REDSHIFT REPORT Subscribe

Exoplanets

Webb Traced the Outline of an Unseen Planet's Air. It Isn't Round.

Using the edges of a transit, astronomers mapped the ring of atmosphere around WASP-17b. Its evening side is swollen and roughly 1,170°F hotter than its morning side.

Oct 30, 2026 · 5 min read · Episode 5

Artist's concept of the hot gas giant WASP-17b, with a thin glowing atmosphere along its edge
Artist's concept of the hot gas giant WASP-17b, with a thin glowing atmosphere along its edge. Credit: Artwork: NASA, ESA, CSA, Ralf Crawford (STScI)

Nobody has ever seen the planet WASP-17b, not even as a single dot of light. Yet astronomers have now measured the shape of the thin shell of air around its edge, and found that it is lopsided. One side is puffed up and, by the best estimate, about 1,170°F (650°C) hotter than the other.

The result comes from the James Webb Space Telescope and a new way of reading a planet's shadow. It is the first time the technique has worked on real data.

A planet built to be studied

WASP-17b is a gas giant about 1,300 light-years away. It has roughly half the mass of Jupiter but is nearly twice as wide, which makes it one of the least dense planets known. On average, its material is less dense than cork.

It circles its star in under four days, and it does so backwards, against the direction the star spins. At that distance the planet is almost certainly tidally locked: one hemisphere sits in permanent daylight, the other in permanent night.

That bloated atmosphere is what makes it so useful. A puffy atmosphere leaves a bigger mark on starlight, and in 2023 Webb used this planet to detect clouds made of tiny quartz crystals.

What they found

The team, led by Lili Alderson of Cornell University, split the planet's ring of atmosphere into different points around its edge and measured each one.

  • The evening edge, where air arrives after crossing the whole day side, came out at about 1,522 K, or roughly 2,280°F (1,250°C).
  • The morning edge, where air arrives after a long trip through the night, came out at about 858 K, or roughly 1,085°F (585°C).
  • The evening side's atmosphere is stretched out about 1.8 times as far as the morning side's. That matches the temperature ratio, since hotter gas expands in proportion to its temperature.
  • Water and carbon dioxide look about the same all the way around the edge.

So the two sides differ in heat and in size, not in what the air is made of.

The likely cause is wind. Hot Jupiters are expected to have a broad jet stream around the equator that carries heat from the day side toward the evening edge. The air then loses heat on the night side and reaches the morning edge cold. The authors note that their map agrees with earlier Webb measurements of this planet's day and night sides.

How they know

When a planet passes in front of its star, it blocks a small slice of the light. For WASP-17b, the dip is about 1.5 percent. Some starlight also filters through the planet's atmosphere on its way to us. Gases absorb particular colors, so the planet looks slightly larger at some wavelengths than at others. Plot that across many colors and you get a transmission spectrum, a chemical fingerprint of the air.

The standard approach treats the planet as a perfect circle and blends the whole ring of atmosphere into one fingerprint.

The new method, called transmission strings, does not. It was developed by David Grant and Hannah Wakeford at the University of Bristol, who are also authors on this study. It uses the very start and the very end of a transit. As the planet slides onto the star, one edge touches first. As it slides off, the other edge leaves last. The precise shape of the light curve in those moments depends on the shape of the planet's outline.

Picture the outline as a loop of string that can bulge out in some places and pull in at others. The method finds the loop that best fits the data, and it does this separately at each wavelength. The result is not one fingerprint for the whole planet, but one for each part of its edge.

The team combined three of Webb's instruments to cover wavelengths from 0.6 to 12 microns, from the edge of visible red light far into the infrared.

Why it matters

There are now more than 6,000 known planets around other stars, and most of what we know about their atmospheres comes from blended, ring-averaged spectra.

If the two edges of a planet are very different, that blend can mislead. A hot, swollen edge leaves a much larger mark in the light than a cool, thin one, so it can dominate the average. Read the blend as if it came from one uniform atmosphere and you can get the temperature wrong. You can also misjudge how much of a gas is present, because a stronger signal may come from hotter, taller air rather than from more of the gas.

Alderson has described the result as proof of concept that planets are three-dimensional objects. Measuring how an atmosphere changes around its edge is a first step toward mapping weather on worlds we cannot photograph. WASP-17b is now the test case for doing that.

What is still uncertain

  • The temperatures have wide error bars. The morning edge is cooler and thinner, so its signal is weaker. The paper gives the morning value as 858 K with an uncertainty of about +434/−233 K. The true difference between the two sides could be noticeably smaller or larger than 650 K.
  • North and south were assumed to match. The team took the planet's northern and southern halves to be mirror images. Models suggest that, but it was not measured.
  • Clouds were not mapped. WASP-17b has quartz clouds, but this analysis could not pin down how clouds or haze change around the edge.
  • This is an unusually easy planet. A bloated atmosphere gives a strong signal. Whether the method works on smaller, denser planets, including rocky ones, is still being tested.

The paper

"A spatial map of the atmospheric limb of a hot Jupiter" by Lili Alderson, Hannah R. Wakeford, David Grant and colleagues was published in Nature Astronomy on 14 October 2026. It uses Webb transit observations from three instruments (NIRISS, NIRSpec and MIRI) covering 0.6 to 12 microns. A free version is on arXiv. The video walks through the method step by step.

More from Redshift