I saw the post from Gresham and I came a runnin'
Very, very good question, I must say.
First things first, action spectra describe the efficiency of light absorption for given wavelengths, not the efficiency of usage of light once it's been absorbed. This is an important distinction, as we'll see.
When a photosynthetic organism absorbs a photon, that photon can be absorbed either by chl a, or by one of several accessory pigments. Chl a has characteristic absorption peaks in the red and the blue/violet. You can see that here (ignore chl b--that's in green plants and irrelevant to corals):
http://en.wikipedia.org/wiki/Image:Chlorophyll_ab_spectra.png
When chl a absorbs a photon at low light levels, it can put that photon into photochemistry pretty darn efficiently.
However, most of the photons that are absorbed by a photocenter get absorbed by accessory pigments and are transferred to chl a to get put into photochemistry. Dinoflagellates have several accessory pigments, each with their own characteristic action spectra (chl c2, peridinin, diatoxanthin, diadinoxanthin, chlorophyllidae). If you have access, you can see a graphic with the action spectra for all of these in this paper:
http://jeb.biologists.org/cgi/reprint/206/22/4041
When all is said and done, dinoflagellates are able to put ~60% of the light they absorb into photochemistry at lower light levels (sub-saturating) if the are otherwise healthy. That is to say, they have a quantum efficiency of ~0.6, give or take ~0.05.
So, a photon comes in, most likely hits antenna proteins (accessory pigments) on the photocenter, gets transferred down the antenna to chl a, and gets transferred into photochemistry (via the e- transport chain)....or at least it gets all the way through ~60% of the time in lower light with healthy zoox.
Once a photon gets to chl a, however, it doesn't matter one iota what the original wavelength of that photon was. Chl a can only transfer energy to the e- transport chain equivalent to essentially a red photon. Any excess energy is lost as heat or fluorescence. So, if a blue photon gets absorbed (much higher energy than a red photon) and makes it chl a to be tranferred, the exess energy is dumped and the energy equivalent to a red photon is transferred. If a red photon is absorbed and transferred to chl a there isn't any excess energy to dump.
Just to be perfectly clear here, once absorbed, the wavelength of the light has effectively no baring on photosynthesis. 1 mol of blue photons is the same as 1 mol of green photons is the same as 1 mol of red photons once they get to chl a.
BUT, the liklihood of initial absorption for any photon by the pigment/protein complex varies with wavelength. This is what the action spectrum shows us--the relative likelihood that a photon of a given wavelength will be absorbed.
The action spectra obtained from most corals looks most like what you see in Fig. 2d and Fig. 3 of the study you cite (you can see a couple in the study linked above). However, most action spectra for corals is actually much more even than most of these. You’ll tend to see a broad peak over the range of ~400-500 nm and a smaller peak at ~680 nm, but you don’t usually see a big lull in between like that for the T. maxima spectrum. The efficiency of absorption at the peak in the blue might be double what it is in the yellow/green, but that means they’re still absorbing a lot of photons in that region. You can see a fairly typical looking action spectrum (though perhaps slightly weak in the blue for most corals) for Favia zoox. isolated by Len Muscatine in Dana Riddle’s article here:
http://www.advancedaquarist.com/2002/2/aafeature/view
The relevance of the action spectrum to photosynthesis depends on light intensity though. Recall that all photons, regardless of what color they are when absorbed, end up being the same when they get to chl a. If photosynthesis is light saturated, the action spectrum for a given plant becomes somewhat irrelevant. Whether they are getting full spectrum light, all blue light, all green light, all red light, or whatever weird spectrum you like, if photosynthesis is saturated with light, it doesn’t matter. Photons are photons are photons.
BUT, because the likelihood of absorption for a photon is wavelength dependent, it will take different numbers of photons (i.e., different PAR) to achieve saturation. We could use diffraction to get monochromatic light (i.e., what we’d use to produce an action spectrum in the first place). If the relative absorption at, say, 450 nm is 2, at 600 nm is 1, and at 680 nm is 1.5, we could achieve saturating light intensities a X umol photons/m2/s at 450 nm, 2X umol photons/m2/s at 600 nm, and 1.333 umol photons/m2/s at 680 nm. The action spectrum tells us how efficiently the plants are able to absorb incident light, but this is only relevant to the rate of photosynthesis at sub-saturating light intensities. If photosynthesis is saturated, is just doesn’t matter what the spectrum of the light is in terms of the rate of photosynthesis.
