Between the fall of 2012 and the fall of 2013, astronomers took a deep, long look into the atmosphere of GJ 1214b, a distant Earth-sized planet in a tight orbit around a cool, red star. Fifteen times it crossed the face of its star as the Hubble Space Telescope watched. Each time, starlight would pass through the thin ring of atmosphere around the planet. Some of it would catch on airborne molecules—methane, carbon dioxide, or maybe even water vapor—leaving a spectral fingerprint.

At least that’s what the astronomers hoped. It had worked for bigger, hotter planets, and now we’d be able to understand the atmosphere of one of the most earth-like exoplanets known. Going forward, we would use the same method to look at ever more familiar worlds, sniffing for chemicals that might have been exhaled by living organisms. With this trial run on GJ 1214 b, however, there was just one problem.

We saw nothing.

Or, to say it in a kinder way, we saw something else amazing: a world enshrouded in alien clouds. The spectrum, painstakingly measured from light filtered through the planet’s atmosphere, didn’t show the telltale spikes and wiggles of molecules. Instead, it was featureless and flat.

In a paper published in Nature, astronomers inferred that an opaque layer of clouds had blocked Hubble’s view. Suddenly, small planets around other stars seemed lot harder to probe, even for the world’s most powerful observatory. “They got a tremendous amount of Hubble Space Telescope time just to study that one planet in exquisite detail,” says Jonathan Fortney, an astronomer at the University of California, Santa Cruz. “And that spectrum came back flat as a line.”

At that point, you could still hold out hope. “It was still just one planet. We weren’t yet at the point where we had a couple others, but it really shows the seriousness of the situation,” he says.

Within the year, though, one planet became several. In 2014, three more small worlds with their own cumbersome names—HD 97658b, GJ 436b, and  GJ 3470b—turned out featureless and inscrutable. To this day, only a handful of molecules have been tentatively detected in the atmospheres above alien worlds smaller than Neptune.

Clouds, as it turns out, are everywhere.

You can find them in every atmosphere in our own solar system, even Pluto, which perhaps should have given us a hint. In endless permutations, they condense out of atmospheric chemistry, or cook up under the glare of the Sun. But in the puffy atmospheres of so-called “super-Earths”—which are common around other stars but absent from our solar system—astronomers were hoping for clearer skies.

“This is a double edged sword,” says MIT’s Zach Berta-Thompson, who is currently trying to use Hubble to study the newly discovered GJ 1132 b, a small, rocky world thought to resemble Venus. “It’s really exciting to know that we have clouds that we didn’t expect in some of these planetary atmospheres,” he says. “But on the other hand it’s really frustrating, because as astronomers, clouds are our natural born enemies—whether those clouds are around Earth or another planet.”

Berta-Thompson’s MIT colleague Sara Seager puts it more bluntly. Right now, Seager is working on building up a library of atmospheric signatures that might herald the presence of life on faraway planets. She hopes to find not just one but a handful of possibly inhabited worlds in the next few decades—a hard task made even harder if we keep seeing cloudy planets like GJ 1214 b.

“That would be a disaster for everything we’re working towards,” Seager says. “If every planet has clouds to the level that one does, it might kill all our dreams.”

* * *

All is not lost, though. A generation of astronomers are pioneering methods to see around and through clouds—or to avoid them all together.

At the front lines is atmospheric modeler Caroline Morley, who is finishing up her Ph.D. in Fortney’s lab in Santa Cruz before moving on to Harvard in the fall. One of the first to explore the chemical compounds that could be cloaking the infamous GJ 1214 b, Morley recently published a paper that reads like a playbook for fighting the cloud problem.

“We need to be a little bit more creative in the observations we do,” Morley says. For one, we can study planets at times outside of when they are passing in front of their stars. Although looking at starlight shining through the ring of atmosphere is a powerful tool, our line of sight takes us the long way through sheets of clouds.

