"Energy loss through an absorbing/scattering medium goes like so...

(I/Io)^(z/zo)

...where I, Io are the intensity and the initial intensity respectively, and z, zo are the path-length and initial path-length. Thus if 50 km attenuates the intensity to 75%, 16 times that reduces it to just 1%. "

But obviously this is wrong.
If the sunlight is 99% reduced, I wouldn't have any problem with the sun's glare driving home.
And light level in general would be significantly different from noon to late afternoon.

In addition our atmosphere varies with elevation- about every 5 km [or something like that] it's density halves.

Quote:
"50% of the total mass of the atmosphere is located in the lower 5km of the troposphere."
http://en.wikipedia.org/wiki/Earth's_atmosphere
Oh, later in article this:
"Atmospheric pressure decreases with height, dropping by 50% at an altitude of about 5.6 km (18,000 ft). Equivalently, about 50% of the total atmospheric mass is within the lowest 5.6 km. This pressure drop is approximately exponential, so that pressure decreases by approximately half every 5.6 km."

If you are correct this would mean that at 18,000 elevation instead of receiving about 1000 watts at noon, one would receive around 1200 watts, and loses you think are due the sun nearer the horizon would be halved. I don't think it's that significant of a change, though I would say there is an significant improvement with higher elevation.

Now, moving a slightly different tack.

Here is graph showing US solar energy in Feb with tracking:
http://rredc.nrel.gov/solar/old_data/nsrdb/redbook/atlas/serve.cgi

It shows a small are of the US getting 10+ kwhr per day

Going to Oct we get same area [plus more area] getting over 10+ kwhr per day:
http://rredc.nrel.gov/solar/old_data/nsrdb/redbook/atlas/serve.cgi

So roughly between Feb thru to Oct you are getting more than 10+ kwhr.
Nov, Dec, and Jan you receiving around 7 Kwhr.
Or if you just look at Annual:
http://rredc.nrel.gov/solar/old_data/nsrdb/redbook/atlas/serve.cgi
It show a large area that gets between 10-14 kwhr per day.

So if average was 250 watts [250 watts times 24] you would only have 6 kwhr per day. Instead what graph show is that in some large region of the US one could on average get about twice this amount. This this is allowing for weather, ect.