It was just before 3 AM on February 28th, 1990, when the space shuttle Atlantis lifted off from Kennedy Space Center to deliver a human skull and a classified satellite to low Earth orbit.

The timing of this unusual shuttle mission meant few people were willing to rouse themselves to view the launch. The notable exception was a small group of amateur astronomers spread across Alaska and the northernmost Canadian territories, who braved frigid temperatures hoping to glimpse the shuttle and its classified payload as they passed overhead.

The astronomers had been assembled at the behest of Ted Molczan, a Torontonian who had spent the better part of the 80s cultivating a rather unusual hobby: tracking government spy satellites. For weeks prior to the launch, the 36-year old Molczan had been pillaging his bank account to make expensive calls into the Arctic, in a desperate attempt to find astronomers willing to help him track the satellite on-board Atlantis.

Many of the astronomers Molczan contacted weren’t keen on the idea of spending hours outside in the freezing, pre-dawn darkness trying to track a satellite that was designed to avoid detection. So Molczan tried a different approach. He told the astronomers that this would, in all probability, be the only time they would ever be able to see a shuttle in orbit. A few stargazers who were unable to resist this once in a lifetime opportunity agreed to help.

Molczan and a small network of amateur spy satellite hunters usually have no problem tracking the objects of their fascination, but the strange orbit of the STS-36 mission meant this time none of them would be able to see it. For the first and only time in the history of American spaceflight, the rules that barred a rocket from flying over land were suspended so Atlantis could deliver its classified payload into a highly inclined classified orbit. This meant the shuttle would fly over America’s eastern seaboard and large swathes of Canada, but only stargazers located north of the Arctic circle would be able to see Atlantis deploy its secret payload.

Officially known as USA-53, but now better known by its alleged codename, MISTY, the satellite belonged to the Department of Defense and was widely believed to be the first instance of a new generation of photoreconaissance satellites. MISTY was reportedly similar in size and technical capabilities to the revolutionary Hubble space telescope. But rather than photographing the cosmos, it would be pointed back toward Earth to surveil America’s adversaries with unprecedented resolution.

In this respect, MISTY’s supposed design was similar to the infamous Key Hole-11 (KH-11) satellites. This latest generation of recon satellites can trace a lineage back to the very first US spy satellites, launched in the 1960s as part of the secretive Corona program. According to Molczan, however, MISTY differed from the KH-11 satellites in a very important way: It likely incorporated stealth mechanisms that made it nearly impossible to track with a radar system or optical telescope, but this wasn't discovered until years after it was launched.

This is, of course, mostly speculation. Nearly every detail of the DoD payload on the 36th shuttle mission remains classified, but it is known that the US government was extremely interested in developing satellite stealth capabilities. When the mission launched, Molczan and his collaborators knew that their only opportunity to see the classified satellite would be right after it was released from the space shuttle. Their calculations based on the shuttle’s launch trajectory suggested this would happen somewhere over the Arctic.

As luck would have it, the Arctic astronomers Molczan recruited to stake out the Atlantis passover deployment did in fact witness the classified satellite being deployed. It didn’t look like much—just one bright light trailed by another bright light—but given that tracking the object was not supposed to be possible, Molczan and his colleagues rightfully viewed this as a success.

Molczan’s Arctic recruits are likely the only civilians to have ever seen MISTY in orbit. Following a handful of initial observations on the night of the launch, there has never been another confirmed sighting. Molczan said that his colleague Russell Eberst, who formerly worked at the Royal Observatory, Edinburgh, may have unknowingly observed the MISTY satellite a few years after it was launched, but there’s no way of knowing for sure.

In fact, a few days after MISTY was launched, a Soviet newspaper reported that Russian astronomers had detected debris in the same orbit that MISTY was in, implying that the satellite may have been destroyed. Molczan told me he isn’t convinced. If anything, it seems plausible that the MISTY satellite shed debris on purpose as a way to further conceal its activities and ultimate orbit.

