Philip MetzgerApr 21, 2014

Forensic Ballistics: How Apollo 12 Helped Solve the Skydiver Meteorite Mystery

The news went viral a couple of weeks ago. A team in Norway announced that a skydiver was almost struck by a meteorite in flight over the ØstreÆra airstrip near Rena, Norway. Their evidence was this video. The rock passes by at about 0:15:

This content is hosted by a third party (, which uses marketing cookies. Please accept marketing cookies to watch this video.

Here is a composite picture of the falling rock:

A falling rock passes skydiver Anders Helstrup
A falling rock passes skydiver Anders Helstrup Image: Anders Helstrup / Dark Flight, photomontage: Hans Erik Foss Amundsen. Used with permission.

The event was actually two years ago. In the time since then, a team assembled and performed an impressive amount of analysis to measure the rock falling in the video and to predict where it had fallen. If confirmed as a meteorite, this would be the first time one had ever been filmed during its “dark flight,” the portion of its trajectory after the fireball when it has gone cold and is falling essentially straight down at terminal velocity. (Terminal velocity is the constant speed a falling object attains when its drag force becomes the same as its weight.)

In the video, it was impossible to tell the distance of the rock from the camera, since a single camera lacks the stereoscopic distance-finding of two human eyes, and thus the team could not tell how large the rock was or how fast it was falling. It could have been a small rock up close to the camera, or a large rock far away. When they compared it to a typical meteorite falling at terminal velocity, it seemed most likely that it was 4.6 meters away from the skydiver, making it 12 by 16 centimeters in size, and an estimated 4.6 kilograms in mass. The shape, albedo, texture, and other details of the rock also reasonably agreed with common meteorites. However, after searching the small region of forest where it was known to have fallen – an area rather tightly constrained by the video analysis – no such meteorite was found. After two years of fruitless searching, the team decided to leverage the power of crowdsourcing to see if others could find something that had been overlooked and thus help solve the mystery. As it turned out, help came all the way from the Moon.

Pete Conrad with Surveyor 3 during Apollo 12
Pete Conrad with Surveyor 3 during Apollo 12 Apollo 12 astronaut Pete Conrad poses next to the Surveyor 3 spacecraft in November 1969. Surveyor 3 landed 2 years earlier in April 1967. The Apollo 12 lunar module can be seen in the background.Image: NASA

On November 19, 1969, the Apollo 12 Lunar Module landed on the Moon’s Mare Cognitum, 160 meters from the deactivated Surveyor 3 spacecraft, which had been collecting dust (literally) for two and a half years. Astronauts Pete Conrad and Alan Bean walked to Surveyor 3, cut off several parts including its camera, a cable harness, and two struts, and brought them back to Earth for analysis. These would tell the engineers how spacecraft materials degrade in the lunar environment: the harsh ultraviolet light, severe temperature swings, vacuum, cosmic radiation, micrometeoroid impacts, and dust. This would prove invaluable. But the astronauts also made a chilling discovery: during their landing, rocket exhaust from the Apollo Lunar Module's main engine blew dust, sand, and rocks at high velocity all around the area, and some of it impacted Surveyor 3.

This was a wakeup call for engineers. Human spaceflight planners would like to put a refueling station on Mars so rockets can land, fuel up, and lift off again. Studies show that making rocket propellants out of the ice and atmosphere of Mars could reduce the cost of missions tremendously. But as Apollo 12 showed, if we land and launch near the refueling station and excessively sandblast critical hardware in the process, we could render it inoperable. And if you can’t refuel your rocket, you can’t come home.

A conceptual Mars outpost making rocket propellants from the local environment
A conceptual Mars outpost making rocket propellants from the local environment Image: NASA

A few years ago my team obtained several of the returned pieces of Surveyor 3 from the Lunar Receiving Lab at the Johnson Space Center. We examined them with a Scanning Electron Microscope and other instruments that weren’t available when they were first brought back from the Moon in the 1970’s. We counted and measured every tiny pinhole that punctured the surface. At the bottoms of these holes, we found lunar sand particles that had been blown there by the Lunar Module's rocket exhaust.

