Why do sounds carry farther at night?
― It's not just because it's quiet
At night, you can sometimes hear a distant train or motorway that's normally silent. It's tempting to think "night is quiet," but that's not the whole story. At night, the very path sound takes changes, trapping it near the ground. Sound that escapes upward during the day bends back down at night. And we can actually calculate that bending.
Open a window late at night and you may hear a sound that should be far too distant to reach you: a level crossing, a main road, a factory, distant fireworks.
But open the same window the next afternoon, and that sound is gone. The source is making the same noise both day and night — yet you can't hear it.
"It's noisier during the day, so it gets drowned out" — true, to a point. But that alone doesn't explain it.
When you actually measure it, the distant sound itself arrives louder at night. It isn't being drowned out — it simply isn't reaching you at all during the day.
Sound travels faster in warm air and slower in cold air. The difference is about 0.6 m/s per degree Celsius.
Waves bend toward the slower side. So which is warmer, above or below decides where the sound ends up.
And day and night have the temperature layers flipped. That's the whole story.
The speed of sound is set by air temperature
Sound is a vibration travelling through air. The warmer the air, the faster its molecules move, and the faster the vibration travels.
The difference is surprisingly small: about 0.6 m/s per degree Celsius. Between 20°C and 5°C, that's only about 9 m/s.
Yet this tiny difference is enough to change the direction. That's the key point.
if it decides which side is faster, changes where the sound ends up.
Waves bend toward the slower side
Picture spreading sound as a bundle of many thin paths. What happens if the speed differs between the upper and lower parts of that bundle?
Think of a marching line. If the people on the right walk faster and those on the left walk slower, the whole line turns left. The faster side gets ahead, so the line bends toward the slower side.
The same thing happens with sound. If the upper air is faster, sound bends downward. If the upper air is slower, sound bends upward. That's the whole rule.
Sunlight warms the ground, so the ground becomes the faster side. Sound bends upward and never reaches distant ground.
Radiative cooling chills the ground, so the air above becomes the faster side instead. Sound bends downward and returns to the ground.
At night, sound travels by bouncing back and forth between ground and sky, again and again. Since it never escapes upward, it barely weakens. That's the real reason you hear distant sounds at night.
Of course, the quiet of night also helps. But quietness only changes how easy a sound is to pick out. Whether the sound reaches you at all is decided by this bending.
Near the ground, friction slows the wind down. So wind is normally stronger higher up.
When sound travels downwind, the upper part gets pushed faster by the wind, so it bends downward. Travelling upwind, the upper part is pushed back harder, so sound bends upward and escapes.
That's why you hear well downwind and poorly upwind. People say "the sound is carried on the wind," but what's really at work is not carrying, but bending. Wind speed is far slower than the speed of sound, so the carrying effect itself is tiny.
So the mechanism is the same whether it's temperature or wind. All that matters is "which side, up or down, is faster."
This article's mechanism works both ways. Sounds you make at night also carry further than during the day.
A car door closing, an engine, voices at the front door, work on the balcony. Sound that would escape upward during the day spreads out hugging the ground at night. The feeling that "it wasn't that loud" is built from daytime experience.
Night-time noise tends to become a problem not just because it stands out in the quiet. It's also because, physically, it reaches much further.
Something you can check for yourself
- Pick the same spot every time (a window, a balcony, etc.)
- Choose one distant sound that's always present: a main road, a level crossing, a train, a factory
- Listen from the same spot for one minute, once in the day and once late at night, and note whether you can hear it
- Repeat over several days, comparing a clear, calm night with a cloudy or windy one
- Clear nights should carry sound better. Clouds stop the ground cooling as much, weakening the reversal.
Steps 4 and 5 are the heart of this observation. If you can tell the difference between "because it's night" and "because it's a clear, calm night," you'll have confirmed for yourself that it's the state of the air, not the quiet, that matters. Note the wind direction too and you should notice you hear it better when downwind.
Summary
Distant sound reaches you at night because the ground cools and the air above stays warm. Sound bends toward the slower side, so on a night when the upper air is faster, sound bends downward and returns to the ground. By day it's the reverse, and sound escapes upward.
It's not that night is quiet that lets you hear it,
it's that the sound has nowhere to escape to, and comes back.
The same mechanism — a wave bending toward the slower side where speed changes gradually — is also part of why whale song can travel hundreds of km underwater. See this article for more.
