Preserving the Astronomical Past

On a small hill in the heart of Cambridge, Massachusetts, 170 tons of glass, imprinted with history, sit in a dimly lit corner. Each rectangular plate is about the size of a sheet of paper and, with the exception of black ink spots, the collection is surprisingly clear, especially when you consider that some date as far back as 1885. A casual eye might easily mistake one of the 500,000 photographic plates for nothing more than a dirt-specked piece of glass.

Slide them into a plate viewer—a device reminiscent of an X-ray light box in a doctor’s office—pick up an eyepiece, and suddenly the universe appears. The spots resolve into stars, planets, even entire galaxies. The Plate Stacks Collection of the Harvard-Smithsonian Center for Astrophysics (CfA) is the largest of its kind in the world. The collection spans more than a century’s worth of astronomical images, from the first taken in 1885 to the last in 1992. Among its firsts, the collection includes an early image of the Andromeda galaxy in 1897 as well as the 1910 sighting of Halley’s comet. Thanks to the foresight of Edward Pickering, director of the observatory in the late 19th and early 20th centuries who began to collect images from observatories in Peru, the collection spans both hemispheres, too. It is an astronomical record of unprecedented proportions.

An image taken at the Harvard Observatory, now the CfA, on July 24, 1952

And one of historical significance. These plates are what Pickering’s Women, known as the Harvard Computers, pored over to observe and record stellar variability, the natural change in brightness that all stars demonstrate throughout their lives. (Similar studies today are called “time domain astronomy.”) Edwin Hubble used many of the images in the collection, like those of the Magellanic Clouds, in addition to the Computers’ work on stellar variability, to hone his model for the universe’s expansion.

“You can really tell what I like to call ‘sea change’ in the understanding of the skies when you look at these plates,” says Alison Doane, curator of the collection. “What we now know as the Andromeda galaxy is labeled as Andromeda nebula because they didn’t know at the time” that it was a galaxy, she adds.

Despite its importance, many astronomers don’t have access to the collection’s rich data—100-year-old glass plates aren’t easily shipped around the world, so scientists must make a trip to Cambridge to dig through the archive. Making sense of it all requires scouring through old notebooks kept by the Computers or spending hours hunched over the plates, meticulously noting distance between stars or measuring their relative brightness.

But now, a project dedicated to digitizing the plates in the collection, is changing that. Known as Digital Access to a Sky Century @Harvard, or DASCH, the project is the brainchild of Jonathan Grindlay, professor of astronomy at Harvard University, and is enabling astronomical discoveries of the future. “This collection is proving to be very important for time domain astronomy, which is the hot field in astronomy right now,” Grindlay says. And the Harvard glass plate collection is, “the only collection that offers both a large duration of 100 years as well as coverage of both hemispheres.”

Historical Markings

Despite the string of superlatives attached to the project, getting astronomers on board with digitization projects like DASCH has proven difficult. When Elizabeth Griffin, an astrophysicist at the Dominion Astrophysical Observatory in Victoria, Canada, speaks to people individually, she says that they see the reasoning behind preserving photographic plates and agree that something needs to be done. “But when the community comes together and needs to make a formal decision, it’s much more in the vein of, ‘Who wants this old stuff when we have all this other interesting new material?’”

To Wayne Osborn of the University of Chicago’s Yerkes Observatory in Wisconsin, preserving astronomical glass plates is of tremendous importance. “I got interested in this because I was surprised to find that a couple of major observatories had taken a series of photographs [with these plates] and had gotten rid of them,” Osborn says. “My initial interest was of rescue and preservation. And I’ve since been trying to encourage people to keep their photographic plates and catalog them.”

For both Griffin and Osborn, plate collections like the one at Harvard contain more than just astronomical information—they are historical artifacts. For instance, Gerard Kuiper, the astronomer and the namesake of the Kuiper belt, moved from Yerkes Observatory to the University of Arizona in the 1960s, taking with him nearly 2,500 plates. These plates were never returned to Yerkes, and so Osborn and his wife made a trip down to the University of Arizona to retrieve them. “We found several notebooks [of Kuiper’s] along with the collection, including some notebooks that belonged to astronomers from Princeton University,” Osborn says. “It included documentation of the 1900 solar eclipse,” one of the first eclipses to be photographed for extensive scientific study.