Now, if we wanted to design a lighting system to be as efficient as possible in terms of the proportion of incident light that is absorbed by the photosystem, we’d simply use a light source that emits all its light at the wavelength of peak absorbanceâ€"that’ll be somewhere in the range of 400-500 nm for most corals (between violet and cyan). However, in practice we don’t care one iota about how efficiently corals use the incident light, we care how much photosynthesis we can get for a given input of electricity into a bulb. This is where measuring PAR/PPFD becomes very relevant.
There are a number of metal halide bulbs that put out most of their light at ~450 nm, which should be close to the optimum wavelength of absorption for the photosystems. Many 12, 13, 14, 15, 20K, etc. bulbs fall into this category. The problem is that, while they produce most of their light where the photosystems are best at absorbing it, they produce relatively little light overall. Something like most 10K bulbs, and good old fashioned Iwasaki 6500K bulbs produce a much more even spectrum. They have their spikes and such, as all halides do, but nothing like the all-or-nothing spikiness of “blue†halides. Not only do they produce a more even spectrum, they tend to produce a heck of a lot more light (much higher PAR/PPFD).
So now we get into the problem of answering, how much of the incident light is actually absorbed and put into photosynthesis over this entire range. This measure is typically called PUR, photosynthetically usable radiation. In order to calculate the PUR you have a good spectrum from your incident light as well as the reflected light (inverse of the action spectrum). The difference in the two (in other words, incident light â€"œ reflected light = absorbed light) is the PUR…or at least it would be if life were easy. In something like a dense phytoplankton culture where light is extinguished within the culture, or with many types of plants, you can get a good approximation of PUR by subtracting the reflected light spectrum from the incident spectrum to get absorbed light. Corals…corals are more complicated, unfortunately.
First, most corals have one or more fluorescent proteins that have nothing to do with photosynthesis at all. They absorb at one wavelength and reflect at another, which can monkey things up a bit. They also tend to have endolithic algae growing below their tissues in the skeleton. These algae absorb light (hence less light is reflected back) and can give the impression that the corals are absorbing and using more light than they actually are.
If we happen to be so lucky as to have a coral without any fluorescent proteins, or especially endolithic algae, it would be far easier to measure the PUR…at least hypotheticallyâ€"we spent days trying to get things to work well last year in Curacao. BUT (there’s always a but), the action spectrum for a given coral is not constant. It depends on the particular mix of chl a and accessory pigments in the zoox. in the coral, and that can change over the course of hours (somewhat) to weeks (substantially). So even if we do a good job of approximating action spectrum for a given coral (we can use software to calculate the PUR from that action spectrum for any given light source if we have the spectrum and intensity), that action spectrum is not necessarily going to work for that coral (either species or individual) under all conditions.
This number that we’d ideally like to have, the PUR, therefore becomes a heck of a lot of work and difficulty to come by, something we can only produce with fairly expensive equipment (many thousands of dollars), and something that we would have to reassess regularly. For these reasons, we don’t deal with this number under normal circumstances.
The best second available to us is just using straight PAR/PPFD. Again, at saturating light intensity it doesn’t matter what the light spectrum is. At sub-saturating light intensity, blue light is the ideal, but the vast difference in intensity between “blue†halides and “white†halides is so vast that, in practice, “white†halides can usually get us to saturating light intensities pretty darn easily for a given input of electricity whereas “blue†bulbs often yield sub-saturating intensity for that same input of electricity.
When people study something that relies on photosynthesis, typically they do a small pilot study beforehand to determine where photosynthesis becomes light saturated with the light they plan to use (sunlight or artificial). Usually people use something with a spectrum close to sunlight to get close to real-world applicability. When people do P/E curves (photosynthesis vs. irradiance) they usually try to do the same thing. In practice, any differences due to light spectrum (since lights are usually pretty close to sunlight) is usually not big enough to measure.
Hope that helps…I’m off to nurse carpal tunnel
Chris
). 
I'll cut and paste your post in a place I know you'll get the replies you seek 
