Thursday, 19 June 2014

Birds and Light - UV Light

CSI Birding

HERE is a useful if somewhat dated summary of UV vision and patterning in birds.

Many flowers have amazing hidden UV patterns (NECTAR GUIDE or honey guide) to attract and direct bees to their nectar.  When photographed with specialist UV equipment, these flowers bare virtually no resemblance to how they appear to us in life.

Wouldn’t it be amazing to find that two, near identical looking birds might have a hidden pattern in their plumage that would make ID far easier with the help of the right optical aid?   Well it turns out that, in some birds where the sexes appear identical to humans there is sexual dichromatism going on in the UV spectrum.  This is also for example the case with many butterfly species.  So a UV camera might be a handy way to tell the sexes apart in these instances.

It would also be nice to discover that hidden UV patterns might actually be responsible for some species divergence.  Perhaps even a cryptic species or two may await discovery.  Taxonomic work in this area is very much in its infancy it seems.

UV vision capability was the default of our vertebrate ancestors.  Humans and many other animals have lost this capability.  UV is blocked from penetrating the surface layers and lens of the human eye.  Humans who have had cataracts removed and replaced with lenses that can pass UV light may find that they can actually peer a little into the hidden UV world!  Humans who can see wavelengths below 400nm report the colour as  violet or purple.  

A number of studies have found UV reflectance and UV vision to be common and widespread in birds. The mechanics of vision in birds is very different to humans.  Many birds have a fourth cone and can see in four colour channels rather than three (TETRACHROMACY).  It is hard to imagine what this might be like but I guess it might be similar to imagining the difference between looking at photos in red, green and blue channels and then comparing that with working only in red and green.  It seems rather simplistic to consider all UV wavelengths as being all the same rather dull violet.  Just as there are a whole range of blue hues to be experienced within the band of wavelengths we call blue, the same must surely be true of the UV spectrum.  For birds that possess the fourth cone, they may well experience colours humans can't even comprehend.

Birds appear to be using UV reflectance and absorption patterning in a rather more subtle way than in the case of the flower and bee relationship.  So, perhaps it is too much to expect such dramatic patterning to occur as commonly in birds as it does in plants.  Hidden UV patterning in birds seems to be quite a rare phenomenon.  There seems to be a clear link between visible colours and UV reflectance.  White feathers reflect all wavelengths well, including UV.  Blue feathers reflect UV more commonly than other coloured feathers, while black feathers don't reflect much UV at all.

We don't really know what bird's actually see when they see in UV.  But it is fun to guess!  In the case of the Cockatiel the UV patterning appears to be confined to the white wing feathers and off-white flesh around the eye and cere.  White feathers appear white to us because they reflect all visible wavelengths, which combine to create white light.  White feathers reflect UV well also.  So does white light and UV light combine to form white light, slightly purplish light, or something else?  I am fairly sure the composite above is incorrect and that birds in fact see white feathers as white, just as we do.  The real question is, how does UV reflectance alter the appearance of other colours.

Birds often use UV in combination with bright colours that are already visible to us, so from an ID 
perspective there is probably no added advantage in being able to see hidden UV reflectance in most cases.  UV patterning for the most part may simply enhance or slightly alter the appearance of colours under different lighting conditions, in a manner similar to how iridescence and similar structural colours allow birds to alter their appearance under different lighting.  Notwithstanding of course our inability to see UV patterns without special optics, just as iridescent colours are a somewhat poor aid to field identification, it may be that UV colours are not very useful for ID purposes either in many cases.  

In this paper (HERE) the authors state that there has been an apparent failure to find spectacular hidden UV-monochromatic patterns in bird plumage.  There is however a really interesting study by Robert Bleiweiss (HERE) which showed distinct differences in UV reflectance markings in the plumage of two very similar Mountain-Tanager species, including differences at sub-species level, suggesting that indeed there may be scope here for further discoveries and possibly even some taxonomic implications.

As the graph HERE shows, the earth’s atmosphere is very good at eliminating most UV, down to mainly UVA (315 – 400nm) with a very small amount of residual UVB (280 – 315nm) and UVC (100 – 280nm) radiation.  It would appear that avian vision and plumage markings are based solely on UVA.

Considering that UV light intensity is generally quite low for a lot of the time, one might expect that birds would only make use of it when conditions allow. There is a known link between UV and mate selection in some birds.  As the summary linked (HERE) demonstrates, the level of UV light is directly linked with overall solar radiation which is in turn directly correlated with latitude and time of year.  During the summer, when most birds in temperate climates happen to be breeding there also happens to be more UV about.  On the other hand, light intensity and UV is relatively more constant in the tropics.  UV intensity is also related to altitude (see graph above), because the atmosphere is thinner the higher up one goes.  It may be therefore that birds in temperate climates and birds living at higher altitudes are more likely to use UV in mate selection than birds in the tropics – I am not sure how much data is available to test this or whether or not it has already been tested and confirmed.  Looking at the wide difference in light intensity between sea-level and upper atmosphere in the UV and visual spectrum up to approx. 450nm, one wonders what advantages high altitude species and migrants might be taking of this very different light signature. One must also wonder what impact Ozone depletion and the resulting increased levels of UV may be having on birds and other animals who can see and use UV and who's biology may be linked to UV in some way.