In the three decades that have elapsed since that fateful February night, MISTY has become Molczan’s white whale. He’s convinced that MISTY is still up there, tracing secret and ever-changing loops around our planet as it does the bidding of America’s intelligence agencies. When I spoke to Molczan he seemed resigned to the fact that, in all probability, he would never see MISTY again. Even if he did, he would never know for certain that he was seeing the same satellite.

Even though MISTY is likely lost to the spy satellite hunting community forever, there’s been no shortage of classified satellites to keep Molczan and his colleagues busy during the past 30 years. In fact, the National Reconnaissance Organization (NRO)—a government intelligence agency formed in 1961 that was so secretive its existence was classified until 1992—launched its most recent spy satellite just last month. Molczan and his colleagues were ready to track it.

So how does one begin tracking an object whose orbit is classified? The first step is to look for clues about the satellite’s trajectory that are publicly available. The most obvious place to start is with the rocket launch itself. Rockets are huge machines and a launch can be seen, heard, and felt for miles around, and details about the time and location of rocket launches are usually published weeks in advance.

To be sure a rocket poses the least amount of risk to any humans in the area, its trajectory immediately after launch is released to aviation and maritime authorities so that boats and aircraft can avoid those areas during the launch window. Molczan and his fellow spy sat hunters can get a pretty good estimate of the rocket’s initial trajectory using the coordinates of these “danger zones.” Afterwards, they can use the principles of orbital mechanics to extrapolate this data and calculate the rocket’s “parking orbit.”

After a rocket boosts into a parking orbit it coasts until it reaches the location of its final burn, where a payload is placed into orbit. Obviously, spy satellite hunters can’t know exactly where or when a spy satellite will be boosted into its final orbit—this is where guesswork based on decades of experience comes in.

Imagine a rocket launched due east from Kennedy Space Center. According to Molczan, that small amount of information is enough to predict the rocket’s payload is “almost certainly” headed to a geosynchronous orbit. This means the satellite will orbit the Earth at the same rate the planet is rotating, so that it remains over a fixed location. Based on this information, Molczan and his colleagues can assume that the payload will be placed in a geosynchronous orbit either during the satellite’s first northbound or southbound pass over the equator.

“Most new satellites are the latest in a series of satellites of the same type, and their parking orbit and subsequent maneuvers often are more or less carbon-copies of past similar launches,” Molczan said. “We draw upon our decades of experience to exploit these situations to accurately estimate search orbits.”

Once an orbit has been estimated, satellite observers can begin planning for observations. Although Molczan said some of his colleagues use telescopes equipped with sophisticated imaging and tracking technologies,

In this respect, the tools of a spy satellite hunter haven’t changed much since the days of Operation Moonwatch, a Cold War-era government program that enlisted amateur astronomers to help track the first Soviet satellites.

The program was originally conceived by the Harvard astronomer Fred Whipple, who realized there was a thriving amateur astronomer community in the US and plenty of support for other observational programs like the Ground Observer Corps, which enlisted citizens to keep watch for Soviet bombers. By combining these two pastimes, Whipple laid the groundwork for a government-sponsored network of amateur astronomers dedicated to tracking enemy satellites. Throughout the late 1950s and early 60s, citizen scientists in towns throughout the United States formed Moonwatch teams that would regularly gather to scan the skies for Soviet (and later American) satellites.

Indeed, Molczan said many of the spy satellite hunters he collaborates with cut their teeth observing Soviet satellites as a part of Operation Moonwatch.

Once a satellite observer has a rough estimate of where a satellite will pass, they use this projected orbit to select two stars in the satellite’s path as points of reference. Then, they can measure how much time the satellite takes to transit between those two stars. The stars are chosen opportunistically based on the satellite’s trajectory, but it’s best to try to select stars that are only about half a degree apart, which is roughly equivalent to the size of the moon as seen from Earth. According to Molczan, picking the right stars to measure a satellite’s pass can take a bit of getting used to for rookie observers.  

