The Steady Light Is an Average

Imagine two bright objects in the same night sky. One flickers; the other holds its light. The familiar advice is that stars twinkle and planets don't.

It's a useful clue. It's also an unfinished explanation. Both beams have to come through the same atmosphere. The air hasn't agreed to leave Jupiter alone.

The difference is partly in what arrives together—and what gets averaged before we notice it.

The air changes the delivery

Most ordinary rapid stellar twinkling is not the star turning its output up and down. As NASA's StarChild explanation puts it, moving air bends the incoming light, so the amount reaching us changes. Stars near the horizon tend to twinkle more because we look through a longer atmospheric path.

That's the accessible starting point. But “the air bends the light” can mean two things that are worth pulling apart.

Temperature and density variations give air a varying refractive index. Different portions of an incoming light wave are delayed differently. The result can spoil the image of a star: instead of concentrating the light into a clean spot, the atmosphere distorts the wavefront. Astronomers call the image-quality problem seeing.

Those distortions can also develop into bright and dark structure in the light arriving at the observer. As that illumination pattern moves across an eye or telescope aperture, the total received light fluctuates. That's scintillation: the brightness variation we call twinkling.

In their 2015 study of atmospheric scintillation, James Osborn and colleagues make the distinction explicit. Wavefront curvature needs distance to propagate into an intensity pattern. High-altitude turbulence is therefore particularly important for scintillation. Turbulence near the ground can be very bad for image sharpness while contributing comparatively little to the brightness fluctuations. The details depend on aperture size and observing conditions; this isn't a rule that assigns every air layer one exclusive job.

Here is the surprise: bad seeing and relatively steady received light can coexist. A messy-looking image and a flickering total signal are not the same measurement.

A planet brings a small crowd

Now return to the steadier object.

An ordinary distant star is effectively a point source for this naked-eye comparison. A nearby planet presents an illuminated disk, or a crescent: light comes from a wider spread of directions. That remains true even when the unaided eye sees something star-like rather than a little resolved world. NASA's Preston Dyches explains this distinction in a skywatching guide: the planet's more spread-out source makes it appear more constant in brightness.

The disk's contributions don't all have to fluctuate together. Their sum can be steadier than the contribution from one tiny angular source. The Blanton–Chase Physics FAQ, hosted at UC Riverside, gives this averaging explanation too. The atmosphere still perturbs the light. The planet hasn't escaped it; we've combined a less synchronized set of fluctuations.

There is a trap in the “small crowd” picture. It does not mean every point on a planet gets its own independent atmosphere. Nearby directions can have correlated fluctuations. A disk is not a magic switch from twinkling to no twinkling, and its mere possession of nonzero angular size doesn't settle how much averaging occurs.

A toy calculation makes the qualification tangible. Suppose four equal contributors each fluctuate with a root-mean-square size of 10 percent of their own mean brightness. If their fluctuations are independent, their average fluctuates at 5 percent of its mean. If all four rise and fall perfectly together, the average still fluctuates at 10 percent. More contributors help only to the extent that their fluctuations differ.

Those numbers are arithmetic, not measurements of Jupiter or a claim that its disk contains four independent optical patches. I checked them by summing the variances and covariances of the four contributions. The point is the contrast: a sum can hide variation without any contributor becoming immune to it.

Nor is reflected light the explanation. The useful difference here is apparent source size, not whether the photons began at the object or bounced off it.

Three ways to make a quieter sum

The planetary disk averages over source directions. A telescope can also average over collecting area, and an exposure can average over time. These are different operations.

Osborn's paper gives a scale that helps make the second one less abstract. For light with a wavelength of 500 nanometres, after propagating 10 kilometres from a turbulent layer, its characteristic intensity-pattern scale is about seven centimetres. The relation is the square root of propagation distance times wavelength; inserting those values gives roughly 7.1 centimetres.

That's not the size of the star's image or the diameter of an air eddy. It's a scale of the changing illumination reaching the observer. An aperture large enough to collect across many such spatial variations can receive a steadier total than one sampling a small part. It still doesn't follow that the resulting stellar image is sharp.

Waiting can help too, but only once the exposure spans enough of the pattern's motion. Osborn and colleagues show a specific on-sky example: a magnitude-eight star observed with the one-metre Jacobus Kapteyn Telescope on La Palma on 21 July 2013. They infer the scintillation component from the total photometric noise after accounting for estimated shot noise. That component stays roughly level at the shortest exposures, then begins to fall beyond a transition around five hundredths of a second. That's a result for those observations, not a universal timing rule for eyes or cameras.

There is another useful catch. In the paper's long-exposure regime, fractional scintillation noise and fractional photon-counting noise both fall as the inverse square root of exposure time. Waiting four times as long halves each, under those assumptions. The total measurement gets quieter, but scintillation hasn't become smaller relative to photon-counting noise merely because we waited.

I like this distinction because “averaging helps” is true without being the whole answer. We need to say what we're adding together, how similarly it varies, and what comparison we're trying to improve.

Keep the clue; lose the certainty

“Stars twinkle; planets shine steadily” remains a handy first clue, not an infallible identification test. The sources support a tendency, not a numerical boundary that identifies any object in any sky. This isn't a report of my own observing session, and the seven-centimetre calculation doesn't predict tonight's twinkling.

What changed for me is the meaning of the steady light. It isn't evidence that nothing is happening along its path. It can be what remains after the disturbances have been added together.

The next question I'd want to answer is where that averaging stops working: how do a source's angular size and the altitude of the turbulent layers determine whether its fluctuations mostly move together? “A planet is bigger” opens the door. It doesn't finish the physics.