The 10 ½ inch round Armagh-Dunsink-Harvard telescope

Many of plates have notes directly on them, too. “Comments, handwriting, markings—these plates have all kinds of annotations,” Osborn says. “So then it becomes a matter of capturing not just what’s within these plates, but also what’s on them.” In the case of DASCH, when Doane and her team of volunteers who work on scanning these plates come upon a plate with markings, they take a digital photograph of the plate itself, to capture the markings, but then erase any markings before the plate is loaded onto the scanner. It becomes, then, a struggle between preservation for the sake of history versus for the sake of science. So far, science seems to be winning.

“The project was not done to preserve history,” Grindlay says. “We’re obviously doing that by making the data available. But in the competitive climate [of funding] these days, that never would have carried the day.”

Digital Discovery

At the CfA, the custom-built DASCH scanner is busy scanning plates. The device is massive, resembling a large microscope attached to a futuristic table saw. It can handle two plates at a time, taking 60 mini-images of each in 90 seconds and knitting them together. The entire process of loading, scanning, and unloading a plate takes about two minutes. Cleaning each plate before the scanning process eats up more time. Now at the rate of roughly 300 plates per day, DASCH has successfully scanned close to 80,000 plates since its inception in 2006. Still, that’s only about 20% of the sky. To accomplish the task, the sky has been split up into 12 sections of 15˚ of galactic latitude each, and the data is released upon completion. In June 2014, the team released the third set of data. So far, the data that has been scanned from the three sets contains nearly 3.5 billion magnitudes.

Alison Doane inspects a plate of the Small Magellanic Cloud taken on a 24

The process doesn’t stop at just creating a digital copy of the photographic images on the plates. The software behind the scanner, which was written from scratch, calculates magnitudes of stellar variability within the plates and compares it with calibrated data from preexisting catalogs of stellar variability. Already, Grindlay and other astronomers analyzing the data have discovered several new objects and new kinds of stellar variables, the existence of which were previously unknown.

“There are anywhere from 30-50 thousand stars in every plate that we scan, sometimes even 100,000,” Doane says. “That’s a lot of information that can be tapped into to learn about our universe.”

Alison Doane places a positive copy of the Small Magellanic Cloud on top of a negative of the same image. The Computers would pore over these composites to discover new variable stars.

For instance, Grindlay and his team have been studying the data from DASCH for stellar mass black holes, which are black holes formed by the collapse of massive stars. Their formation is often preceded by a sudden flare-up, from a nova or supernova, so Grindlay began to look for these stellar black holes by looking for outbursts. “We started looking at the ones in regions of the sky we had already scanned and, lo and behold, we found four of these objects that have had previous outbursts,” he says. “They had occurred 50 to even 100 years before the outbursts that occurred in modern times.” One of these objects was only discovered because, in 1999, a modern X-ray telescope detected an outburst—but Grindlay was able to identify its 1901 outburst using DASCH data. Another similar object, which was discovered in 1999, seems to have had an outburst in 1928, according to Grindlay’s findings.

“It’s not just on one plate—each of these has maybe three or four—in some cases even eight—observations that allow us to measure the duration of the outburst,” added Grindlay. Given that these are rare events, “if you only see them 1% of the time, and you see a few that are very similar in their outburst duration and intervals between outbursts, then you can immediately infer that there are many, many more than you would have otherwise thought.”

Given that less than a fifth of the sky has been scanned at DASCH, it remains to be seen what else might emerge from the plates. And while the DASCH project holds tremendous promise for the discovery and better understanding of astronomical phenomena, it’s not yet a success story. Astronomers have been trying, largely unsuccessfully, to preserve historical plates ever since they went out of use in the late 1980s.