The idea is that in other phases of the planet’s orbit, as it moves beside its star and then behind it, we can pick out starlight that reflects off the planet’s atmosphere. Or perhaps we could even see a glow from the atmosphere itself.

For example, hotter planets shouldn’t be able to form the hazes that blanket colder bodies like Saturn’s moon Titan. Instead, molecules in these same hot atmospheres should emit their own faint light that would pass through clouds, which might be measurable with next-generation telescopes. Or on planets much colder than our own, clouds made from frozen grains of water, methane, or ammonia should be highly reflective, which will help us pick up the signature of those molecules.

And for a vast swath of lukewarm planets, even those are completely covered by clouds, we can at least figure out what kind of clouds we’re dealing with by looking at how the reflected light changes with angle. “We should be able to tell the difference between different types of clouds or hazes,” Morley says.

Across the country, on a team with Seager at MIT, postdoc Alexandria Johnson is already at work on the last part.  On a recent visit, she showed off a small, cylindrical chamber. Previously, the same device has been used to make the kind of ice crystals that form wispy cirrus clouds high in Earth’s atmosphere; another time, it conjured up ersatz Martian clouds. Now Johnson plans to make truly alien clouds from particles of zinc sulfide or potassium chloride, the molecules Morley’s work suggests may obscure GJ 1214 b.

“We’re going to take the substances we think are in exoplanet atmospheres and test on them,” Johnson says.

Her experiment works by levitating a tiny chemical grain in a magnetic field, like a single suspended cloud seed. Then, Johnson shines laser light at the grain at different angles. Eventually, the plan is to match a cloud particle’s pattern of reflecting light with the light bouncing off a planet as it circles its star. With that, we’d be able to guess at the composition and size of the clouds there, and we could try to work out the atmosphere underneath.

* * *

As these experiments and others make progress, the clock is ticking. In just two years, Hubble’s successor, the James Webb Telescope, is planned for launch. Because it can see at infrared wavelengths, the Webb should have an easier time seeing through clouds.

But once it launches, the Webb will become the most precious resource in astronomy—an enormous taxpayer-funded observatory coveted by everyone, not just those who study small, enigmatic planets. And unlike Hubble, which had its lifetime extended five separate times by visits from the Space Shuttle, the Webb will be too far away to be fixed. Mindful of Hubble’s experience with GJ 1214 b, nobody wants to waste time staring at a veiled rocky planet if we can’t be sure it will work.

“It’s going to be a tough thing. It would take an inordinate amount of [Webb] time—I mean, months and months and months,” says Nikole Lewis, who is planning planet observations with the new telescope from the Space Telescope Science Institute in Baltimore, Maryland. “In that span of time you could have characterized hundreds of other exoplanets.”

That means we’ll be screening for transparent worlds, and skipping the most opaque planets. For worlds a few times the size of Earth, astronomers hope Hubble will still be around so that they can use it to check for clouds first, before bringing in Webb if the skies are clear.

But for tiny rocky planets like ours, which life-hunters are most curious about, the near-term forecast is grim. In our own solar system, Venus and Earth are roughly similar in size, although the former is a hellscape, with temperatures high enough to melt lead. A recent study imagines putting first one and then the other around a small star, the best-case scenario for us to study.  Even then, the Webb would have to look at roughly 100 transits across the star to even tell the two apart. That could add up to over 300 hours of fixed staring from the world’s most valuable telescope: more time than even the famous Hubble Deep Field, an unprecedented look at galaxies in the distant universe.

A planet will have to be incredibly promising to justify that kind of investment, according to Fortney. And even then, we might not be sure if we’ll find anything after a hard look. “I personally think the abilities of James Webb to probe earth-like planets have been oversold,” he says.

In the long run, though, planet hunters are optimistic. Outside of small, cloudy worlds there are other types of planets the Webb telescope can characterize easily. And for these stubborn planets, we’re still learning plenty about them, even if much of that is just how hard they are to probe. “I think what we’re seeing is that planets are really complicated objects,” Fortney says. “They have a lot of character.”