Wow, great discussion going on here!
I have a couple things I'd like to add.
Light differs from sound. Sound is a wave that effects matter. When two sound waves hit 1 particle, both waves will effect it.
Light moves through space. When 2 lightwaves hit each other, they do not have the same "added affect" as a sound wave.
So, light pressure is different than sound.
Ok , now we've all heard this before: "Lumens do not add up", "Watts do not add up" or "Lights do not add up".
However, if you center a lux meter at a set point and turn lights on one at a time around it, the reading will rise.
Here's where it gets interesting:
If you take a measurement from 1000 watts of CFL at say 6", you'll get a high reading. You'll keep this high reading all throughout the "coverage area" that the lights are covering.
If you do this with a 1000w HPS, you'll get a much higher reading, but as you move from side-to-side along the "coverage area", It will drop.
(Since in gardening, we generally work on a flat plane, the "coverage area" will be regarded as a flat plane.)
So, we establish that a single stronger source of light will produce a higher lumen at a set distance, but an uneven pattern that tapers from this single source. Many smaller sources of light will not ever produce a higher measurement than the single source, but they will evenly blanket an area with light.
Now, let's move the other way, up and down.
Let's say you take an HPS bulb rated at 160000 lumens. At one foot, you'll read about 160k lumens. At two feet, you'll read a 1/4 of that. This is called the inverse square law. Because light radiates evenly, at double the distance, you quadruple the area that the light is now covering. (Imagine the surface area of a balloon that's 1/2way blown up vs. one that's completely blown up - Light spreads in the same way the balloon does)
So, (for our example), at 2' you have about 40k lumens. At 3', You'll have 10k, 4' = 2500 lumens...
Ok, so that's penetration. At 4' away from the light, you still have a usable amount of light (2500 lumens) that will support growth.
Now, let's try that with many smaller points of light.
Let's say you use 42 w CFLs rated at 2700 lumens.
So at 1', let's say you have so many, so close that you are recieving a total measurement of 10000 lumens - not bad right? Especially since you're getting the same amount of light that you' be getting at 3' with the HPS.
Now here's the problem: Each of these smaller sources are suseptible to their own inverse square law.
So, You can't just divide the 10000 by 4 to get 2500 at 2'. You would need to take each of the 2700 lumen sources and start quartering those. So now, you've gone from many sources adding 2700 or less to the measurement, to many sources adding 700 or less. Bulbs that are further away, now have little to no effect and the ones that are close, can't get their light far enough to really bring that number up.
Whether the source is from an HID, CFL, or LED, this rule holds true.
If a 200w CFL has a higher initial lumen than a 150w HPS, the CFL will have better penetration than the HPS.
if you want proof, google "inverse square law". This explains it all...