Laser-scanned map of the surface of Surveyor 3
Laser-scanned map of the surface of Surveyor 3 A laser-scanned map of the surface of Surveyor 3 showing pinholes from high velocity sand penetrating the surface, and cracks that radiate away from the penetrations. In the bottom of each hole is a lunar soil particle.

We also observed the damage that dust particles did as they crushed and scoured the surface of the hardware at a microscopic level. This provided a quantitative measure of the sandblasting effects.

Scanning electron microscope images of two pieces of Surveyor 3
Scanning electron microscope images of two pieces of Surveyor 3 Left: a part that was not exposed to the sandblasting of Apollo 12’s lunar module, showing the original form of the paint texture. Right: a part that was sandblasted by Apollo 12’s lunar module, showing paint texture that has been crushed and mixed with lunar dust, with a crack propagating across its surface.Image: NASA

We obtained high-quality copies of the videos looking out the windows of the Lunar Modules during all six lunar landings and we measured in them the blowing dust and rocks using photogrammetric methods (similar to those used by the team in Norway investigating the skydiver’s rock). These data tell us how much soil is blown by a Lunar Module and into what trajectories. But to make predictions for other spacecraft, we needed a software model, so the team developed a code that predicts how dust, sand, and rocks are blown by rocket exhaust as it expands into a vacuum at extreme velocities. The software models were compared with the Apollo and Surveyor data for validation.

The first real test of the software came in May, 2008, after launch of Space Shuttle Discovery for mission STS-124. The exhaust of the Shuttle’s solid rocket boosters pulled several thousand bricks from the wall of the launch pad’s flame trench, slammed them into each other, creating many thousands of fragments, and blew them as far as a kilometer from the pad.

Brick damage
Brick damage On Space Shuttle launch STS-124, thousands of bricks were sucked from the wall of flame trench by the flow of gas from the solid rocket boosters.
Brick fragments
Brick fragments Brick fragments were found strewn over vast acreage after the launch. Scientists John Lane and Ryan Clegg are measuring brick fragments to study how the plume blew them around the launch site.

At the same time, an experimental infrared camera spotted at least one piece of debris flying very high into the air near the Space Shuttle. The timing of this “high flyer” and its apparent origin implied that it was possibly one of these brick fragments. If so, this was very scary because the launch pad was designed to channel debris away from the vehicle, not high into the air. If indeed a brick fragment was flying high, then something was wrong with the system to keep the Shuttle safe.

Infrared video of “high flier” debris
Infrared video of “high flier” debris Infrared video shows the “high flier” debris traveling upward near the Space Shuttle during launch.

We had only a few days before the Shuttle needed to land to decide whether the vehicle had indeed been at risk of a damaging impact from a brick fragment. This was when we realized the software developed to predict the Surveyor 3 damage and to design Mars outposts could be applied to identify the Space Shuttle’s high flyer. We quickly measured its trajectory in the launch pad videos and used the software to predict what kind of a path it should have taken. We found that the path it actually took was possible only if the object was extremely light weight. In fact, it had to have the exact same density as the foam that plugs the nozzles of the solid rocket boosters prior to ignition. With this information, we knew that it was simply the foam blown out of the solid rocket boosters and thus its high trajectory was probably an ordinary event for every launch – seen for the first time because of the experimental infrared camera – and so fears that the flame trench was not doing its job were alleviated.

And this brings us to the skydiver’s falling rock in Norway, which is a very similar problem. We knew a rock was flying because the videos very clearly showed its texture and shape, but we didn’t know its size or origin. The size would be proportionately larger or smaller depending on its distance from the camera, and as it was almost certainly a rock its density was probably about 3 grams per cubic centimeter. With this information we could predict how fast the rock would fall as a function of its distance from the camera. This graph shows our model predictions.

