Illumination data needed for available light calculations
Posted 21 February 2005 - 04:51 AM
I am fiddling a bit with calculations of available light spectra at depth and use of filters. I am trying to compile as much data into my calcs as possible -- for instance, using measured absorption coefficients and tables of filter transmittance. The two bits of the problem that I am still looking for data on are
1. Actual illumination spectra for different weather conditions (e.g., clear versus cloudy), times of day, and latitudes. Ideally, I'd like to have intensity vs wavelenght at more-or-less 10 nm spacing over the visible (400-700 nm) range. My fallback is to use CT or CCT -- but measured spectra would be a nice adjunct.
2. Reflectance data -- still looking on this. Would like to have it for a variety of conditions, also.
In both these areas, I am googling away, and will also make another trip to the library Tuesday -- but, strangely, tables of illumination have been harder for me to come by than attenuation were (and the filters were easy, thanks to Kodak)...
I am looking for tables of data -- either digital or hard copies that are clean enough for OCR. I'd rather not introduce (more) errors by reading points from graphs.
As I mentioned, I already have a bunch of attenuation data and filter data loaded. I hope to add the illumination and reflectance data, along with some very simple calculations, and post later this year. Any help appreciated!
All the best,
Subal ND70 + 2X DS125 ... mostly 10.5, 16, and 60
Posted 21 February 2005 - 06:21 AM
I don't have any data. But best of luck. A most commendable project.
I do not want to sound too negative but I fear that you might be trying to over simplify an incredibly complex natural system. And for photographic purposes, while we want the filter to most of the work, as long as it gets us close, white balancing will do the rest. Further more changes in your light path (depth + subject to camera), which can be difficult to control when actually using filters in the ocean (the damn whale shark was swimming 1 m too deep) make much more difference to colour temp than many of these variables.
Other parameters that change the colour temperature of UW light at a certain depth are:
1) depth of the seabed. If it is close to your working depth and is white sand there is much more blue upwelled light.
2) time of day. Changes in colour temperature and also refraction of light. When the sun is low much of the direct sunlight is bounced off the water. Instead much of the UW illumination is reflected off the clouds or from the atmosphere - this is cooler colour temp than sunlight.
3) weather conditions. Sun, hazy clouds, thick clouds, rain all effect colour temp above water and hence UW. As you say. But these are not decrete events, but a sliding scale.
4) sea surface state. A rough sea state will reduce the amount of direct sunlight underwater. Instead a higher proportion of light reflected off clouds. Also more refraction at the seasurface means more diffuse light UW. The slightly longer light path (because the light is more physically scattered when it entering the ocean means cooler colour temp.
5) water properties. Such as phytoplankton content, sediment load and gelbstoff also effect its optical properties. As underwater photographers we often talk about green water and blue water as seperate entities when discussing filters. They are of course ends of a continuous spectrum.
You best bet for data that i can think of is in ocean colour research. Satellite Oceanographers use ocean colour satelites to study many aspects of the oceans. Most notably phytotplankton blooms. They have done a huge amout of work on the optical properties of water. Try googling "underwater spectrophotometer"
When it comes to UW photography and filters I really think an empirical approach is more sensible than a theoritcal one. But I would love to be proved wrong. Please share your results. I really will be fascinated to see them.
Posted 21 February 2005 - 10:04 AM
I have little prospect of proving Alex wrong. As a modeler of natural systems in my day job (oil and gas reservoirs, aquifers, and sediment transport), I have learned to be humble and cautious. However humbled, I still find that modeling helps me think through the relations and "what-ifs". And I can take a break (as I am now, from the dreadful task of marking exams!) and fiddle with a model for a bit -- I can dive "my sort" of sites only a couple weeks a year.
As Alex suggests, I have found a cornucopia of data in the physical oceanography literature -- so far, mainly for attenuation of pure water and plankton assemblages. I am still looking for good reflection and illumination data, as I mentioned in my post.
So I am using a simple but data-rich approach: use real measured spectra (dependent on weather, latitude and time, as you say), and apply measured reflection models (depending on sea state and sun angle, mainly) to get the illumination model just below the water's surface. Then apply nonuniform attenutation models (again, from measurements in the literature) to get the spectra at depth. Then filter these (using published transmittance tables) to try to tune to a desired effective illumination. It's a game, entertaining to troubled modelers such as myself, only as fun as it is calibrated to measurements...not as good as halo, maybe, but fun all the same.
So, in all humility, what do I expect to get from this?
* An understanding of how useful color temperature is as a concept, when nonuniform attenuation is applied to typical (nearly ideal "black body") illumination sources. That is, does color temperature really mean anything at depth? (I suspect the answer is "not really").
* For reasonable assumptions of illumination, reflection, and attenuation, which filters might one use to get near to a target illumination spectrum? I suspect the answer is similar to what Alex and Craig Jones have converged on (e.g., combinations of warming conversion and correction filters). But it'd be interesting to see if some different combinations of filters (e.g., weak M and R, rather than a strong R or M) might work well.
I am sure, at the end of the day, that the empirical approach is best. Those who dive often, shoot smart, and critique thoroughly will have the best advice to offer. But empiricism informed by modeling is better yet!
One useful outcome, which I plan to post when reasonably complete and organized, is a set of tables of illumination spectra, reflectance and attenuation models, and filter specifications -- these are data, and data are good (as a modeler, I live on the data collected by others).
And when one can't get wet, one can still ponder underwater spectra...
All the best,
Subal ND70 + 2X DS125 ... mostly 10.5, 16, and 60
Posted 21 February 2005 - 10:35 AM
Thanks for the extra information. Now that I know your background I am much more positive.
I do think that your model will give some very useful guides for filter choice (particularly as conditions change). These are things I notice with my eyes when I am underwater, but do very little to compensate for in terms of camera settings (instead relying on white balance).
It would be great to be able to ask the model - so at midday I am using a UR-Pro filter at 7m in clear blue water. If I am diving in the late afternoon should I add additional gels? Or if the water is a bit green? Or if it is cloudy?
I am really fascinated. Of course it will only be a guide. But that is all we use anyway. Guides that are approximations of the accepted attenuance charts mixed in with experience.
In terms of additional data. I suggest looking up PAR - photosynthetically available radiation in an oceanographic context. PAR is usually measured at 1m below the surface. Usually only light intensity but someone is bound to have measured spectral properties. Again phytoplankton/primary production work is likely to be the most instructive. In addition work with the relatively new FRRF - Fast Repetition Rate Fluorimeter - often has accompanying datasets with detailed light information (to accompany their pigment absorbtion spectra).
The other area you might find useful data are within some of the big oceanographic timeseries. The BATS - Bermuda Atlantic Time Series - may well have some useful data. As might the HOTS - Hawaiian version. Both in good blue tropical ocean water. Dunno what the data availability will be. But most academics I know would happily send you some data for an honest request like this.