Birds are also known to use UV vision to aid in foraging.  The Common Kestrel is believed to be able to see UV reflected from urine trails left by voles, which it uses for tracking and hunting.  Interestingly urine phosphoresces under UV (UV light is turned into visible light by phosphors in urine).  I wonder if this partly explains the Kestrel's ability?  UV reflectance and absorption by different fruits may also assist birds during foraging.

UV imaging of birds

By enhancing UV patterns and blocking out visual and IR wavelengths we can study UV patterning in birds.  As an alternative to simply displaying UV patterns, it may be possible through trial and error to try and capture and present images of birds with both visual light and UV patterning displayed together in a way that might mimic how birds see eachother and the world around them (eg. Cockatiel above).  Such images might help us to find explanations for some aspects of avian behaviour and biology and just how birds benefit from their ability to see in UV. 

As regards the equipment for delving into the UV world.  The standard approach of the UV photographer in the digital era seems to be to have an off-the-shelf digital camera or camcorder modified by removing the UV and IR filters over the digital sensor and replacing them with an appropriate glass which can pass full spectrum light (from low wavelength UV to high wavelength infrared, and everything in between).  Nikon cameras work best apparently and some models (eg. D70S) can record UV images without any modification.  The irony is, all CCD and CMOS digital camera sensors are already fairly good at capturing UV (to an extent) and IR (much better).  Manufacturers put filters in the path of the sensor and apply coatings to remove UV and IR.  For most photographers UV and IR wave bands are unwanted as they can create artifacts in digital images. 

Special UV pass lenses are expensive though older lenses with fewer glass elements and less coatings often work well in UV photography.  Some photographers have taken to scrubbing the coatings off lenses as an alternative to buying specially designed lenses.

The final but most critical element in the camera setup is the choice of filter or filters over the lens to selectively pass UV and eliminate visible and IR wavelength radiation as required.  These filters are not cheap  and may even turn out to be the most expensive element.  The best UV pass filter is said to be the Baader-U Venus filter - originally designed for use by amateur astronomers to photograph the clouds of the plant Venus.  It passes UV radiation in the range 300 - 400nm without any "bleeding" of IR and visible spectrum light.

One could spend an absolute fortune modifying existing equipment and paying for additional lenses and filters but some expense might be spared by zoning in on the specific wavelengths of UV that are of interest.  Birds see and appear to display plumage patterning mainly in the upper end of UVA (also called near UV or long wave UV) roughly at wavelengths from 350 – 400nm.  From what I read, it should not be necessary to pay for the most expensive quartz glass lenses to be able to pass lower wavelength UV light if such wavelengths are of no particular interest for avian study.  I have started experimenting with black glass and UV LED lamps to see if I can transform one of my old camcorders into a UV camera.  I am working on the assumption that the older camcorders, many of whom came with IR, Night Vision features may be quite good at recording UV.  Hopefully I will be reporting success soon!  

Perhaps the most important considerations of all are camera exposure and focus.  If the images to be captured consist only of UV wavelength light, the level of UV illumination, even on a sunny day in mid-summer may be too low for reasonable exposure of a moving target like a bird.  The solution might be to go with a digital camcorder instead of a DSLR as these can often out-perform DSLRs in low light.   The final solution may be a partial filtering of the visual spectrum as opposed to full-filtering.  This would provide a better mimic for what birds actually see, as clearly their vision is predominantly based on the visual spectrum, similar to ours.  This method would allow some of the camera exposure needed to create the image to come from the visual spectrum, while hopefully leaving enough of a clear impression of the UV patterning so that it stands out and can be appreciated in the image.  So far however, from what I can gather this doesn't appear to be very feasible.

Not only is exposure a problem, UV focuses differently to visible light and if the image being viewed is monochrome and difficult to see well it can be hard to get the focus of the image right.  

So, lots to be overcome, but hopefully all the effort will be worth it!


Thanks to Kevin J. McGowan for directing me to the following paper (HERE).  It seems that there has been at least one systematic study confirming that UV reflectance is ubiquitous across the whole bird kingdom.

Thanks also to Bailey D. McKay for his comment to the blog which references his paper on the use of digital photography in systematics (HERE).  That paper encapsulates a lot of the work I have been doing with colour on this blog.

Success!  Having purchased a Baader-U filter and tried it out with various cameras I have arrived at what I think might be the ideal UV imaging equipment for the birder.  For nore see HERE.