Rock velocity versus distance from camera
Rock velocity versus distance from camera Rock velocity (vertical axis in m/s) versus distance from camera (horizontal axis in m). The solid line is the prediction using the software developed for space applications. The dashed lines are the measured velocity in the video, assuming the skydiver was falling at several different velocities. (The skydiver repeated the jump with instrumentation on his body to determine how fast he was falling, but the results showed a lot of variation from one jump to the next.) Where the dashed and solid lines cross show the possible distances of the rock from the camera.

It turns out the proportionality (the straight line in the figure above) crosses the predicted value derived by the software at only two distances from the camera, so the rock was either quite close – about a meter away and only a few centimeters in diameter – or else quite far – about 13 meters away and very large in diameter. Thus, it was either a very, very large meteorite, or it was just a small piece of gravel. The reason it could not be any in-between value was because the model predicts a rock at those in-between sizes would be experiencing something known as the “drag catastrophe.” That is when an object is falling so fast that the boundary layer of gas separates off the object and the drag force suddenly drops by a factor of almost 10. The reason why golf balls have dimples is to cause this drag catastrophe to happen at slightly slower speeds, so the ball will travel a lot farther. The rock in the video was falling too slowly for the drag catastrophe at the intermediate sizes of rocks, so it had to be either much slower and below the drag catastrophe, or a much larger rock. A meteor at the larger size would have made a brilliant fireball in the sky and would have been much easier to find on the ground. This leads to the idea that the smaller sized rock was more likely the correct solution, so it might have been just a stowaway piece of gravel that had fallen out of the parachute pack.

Coefficient of drag as a function of Reynolds number
Coefficient of drag as a function of Reynolds number Coefficient of drag (y-axis) as a function of Reynolds number (x-axis). Reynolds number is a dimensionless ratio that characterizes fluid flow. In this case it is the velocity of the rock times its diameter divided by the kinematic viscosity of the air. The sudden drop in drag that occurs between Reynolds numbers of 105 and 106 is the “drag catastrophe”.

It was also suspicious that the rock flew by the camera so near the time the parachute was released. The drogue chute was released first, then 3 seconds later the main chute opened, and 9 more seconds later the rock flew by. Using the same software, it was possible to simulate the full trajectory of a rock that fell out at various times. Assuming it fell out when the drogue chute was first released, the simulation can be performed using different size rocks to find out what size rock will pass the skydiver 12 seconds later. The model shows that a small rock will at first slow down and fall behind the skydiver because its terminal velocity is less than a human’s, even though it is experiencing the drag catastrophe when it is first released. It then slows down to below the drag catastrophe and so begins slowing down even more abruptly. However, just three seconds later the skydiver opens the main chute and slows down so much that the rock – now at its own terminal velocity – begins to catch up. It turns out that when the rock size is set to about 3 centimeters in the simulation, it passes the skydiver at 12 seconds. This rock size just happens to have the same terminal velocity as the solution we found by matching the velocity seen in the video. This provides good verification and thus real feasibility that the rock was a piece of gravel released from the parachute pack.

Simulated trajectories of the skydiver and the rock
Simulated trajectories of the skydiver and the rock Simulated trajectories of the skydiver and the rock, beginning in the upper left at the moment the drogue parachute is release and ending at the lower right as the rock passes the skydiver. Red – skydiver. Blue – rock. Distances are in meters. Vertical and horizontal axes represent vertical and horizontal distances traveled, respectively.
Separation between the rock and the skydiver
Separation between the rock and the skydiver Vertical axis – distance in meters. Horizontal axis – time from release of drogue parachute in seconds.

To be honest, I really wanted the analysis to prove that it was a meteorite, since a video of the dark flight would have been so cool! I made it clear that the team in Norway would need to critique my results and compare to the other data they had. They contacted me and – although I’m sure they were disappointed like I was – they thanked me for the analysis and said that Occam’s Razor demands the simplest explanation and so it probably was a stowaway piece of gravel. They noted that crowd sourcing paid off. In this case, crowd sourcing tapped into a line of research that began in the 1970’s with two astronauts walking on the Moon.

Let’s Go Beyond The Horizon

Every success in space exploration is the result of the community of space enthusiasts, like you, who believe it is important. You can help usher in the next great era of space exploration with your gift today.

Donate Today