On 08/02/2009 05:32 PM, Kelly Jones wrote:
> In sun.frink, I noticed that sunset[] doesn't take elevation as a paramater.

(For reference, this is in regards to Frink's high-accuracy
sun/moon/astronomical calculation library, located at:
http://futureboy.us/frinksamp/sun.frink )

> However, elevation does affect sunrise/set time: the greater your
> elevation, the longer the sun (or any other object) stays up.
>
> I normally wouldn't complain, but if you're going to take pressure and
> temperature into account, elevation should be included too?

Yes, elevation affects sunrise and sunset time. However, the
tallest things between you and the horizon do too! So it's not enough
just to know your elevation. The effects of elevation are different if
you're on a plane or on a plain! If you're sitting on a airplane a mile
above ground, then sunset will be later by about a minute. If you're
sitting on a plain (like in Kansas or Kazakhstan or Kathmandu) a mile
above sea level, then sunset time will be hardly affected because
everything on the horizon around you is also higher!

So height-above-ground would be a much more useful measure than
elevation, but to calculate when the sun peeps over *your* local
horizon, you have to take into account everything between you and the
horizon. To do it right, you'd need to consult digital elevation maps
or the like, and not only enter elevation (i.e. distance from the geoid)
but also height-above-ground.

As Jean Meeus says in his book Astronomical Algorithms, predicting
the exact time of sunset/sunrise to more than a minute for any
particular day probably doesn't make much sense, due to refraction.
None of his algorithms correct for elevation or height-above-ground.

If you'd like to contribute a modified algorithm with citations, I'd
be happy to include it!

--
Alan Eliasen
eliasen@...
http://futureboy.us/