For Griffin, the Dominion astrophysicist, “It’s been a lifelong involvement.” “When I started my career in the ’60s, these were the only way to record such images.” Griffin, who studies stellar spectra and the evolution of stars, really began to champion the preservation of these plates since much of her work also depended on looking at data from the past. “I often found myself thinking, wouldn’t it be great if I didn’t have to track down these plates, then make the trip to get them, and then scan them before I even got to analyzing the data.”

Plate Shelter

The CfA’s plate collection sits in a crowded corner, and there they will stay for the foreseeable future. But many other observatories have not been able to afford even that. One of the main reasons only about 50 plate collections exist in the entire world is because of space constraints. “The challenge is storing them correctly and having a place to store them,” Osborn says. “The conditions under which they’re stored are critical. They can’t be in very high temperatures nor can they withstand high humidity.”

Fortunately, there is one place that’s interested in storing as many plates as possible—the Pisgah Astronomical Research Institute, or PARI. Affiliated with the University of North Carolina system, this educational facility has opened its own center for the preservation of photographic plates, the Astronomical Photographic Data Archive (APDA), largely to help other institutions that are looking to get rid of their glass plate collections.

“I got a call in 2004 about an astronomer, Nancy Houk at the University of Michigan, who was retiring and looking to donate her collection,” says Michael Castelaz, the science director at PARI and an associate professor of physics at Brevard College. “We took her collection and word started to spread about our place taking collections.” APDA was formally established in 2007 and is now home to more than 40 collections—nearly 220,000 plates—from all over the country. And though the archive’s primary goal is to accept and preserve collections, the group is also making the move toward digitization. But unlike DASCH, which is supported by National Science Foundation grants, APDA has largely had to fend for itself.

The Harvard Observatory in 1899

Castelaz and PARI set up a Kickstarter campaign in 2013 to help raise funds to begin the digitization, but less than a third of its $67,000 goal was met, so the group had to return the money. They then turned to Indiegogo in February 2014, knowing that they would be able to keep whatever money they did raise, and were able to raise enough money—more than $20,000—to restore one of the two scanners they had acquired to help with the digitization.

“We finally restored this machine,” Castelaz says, “and only just started our first round of scans. We don’t yet know what kinds of things we will find.” But for many, like Osborn and Griffin, discovery isn’t the sole motivation. “Astronomical data is part of our scientific heritage,” Griffin emphasizes. “It’s very wide and broad. You just do not know how science will evolve, and we don’t know which of these data will be important.” Osborn echoes this sentiment, adding that simply digitizing these records to make the data freely available isn’t enough.

“I think it’s good to keep the original record,” Osborn says. “My preference would be to keep several storage locations, perhaps divided up based on specialties,” he says, referring to the different astronomical phenomena captured on the plates—eclipses, asteroids, galaxies, and so on. Because many of the plates were not stored under proper conditions even when they were in active use, he says it would be beneficial to “get the data off of them as soon as possible” before they deteriorate in condition.

Digitizing the data could also open up a treasure trove of historical data to a new generation. Given how difficult the plates have been to access, few newly minted astronomers think to use them. “These collections are stored away somewhere, and younger astronomers tend to not have an interest in these,” says Castelaz, echoing a sentiment shared by nearly everyone who works with these plates. “But I expect it to be more dynamic as time goes by and these digitized files gain more traction.”

“The best way at all to generate interest is to discover very unexpected and possibly controversial things,” Griffin says. Currently, she adds, “They’re finding hundreds of variable stars, which are still only getting the interest of people that are really interested in that particular field of study.”

For Grindlay and his team, however, discovery is enough. “The average person wouldn’t have much reason to believe why there would be more black holes than we originally thought,” says Grindlay, adding, “but this is basic science. These are very exotic objects that are being formed in very bizarre ways and we’re trying to peel back layers of the onion to understand how nature produces these things.”

For the others, preservation of these plates for the potential they hold is worth fighting for. “One of the biggest misconceptions is that they’re old and therefore useless,” says Castelaz. “But I would say that they are old and therefore timeless.”