At the end of 2020, planetary scientist Marek Slipski found himself glued to his computer, spending countless hours—more than he’d like to admit, he says—poring over image after image of the Martian atmosphere: zooming in, adjusting the contrast, upping the brightness, and playing around with color. Slipski, a postdoctoral researcher at NASA’s Jet Propulsion Laboratory (JPL), was looking for clouds. Although he’d written up an algorithm for the task, it was yielding mixed results, so he’d resorted to eyeballing the data instead.
But this quickly became overwhelming. Even in the small chunk of data Slipski was studying, there were so many distinct cloud populations, each varying in height and brightness. “After I did this for a week, I was like: ‘Okay, this is going to take a bit more time,’” he recalls. “And it’d be nice to have some help.”
Serendipitously, NASA had just put out a call for its Citizen Science Seed Funding Program, which gives space fans an opportunity to get involved in cutting-edge research. Slipski and Armin Kleinböhl, an atmospheric physicist at JPL, immediately started crafting a proposal. Perhaps the crowd could tackle what Slipski had mostly been attempting to do alone: identify mesospheric clouds. These float at altitudes between 50 and 80 kilometers from the surface, and can be seen in data from the Mars Climate Sounder, an instrument orbiting the planet to measure its atmospheric temperature, ice, and dust content. “We actually got selected as the only planetary proposal,” Kleinböhl says. “I guess the stars aligned—or the planets did!”
After weeks of beta testing, in late June the Cloudspotting on Mars project launched on the Zooniverse, a platform that hosts hundreds of citizen projects. So far about 2,600 volunteers have joined the effort, introducing themselves on the forums (“I am ready to chase the clouds,” a mechanic from France wrote) and digging into the climate sounder’s maps of the atmosphere at different heights, locations, and times of day. Participants need only a computer and internet access to contribute, since the data is viewed using a browser-embedded visualization tool that comes with a quick, optional tutorial.
The five researchers making up the Cloudspotting team hope that this work will shed light on the Red Planet’s global weather patterns and why its atmosphere is so thin compared to our own, and even help them understand how liquid water, once present on Mars’ surface, escaped into space. “The climatology that we will get through the citizen science project will be a lot more comprehensive than what has been in the literature so far,” says Kleinböhl, the sounder’s deputy principal investigator.
He’s particularly interested in the processes driving the formation of Martian clouds, which are composed of either carbon dioxide (dry ice) or water ice. “The CO2 clouds will tell us something about the structure and dynamics of the atmosphere, and the conditions that lead to very low temperatures,” he says, since carbon dioxide condenses at a temperature typically colder than that of the Martian atmosphere, “while the water ice clouds might tell us something about the presence of water vapor and the processes that might be responsible for transporting water vapor to these high altitudes.”
The sounder is one of six instruments aboard the Mars Reconnaissance Orbiter, which has been circling its host planet since 2006. It’s equipped with nine channels, each tuned to a different wavelength of visible or infrared light, and maps the heat radiating from the Martian surface up to 80 kilometers in the air. While the sounder wasn’t specifically designed to study clouds, the mission scientists immediately noticed prominent, archlike features in these heat maps that implied their presence. These arches, they worked out, result from the changing angle between a cloud and the infrared sensors as the spacecraft travels along its orbit. As the sounder approaches a cloud, to the sensors, the cloud appears higher in the sky. As the sounder moves beyond it, the cloud appears closer to the ground. (It’s akin to how we see the sun rise and fall in an apparent arch in our own sky as the Earth spins.) The peak of the arch, then, represents the cloud’s true altitude above the Martian surface.
For the Cloudspotting project, citizen scientists use a point-and-click tool to identify the peaks of any arches they can find in the sounder’s heat maps at different altitudes and times. Each image is shown in four different frames (the original, and three others with varying contrast and brightness levels). Users can also invert the color to spot arches that are particularly faint. To account for human error—after all, there’s some variability in what people might consider an arch—20 different users must classify the same image before it is marked complete.
Slipski and Kleinböhl initially uploaded about four-and-a-half months’ worth of images to the Cloudspotting website, expecting it’d take a couple of months for people to parse this batch of data. “But we had an overwhelming response,” Kleinböhl says. “It was really fantastic—much better than we had anticipated.” In just two weeks, citizen scientists scrutinized over 6,000 images (that’s more than 120,000 classifications), finding, on average, three to four clouds per image.
Enlisting the public to analyze massive datasets isn’t new: Scientists have used the Zooniverse to classify space photos, digitize rainfall records, and more. It’s an advantageous approach for research that involves looking for features that are too hidden, or too complicated, for a computer to identify. “Humans have evolved over millennia to be really good at pattern recognition and filtering out extraneous information,” says Haverford College astronomer Karen Masters, the principal investigator of Galaxy Zoo, the project that led to the Zooniverse’s creation. “But it’s still relatively easy to fool a computer.”
But working with the public, Masters says, comes with its own challenge: namely, getting—and keeping—people interested. To sustain engagement, the Cloudspotting team stays active in the forums: troubleshooting, responding to interesting arch features people have discovered, and sparking discussion about the science of Mars. (One volunteer pondered about how identifying the arches might help future piloted missions navigate the planet’s atmosphere: “Flying into ice clouds is a no-no!”) On July 15, Slipski hosted a webinar to give citizen scientists a chance to meet the research team as well as each other.
The Cloudspotting team doesn’t expect the volunteers to get through all of the data available from the climate sounder, taken over eight Martian years. (That’s about 16 Earth years.) Rather, once a few representative years have been analyzed, Slipski hopes to be able to use the cloud characterizations to train his algorithm to achieve more reliable results. Eventually, this would provide one of the most comprehensive, long-term datasets that scientists can use to learn about the atmospheric past and present of the Red Planet.
.jpg)
“Sixteen years of data—that is not trivial,” says Majd Mayyasi, a planetary scientist at Boston University who is not involved in the project. “It will definitely inform not just the mesospheric community, but also the lower atmospheric community and the upper atmospheric community about the properties of water and clouds, and how they’re connected.” Mayyasi, who studies how water escapes the Martian ionosphere into space, notes that clouds play an important role in how water gets transported from the surface up to higher altitudes.
“That’s been a really important part of the evolution of Mars’ atmosphere from a warm and wet planet to the cold and dry one we see today,” Slipski says. With the help of citizen scientists, the Cloudspotting team hopes to release preliminary results early next year. And once the entire dataset is analyzed, they’ll be able to expand this work to a full characterization of cloud populations and climate patterns across the planet, as well as a detailed understanding of how dust, water vapor, and carbon dioxide move through the Martian sky.
But there’s still a pile of data to analyze between now and then. Two weeks ago, the Cloudspotting team released a second batch—about 12,000 images, or eight Earth months’ worth—for people to continue classifying. They expect the citizen science project to stay active for the next two years, and the Mars Climate Sounder will keep sending back information through the end of 2022 (or longer, if NASA decides to extend the mission). “I hope that we will be able to accumulate a few more Mars years, hopefully all the way to the end of the decade,” Kleinböhl says, “to really establish a resolved, detailed climatology of the Martian atmosphere.”
