# Why do sound waves hang together when traveling at different speeds? (i.e., being refracted)

I've heard that if a cold air layer exists above a warm one, sound may be refracted upward, causing it to miss hearers that otherwise could hear the sound.

My understanding (which may not be correct) is that as the wave leaves the source, the top of the wave front encounters the cold air and starts slowing down. This drags the faster part of the wave toward itself (since it's going more slowly), eventually lifting the wave off the ground.

If this is so, I'm confused about why the wave hangs together. Why doesn't the bottom half of the wave front simply split away from the slower portion, and continue at normal speed and direction, reaching the hearers on the ground like normal? What "glue" holds the wave together, allowing the slower portion of the wave front to drag the faster portion upward?

Any help is appreciated.

every single point on the wave front of a propagating sound wave is behaving as if were a tiny point source of sound radiation, radiating into the hemispherical volume of space right ahead of it.

if the propagation speed of this radiation slows down on one end of the wave front, the hemispherical waves being propagated at normal speed nearby outrun their neighbors in the slow zone, and those little hemispherical waves then have a chance to spread slightly sideways towards the slow zone, ahead of the slower waves.

So it is not that the slow waves somehow "drag" the faster ones in toward them.

• So is it correct to say that the "points" along the wave tend to travel toward the area of greatest/easiest expansion? I'm picturing a long line of cars driving in tandem. When some of the cars on one part of the line hit mud and slow down, the cars on the faster part of the road turn that way, simply because they can expand/turn in that direction more easily than in the still-crowded fast portion of road. Feb 12, 2019 at 15:23
• that's a good way to think of it. Feb 12, 2019 at 18:48
• I found this picture today in the Wikipedia article on the Huygens-Fresnel principle. Is it stating something slightly different? It seems the wave which hits the slower medium first propagates farthest into the medium, before the rest of the wave gets there? wikiwand.com/en/Huygens%E2%80%93Fresnel_principle Feb 13, 2019 at 21:43

When we have cold air above hot air it is not a abrupt change but a gradual one. As such you can not expect a breaking apart of the waves without a discontinuity in the underlying medium.

That being said even where there is an abrupt change there will be boundary effects. A wave going parallel to a reef edge will effectively split apart but there will be trailing and leading ripples across the boundary due to diffraction effects (ie those mentioned by niels, that each point on a wave is a source of secondary waves).