Calculate ambient light level?

November 22, 2005 9:19 AM

How can I calculate the ambient (natural) light level for a particular location and time?

The U.S. Naval observatory can tell me when the sun will rise and set for any particular lat-long.

But how can I find the light level, including any contribution of reflected moonlight, assuming no clouds, and "normal" weather both on the Earth and Sun, for any arbitrary date and location? (By location, I mean lat-long and height above ground, assuming "average" or flat terrain.)

posted by orthogonality to science & nature (9 comments total)

not a direct answer, sorry, but in astronomy there's a thing called "limiting magnitude" which is the faintest start visible (star brightness is measured in magnitudes for stupid historical reasons). this varies with the phase of the moon and is a measure of the scattered light from the sky.

so it won't get you absolute brightness, but it does indicate how the brightness varies over the night and from one night to the next. and calculators are available that will calculate it for any lat/long/date, because astronomers need this information to plan observations.

so for a long shot, googling for "limiting magnitude calculator" will give you a start. and note that magnitudes are a logarithmic measure, so you need to correct for that.
posted by andrew cooke at 9:25 AM on November 22, 2005

Not a direct answer - but one of the features built into some architectural rendering programs (such as Sketchup) is the ability to display the correct light level and direction for a particular time and location on the planet. Stellarium will also do this in an astronomical contect (and is free)
posted by rongorongo at 9:34 AM on November 22, 2005

The Clear Sky Clock takes moonlight into account in its darkness value, and presents a graphical representation of this limiting visual magnitude, which also assumes no clouds (these are taken care of by other predictions made by the Clock). But it's not arbitrary. It predicts conditions for the next day or so for a specific (but growing) set of locations.
posted by Songdog at 9:39 AM on November 22, 2005

Here's (pdf) a paper by Klassen on modeling atmospheric light. Not sure if you'll be able to access it though.

This page has lots of links to information on rendering the atmosphere for computer graphics. Doesn't look trivial.

Tomoyuki Nishita has also done a lot of work on modeling atmospheric effects.
posted by driveler at 9:39 AM on November 22, 2005

Ecotect will do it, but there's a price and a learning curve and the site is down for maintenance, so no cigar, I guess.
posted by signal at 9:46 AM on November 22, 2005

Agreed. Great excuse to buy sketchup! That software is freaking amazing.
posted by zpousman at 11:18 AM on November 22, 2005

actually, i guess limiting magnitude will give you absolute flux if you do the maths. but it's nasty (complex details, easy to get wrong, but not "hard") maths.
posted by andrew cooke at 11:58 AM on November 22, 2005

This has something close to what you want. See Eq. 1 in the attached paper. It gives the luminance at a point in the sky for a given sun position. However, to get a value for the entire sky, you might have to integrate the equation over a hemisphere.

It seems like things can get really tricky depending on how realistic you want your answer to be. Check out this recent paper. The pictures look great: they can model things like rainbows, halos, and sunrise and sunset effects, but the model has about two dozen parameters.
posted by driveler at 12:21 PM on November 22, 2005

AFAIK, Shetchup will calculate sun position (and maybe relative brightness?), but not actual light level, expressed in Lux or whatever, which seems to be what the question calls for.
I would gladly be proven wrong, though.
posted by signal at 12:41 PM on November 22, 2005