Want to know more? ― Terms, numbers, and how this connects to textbooksFrom middle-school science to topics still under active research — each level is labelled
- MSCovered in middle-school science
- HSCovered in high-school "Physics Basics"
- HS+High-school "Physics," or advanced/sidebar content in textbooks
- Univ.Not taught in high school — content from a university specialist course (atmospheric acoustics)
- ResearchNot yet settled even at university level — something researchers are actively investigating
MSTerms: sound and air
- Speed of sound: how fast sound travels through air. Changes with air temperature.
- Refraction: a change in direction of travel caused by a difference in speed from place to place. Happens with both light and sound.
- Radiative cooling: the ground losing heat to the sky and cooling down. Strongest on clear, calm nights.
- Surface temperature inversion: a state where only the layer near the ground is cold, and it gets warmer higher up. This is the "reversal" in the main text.
- Sound shadow: a place sound can't reach because it has bent away.
HSWorking it out with a formula: what radius does sound bend on?
The main text said sound "bends and comes back." We can actually calculate how much it bends. All we need is the relationship between the speed of sound and temperature.
v = 331.5 + 0.6 × t
| v speed of sound | units: m/s |
| t air temperature | units: °C |
| 0.6 | increase per °C [m/s] |
This is a commonly used approximation. Since it increases in a straight line with temperature, once you know the temperature above and below, you immediately get the speed of sound above and below.
For example, at 20°C: 0.6 × 20 = 12, so 331.5 + 12 = 343.5 m/s. The familiar "about 340 m/s" figure is roughly this temperature.
On a night with radiative cooling, the ground ends up colder than the air above. Let's take the ground at 5°C and 10°C at a height of 100 m.
| Speed of sound at ground (5°C) | 0.6 × 5 = 3, so 331.5 + 3 = 334.5 m/s |
| Speed of sound at 100 m (10°C) | 0.6 × 10 = 6, so 331.5 + 6 = 337.5 m/s |
| Difference | 337.5 − 334.5 = 3 m/s |
| Change per metre | 3 ÷ 100 = 0.03 m/s |
Just 0.03 m/s faster for every metre you rise. Against a speed of sound of 334.5, that's less than one part in ten thousand. Can that really bend anything?, you might wonder.
The radius of the circle traced by the sound's path is found by dividing the speed of sound by the change per metre.
R = v ÷ (change in speed of sound per metre)
| Substituting | 334.5 ÷ 0.03 ≒ 11150 m |
| In km | about 11 km |
A gentle arc with an 11 km radius. It seems "almost dead straight." But as we'll see next, that was enough.
Let's follow a sound that leaves the source just one degree above horizontal. Without bending, it should keep rising forever and vanish.
| Height risen = R × (1 − cos1°) | 1 − 0.99985 = 0.00015 |
| Substituting | 11150 × 0.00015 ≒ 1.7 m |
| Maximum height reached | just 1.7 m |
Barely higher than a person's height. You'd expect even one degree upward to escape into the sky, but the 11 km arc bends it back before it ever climbs that far.
We can also work out the horizontal distance before it returns.
| One-way horizontal distance = R × sin1° | 11150 × 0.01745 ≒ 195 m |
| Round trip back to the ground | 195 × 2 = 390 m |
Roughly every 400 m, it returns to the ground, bounces, and repeats. Since it never escapes upward, it barely weakens even as the distance grows. That's what "carries far" really means.
A speed difference of less than one part in ten thousand is enough to tie sound to the ground. A tiny difference, once it decides a direction, can produce a huge result. That's what this article set out to show.
| Speed of sound at ground (20°C) | 0.6 × 20 = 12, so 331.5 + 12 = 343.5 m/s |
| Speed of sound at 100 m (15°C) | 0.6 × 15 = 9, so 331.5 + 9 = 340.5 m/s |
| Difference | 343.5 − 340.5 = 3 m/s (opposite direction from night) |
The size is the same 3 m/s as at night, but which side, up or down, is faster has swapped. So sound bends upward, and with the same 11 km radius moves away from the ground.
The result is a "sound shadow" beyond a certain distance from the source. It's not just that daytime is too noisy to hear distant sounds — they simply aren't arriving at all.
※ Real temperature profiles aren't perfectly linear — the rate of change varies with height. Ground reflection and atmospheric turbulence also play a role. The calculations here are meant to capture the mechanism and the rough scale.
