The Physics of Sunsets

Thorsten

Active member
Joined
Dec 7, 2013
Messages
785
Reaction score
56
Points
43
I've long been interested in light passing through the atmosphere, particularly in sunsets. Initially this came from implementing low light halfway correctly in real time 3d rendering, but I've quickly discovered that many multiple scattering processes are simply too complex to include in a fast rendering framework, yet they lead to quite beautiful results in nature.

Over the years, I've taken (literally) hundreds of sunset photographs - usually from the same location - and in the last years I've also started to film sunsets on video, and I've used that reference material to teach myself what mechanisms actually drive what we see. Of course the basics are simple - Rayleigh scattering colors the light, Mie scattering creates a halo - but then? There's actually way more to it in the way indirect light plays out for instance, or in the way the view ray (as opposed to the illuminating ray) is affected by scattering processes.

In various conversations I've learned that many science people are simply not aware of some phenomena - for instance that Mie scattering works out in both directions - people recognize the silver lining phenomenon, but are oblivious to the fact that there is dark fringing if one looks the other way. Or that Rayleigh scattering is rather generic - the blue sky of Earth is not a particular property of Earth's atmosphere, but driven by the fact that air molecules are small-sized scattering centers, and any colorless gas atmosphere would also appear blue once optically thin.

So I've stared to organize the material that I have into a series of explanations of what we see in the hope that it helps more people to appreciate also the intricacies of what happens in the sky.

Here's a gallery of the vast differences in appearance sunsets can have - this is what needs to be explained:

tableau.jpg

Please enjoy the first parts of The Physics of Sunsets
 

steph

Well-known member
Joined
Mar 22, 2008
Messages
1,393
Reaction score
713
Points
113
Location
Vendee, France
Well, there is a lot of variety, but one always has to include clouds (or their absence). Also, the airmass matters, and even stuff that is below the horizon, due to reflection etc. I like weather, pretty much any aspect of it, including sunsets. But sunsets are complex exactly due to them happening in all types of weather
 

clipper

Well-known member
Joined
Feb 27, 2018
Messages
256
Reaction score
388
Points
63
Slightly off-topic but this reminded me of your amazing work on atmospheric light scattering in FlightGear - I haven't used it in years now but I distinctly remember how playing around with different ALS sliders combined with various weather conditions was my absolute favorite part of that sim as it could result in some beautiful atmospheric vistas with fog and haze that surpassed the go-to sims at the time in that regard like FSX and XP10 even with layers of relevant add-ons on top.
 

Thorsten

Active member
Joined
Dec 7, 2013
Messages
785
Reaction score
56
Points
43
Well, there is a lot of variety, but one always has to include clouds (or their absence). Also, the airmass matters, and even stuff that is below the horizon, due to reflection etc.

That's kind of the point of the project - to illustrate what matters how.

I found something called precomputed atmospheric scattering.

As you'll figure out quickly when you actually try this - a bare atmosphere is easy to compute (that works with decent accuracy even in real time). The reason is that you can see through the atmosphere, it is optically thin and so single scattering usually gives you most of what you see.

It's the clouds and haze layers which do most of the magic - and bring nearly all of the complications into the equation - not only is their distribution not simply an exponential decay in altitude, but they may be thick and so the single scattering approximation breaks - solving a diffusion equation works better.

Of course you can still use the same basic scattering equations and with enough computational brute force you get results, but this is the part where understanding what happens and introducing just the right simplifications into the problem can give you lot of results.

Anyway - as mentioned above, I have done my share computing this (some examples found here), I even know to apply it from space (see here for an example), but this thread really is not intended about computational frameworks but about some of the more intricate things that happen in reality (and that I don't find so easy to treat in a computational framework).
 

Thorsten

Active member
Joined
Dec 7, 2013
Messages
785
Reaction score
56
Points
43
If anyone is interested, the article is now finished with the discussion of a few exotic phenomena - light scattering on ice crystals for instance, or the question of green sunsets.halo04.jpg
 

Linguofreak

Well-known member
Joined
May 10, 2008
Messages
5,017
Reaction score
1,254
Points
188
Location
Dallas, TX
or the question of green sunsets.

My parents have a story about seeing a green sunset after a tornadic thunderstorm, and I've heard other reports of green skies associated with tornadoes. Some quick checks on Wikipedia seem to indicate this really has more to do with the presence of hail-friendly conditions. I've seen a slight greenish cast to the clouds in severe thunderstorms from time to time, but never anything that I would describe as a green sunset.
 
Top