“While tracking a satellite, an observer has only a few seconds to judge whether the satellite is about to pass between a pair of suitable stars and stop tracking to make an observation,” Molczan said. “Ideal stars are seldom available so the observer cannot afford to be too picky, lest the satellite pass without making a measurement.”

When the satellite is thought to be between the two stars, the observer hits a stopwatch to mark the time it passes between them and estimates the satellite’s fractional distance. Molczan said seasoned observers can make several of these observations during a single pass. After determining which pairs of stars were used as observation markers with a star atlas, an observer can plug the data into a computer program that will help them determine the satellite’s orbital characteristics.

These computer programs run calculations based on data and orbital models the US Air Force uses to determine a satellite’s orbit. The programs make predictions about where and when the satellite will make passes in the future, so observers can make further observations that will refine the orbital predictions. Molczan said most observations made by satellite observers are in the service of refining orbital predictions to make up for slight changes in a satellite’s orbit caused by atmospheric drag and other forces.

According to Eberst, who spent decades observing satellites professionally and as a hobby, one of the most significant changes to the way the satellite tracking community operates in the past half century was the arrival of the internet.

“The internet has had an enormous effect on the tracking of satellites,” Eberst told me via email. “Among other advantages, it allows us to overcome the frustration of cloudy skies by passing on orbital details to overseas colleagues who have clearer climates.”

In addition to enabling rapid correspondence between individual collaborators, satellite observers use the web to share tracking details, photos and other information on sites like satobs.org and Heavens Above. The ability to rapidly compare notes and organize observations, coupled with the emergence of dozens of satellite tracking software programs based on decades of tracking data, has transformed the hobby. Indeed, teaching yourself how to keep tabs on the most secretive objects in the night sky has never been easier. Nevertheless, Eberst said he’s concerned about the future of the hobby.

“What frustrates me more is the lack of enthusiastic young observers who will take over this hobby of satellite tracking from the now ageing band of sky-watchers who have kept up with the activities in space for more than 60 years,” Eberst told me.  

Knowledge of US spy satellites’ orbits is incredibly valuable information for other nations and terrorist groups, who can use that information to decide when to take actions they wouldn’t want seen by America’s eyes in the sky. A memorable instance of this occurred back in 1998, when India timed its nuclear tests so they would fall between passes of US spy satellites. The US only learned of these tests later, when they analyzed photographs of the test site, which contained hints that the Indians were preparing for a test, but the satellites missed the test itself. As such, it’s hardly surprising that intelligence agencies would prefer the hobbyists would keep their tracking information to themselves, as an NRO spokesperson told the New York Times in 2008.

Molczan said he wasn’t concerned that his activities pose a threat to others’ safety and said he’s never been contacted by a government official about his pastime. After all, if he and a loose network of amateurs are able to track these satellites so efficiently, it would be trivial for a national intelligence agency to do the same. In fact, this information is now so widely available that there are smartphone apps to tell you which spy satellites are overhead at any given time.

This wouldn’t have been possible, of course, without the efforts of people like Molczan and Eberst. Today, amateur spy satellite hunters are tracking more than 360 satellites, rocket bodies, and other launch debris in their secret orbits. In the future, satellite hunters will have their work cut out for them as low Earth orbit becomes increasingly crowded with new spy satellites launched on a new generation of commercial rockets. The US military manifest already has dozens of launches on its schedule through 2023, DARPA is planning to launch a fleet of cheap spy satellites called Blackjack starting in 2021, and now the private sector is even starting to get in on the spy satellite business.

After 40 years honing his satellite hunting skills, Molczan told me he’s now too busy running his energy consultancy to go outside and do observations. Still, he helps his colleagues calculate the orbits of satellites others have observed, when he has the time. Some of these satellites have well-known names and missions, but for Molczan, the most interesting targets will always be those nameless vehicles whose purposes are condemned to be eternally obscure.