Sunday, 15 June 2014

Birds and Light - Lighting under the microscope

Grey Card Comparisons

I have been playing around with a test rig consisting of an X-rite Colorchecker and some additional targets mounted on a white board.  HERE I took a closer look at White Balance correction and explained a bit more about the rig.  I first put the rig to the test in a verdant forest setting and was really amazed by the results as outlined HERE.

Below I compare a variety of commonly encountered lighting settings.  I will continue to add to these in due course.  Here is a summary of what I have been uncovering.

A Tale of Two White Balances

Firstly an important point.  It is possible to have more than one correct white balance in any given image.  When there is more than one source of illumination in a scene, each illuminating a different part of the scene more or less independently, then it is indeed possible to correct for each of these light sources independently.

Surely the sun is the only source of illumination in an outdoor scene, one might ask?  While the sun may be the ultimate source, different entities within the environment, which reflect or transmit "modified" sunlight, themselves act effectively as alternative light sources.  Examples include the sky which transmits a powerful blue light, foliage which can transmit an intense greenish light and water which can reflect a virtual mirror image of an environment onto a subject, including indeed a mirror of the subject itself, in all its various colours.  Obviously, in order to correct for different sources the grey card must be placed in the path of each source individually.  The x-rite colourchecker allows for the simultaneous correction of two different light sources in the same scene as outlined below.

GOLD - Optimum Lighting 

The optimum light for observation and photography is a bright and overcast day.  This is due mainly to the light scattering affect of moisture droplets that go to form clouds.  Unlike the scattering created by gas molecules in the atmosphere which mainly just scatters the lower wavelength blue portion of daylight, moisture droplets scatter light of all wavelengths equally well.  The result is a soft, diffuse, neutral light, which baths a subject quite uniformly, even scattering into deep shadows. 

SILVER - Harsh Daylight 

There are a couple of issues with bright sunlight.  Cameras are limited in their ability to simultaneously deal with high luminance bright sunlight and the low luminance shadows left in its wake.  The ability of a camera to deal with light of variable luminosity is referred to as it's Dynamic Range.  For most digital cameras the result of taking a photograph in bright sunshine is a high contrast image with some burnt out details and colours.  Parts of the image are generally underexposed and parts are generally overexposed.

There is also an environmental affect.  The high intensity sun heats up the molecules in the air causing them to bounce around resulting in heat haze, which affects image focus.

From a white-balance perspective we also have, in effect, dual illumination going on as shown in the "tale of two white balances" example above.  The sun is casting a highly polarised, direct white spotlight onto a subject, while, at the same time, the blue sky canopy is casting scattered blue light on the subject from various directions.  As this blue light is much dimmer than the sun it is overpowered by the white light on that portion of the subject illuminated directly by the sun's rays.  However, in the shadows, where the sun's direct rays cannot penetrate, the blue light from the sky creates a clear blue colour cast.  

A single grey card cannot simultaneously deal with dual sources of illumination, unless of course the card happens to be positioned at the boundary between two independent light sources (e.g. partly in sun and partly in shade, as the example above illustrates).  Either way however, we can only make one correction to an image at a time.  Either we live with the blue shadows or we correct for those and end up with a yellow colour cast in the mid-tones and highlights.  Harsh daylight really is not ideal, either for observation or photography.  When trying to judge colours accurately the key point to remember is to verify highlights are properly white balanced and remember then that shadows will contain a bluish colour cast.

BRONZE - Deep Shadow

Deep shadows tend to be associated with bright sunny days.  They represent a gloomy juxtaposition to harsh, bright daylight.  As the above test proves, deep shadows are naturally bluish in colour due to the influence of blue sky canopy light.  However if the whole subject and grey card are both in shade there is a chance that a single grey card white balance correction will eliminate this blue cast.  The results below show that, while there is a dramatic improvement, there does appear to be a residual blue cast even after white balance correction.  The other interesting finding is the tendency for targets to show a much more obvious reflection, or projection of the bright background, i.e. the sources of light peering into the shadow.  The overall result therefore suggests that shadows are not a uniform blue cast as they might appear, but are rather more akin to a cinema projection of the surrounding environment.  While it may not appear as dramatic as this (it appears as a subtle dappled affect) I think the presence of complex and bright background colours might impact our ability to judge colours accurately in the shade, even after a good white balance correction.  It would seem therefore that an image of a bird in direct sun offers a better chance of judging colours correctly than an image of a bird that is in deep shade.



Foliage Canopy
Though this particular foliage canopy test possibly out-performed either the bright sunlight or the deep shade tests, typically birds are not too cooperative in these environments, so shots tend to involve dappled light or tend to be under-exposed due to extreme shade.  A useful test nonetheless, particularly the revelation that reflected and transmitted light from foliage can be quite uniform in nature and can be more or less eliminated with a grey card white balance correction.