HS+Exactly the same law as the refraction of light
The bending we used in ③ comes from the same law that bends light entering glass. A single principle — "where speed differs by place, a wave bends toward the slower side" — explains both light and sound (the light side is covered in our article on rainbows).
The difference is how the boundary behaves. At a glass-air boundary, the speed changes abruptly, so the path bends sharply like a kink. The atmosphere's temperature, by contrast, changes gradually, so the sound's path traces a smooth arc. A sharp kink or a smooth curve — that's the only difference.
There's also the fact that higher-pitched sounds are absorbed by air more easily. The muffled, low-pitched quality of distant sounds comes from losing their higher components along the way. This is thought to be one reason the distant road noise you hear at night is a low "roar."
Univ.Heard hundreds of km away, but not in between
Apply the same refraction on a far larger scale and something strange happens. Atmospheric temperature rises and falls repeatedly with altitude. As a result, there are several layers high in the atmosphere that bend sound back down.
With large explosions or volcanic eruptions, sound can rise tens of km, get bent back, and land far away. The result can be a phenomenon where nothing is heard nearby, yet windows rattle hundreds of km away. There are known cases of historic major eruptions where only distant records of the sound survive.
This long-range propagation is also used to monitor nuclear tests. Infrasound — pitched too low for humans to hear — travels on a global scale. International monitoring networks operate on this very principle.
If you think of sound as something that "travels in a straight line and simply weakens with distance," this phenomenon makes no sense. The key is that the path itself bends.
Research"How far will it carry tonight?" still can't be answered precisely
- Atmospheric turbulence makes prediction hard. Real air isn't uniform — eddies large and small are constantly moving. Sound scatters off them, so even with the same temperature profile, what you actually hear varies moment to moment. Sound is known to leak into the "sound shadow" through this turbulence.
- This effect is debated in the context of wind-turbine noise. Because sound travels far through a night-time inversion layer, predictions based on daytime measurements can fail to match what people actually experience. How to properly assess this is still under discussion.
- The evaluation of how low and infrasonic sound affects people is not settled. Perceived loudness and the discomfort felt in the body don't always match, and the very way of measuring it is a subject of debate.
- Differences due to ground surface aren't simple either. Reflection behaves differently over grass, snow, and asphalt, and in particular snow is known to absorb sound very well. "It's quiet when it snows" has causes beyond just fewer sound sources from reduced activity.
The basic mechanism of how sound bends was understood in the 19th century. Even so, precisely predicting "how far that sound will carry tonight" still can't be done. Understanding the principle and being able to predict it are, after all, two different things.
How this connects to textbooks (by level)
| Level | Subject / unit | Where in this article |
|---|---|---|
| MS | Science: properties of sound / weather | Sound travels through air; radiative cooling |
| HS | Physics Basics: waves and sound / speed of sound | Calculation using v = 331.5 + 0.6t |
| HS | Physics Basics: reflection and refraction of waves | The property of bending toward the slower side |
| HS+ | Physics: wave propagation (Huygens' principle) | The smoothly curving path, the radius calculation |
| Univ. | Atmospheric acoustics / acoustic engineering | Long-range propagation, bending back high in the atmosphere |
| Research | Environmental noise / atmospheric acoustics (unresolved) | Turbulent scattering, night-time noise assessment |
| ― | Everyday life | Your own night-time noise carries further than you think |
- Explanatory material on outdoor sound propagation from the Acoustical Society of Japan (日本音響学会).
- Salomons, E. M., Computational Atmospheric Acoustics (a standard textbook on how sound travels through the atmosphere).
- Attenborough, K. et al., a series of studies on prediction models for outdoor sound propagation.
- Explanatory material on the infrasound monitoring network from the Comprehensive Nuclear-Test-Ban Treaty Organization (CTBTO).
- Material on noise measurement and assessment from Japan's Ministry of the Environment (環境省).
※ The speed-of-sound formula is an approximation and also varies slightly with humidity. Temperature profiles and ground properties vary widely from place to place.
※ This article is a general-audience science explainer. The figures given are approximations meant to illustrate the mechanism; actual sound propagation varies greatly with weather conditions, terrain, and ground surface. For specific judgements or measurements relating to noise, please consult your local environmental authority or a specialist body.