (left) and red (right) channels (B and R from the RGB (tiff) composite) from NASA'a western hemisphere Blue Marble. They show particularly clearly the effects of Rayleigh scattering from the clear parts of the sky and also the relatively higher luminosity emerging from the ocean in the blue part of the spectrum (compare the sea with the brightness of Florida).

The light coming from the clear sky is almost (modulo a different scattering angle) the same as a clear zenith sky seen from sea level - which is Rayleigh scattered sunlight. There will be a contribution from Raman scattering as well (the Ring Effect) but this will have little perceptible effect on the colour we see since the main result of this is a slight filling in of Fraunhofer spectral absorption lines. Since we have assumed no water vapour, the only remaining 'selective' (Raman's terminology) absorption will be due to O2, O3 and O4 which, apart from O2 in the far red, is very small. The attenuation of the scattered sunlight by extinction is small since both the Sun and our viewing angle see an airmass ~ 1. The atmosphere is essentially a single scattering screen.

The surface of the calm sea will reflect about 2% of the skylight back. The specular reflection of the Sun can sometimes bs seen as a 'glint'.

Within the body of the (pure) water we have the processes that Raman describes in his 1922 paper: selective absorption by liquid water at redder wavelengths, due to vibrational overtones in the water molecule, and elastic molecular scattering from fluctuations in the density of water molecules. This latter effect is like Rayleigh scattering from gas molecules but includes the interference resulting from phase relationships (coherence) between photons scattered by neighbouring molecules. There is also, of course the (inelastic) molecular scattering that Raman had not at that time discovered!

As in the atmosphere, the main effect of Raman scattering is to fill in absorption lines but this has little effect on the colour. Raman scattering is, however, sufficiently important in the sea that it is taken into account when measuring sea colour from satellites and retrieving information about particulate matter, including chlorophyll ( onlinelibrary.wiley.com/store/10.1029/2002GL014955/asset/... ). By down-scattering copious blue photons to redder wavelengths, the Raman scattering has its greatest fractional effect on water colour at red wavelengths where the selective absorption by water is large and the scattered red light is very weak. It is fascinating that in deep water (> a few 100m), fish have developed effective camouflage for downwelling blue light. However, some predatory fish have developed extreme red visual sensitivity that allows them to break this camouflage by using the very weak, isotropic red light that results from multiple Raman scattering of the copious blue photons (see: "The Optics of Life", Sönke Johnsen, Princeton University Press, 2012, p198).

The scattered light emerging from Sun-illuminated water comes mostly from around unit optical depth (tau) at each wavelength. The physical depth associated with tau = 1 in pure water is at its greatest value of about 50m at a blue wavelength of 480nm. This means that there is very little light emerging from deeper than a few hundred metres. In a way, the resulting spectrum is analogous to the sweep of twilight colours from red through pale apple-green to deep blue, the combination of the scattering source function tau(lam) ~ 1/lam^4, and the extinction sink function exp(-tau(lam)) results in the green-blue peak in the backscattered light that Raman describes.

This is illustrated in the plot above which is derived from Table 1 in Raman's paper where he presents the expected luminosity emerging from deep, sunlight-illuminated water expressed in terms of the brightness of a clear zenith sky (orange line) at each wavelength: corresponding to an 8km atmospheric path. In our diagram, a measurement of the flux from a blue sky (blue line) has been shown, normalised to 8km near 555nm. The green line is the product of these two curves (with the sky normalised now to unity near 555nm) and therefore represents the emergent flux/nm of light from an illuminated ocean. The weakness of the red light is due to the intrinsic absorption of water in the vibronic overtone bands - which increase in strength towards the red - and the the blue elastic scattering of water molecules in the green and blue. I reiterate that there is no Raman scattering in this plot and, if there were, it would make little difference.

So the bottom line is that the blue of the water is due predominantly to elastic molecular scattering, strongest in the blue, and to the intrinsic absorption of the red light exciting the internal vibrations of the water molecule. This results in a markedly different spectrum from that of the blue sky.

The addition of water vapor and aerosols to the atmosphere and particulate matter to the ocean will produce a range of effects that keep the Earth Resource community gainfully occupied, but I don't intend to discuss those here. However, look at some of the brilliant turquoise shallow waters on the Blue Marble!

Flickr page for this photo.

Visit the Eye for Science Flickr group.

Get the "Eye for Science" widget here and install it on your own blog. Spread the scienticity!

Like this image? Want it for your smartphone wallpaper?
Use this QR Code to load it on your smartphone and enjoy!