Hyperbolic meteors: is CNEOS 2014-01-08 interstellar?


This page served as a basis for this scientific article: J. Vaubaillon, 140 WGN, the Journal of the IMO 50:5, 2022, publicly available at arxiv.

In the past few months, several announcements claiming that the meteor CNEOS 2014-01-08 observed by some “US sensors” (no specification is given) is of interstellar origin. However, the scientific community has not validated this claim, mainly for lack of scientific evidence. Apparently, “someone” (unidentified person, to our knowledge) at the US Department of Defense declared that the accuracy of the data is good enough to confirm the interstellar origin. However, the data acquisition and reduction process are not described, preventing anyone from building on such a claim.

The goal of this page is to give accurate information, based on the science of meteors, and examine if enough scientific proof are given to derive a hyperbolic nature of the CNEOS 2014-01-08 object. This page was created on the 23rd of Aug. 2022: future scientific proof regarding this object may naturally change the content and conclusions described here.

2020 Geminid (author: J. Vaubaillon/<wbr>IMCCE)
Geminide observed from the OHP in 2021. Credits J. Vaubaillon/IMCCE

Introduction to meteor sciences and observations

Basics of meteor phenomena

When a meteoroid (small rock in the interplanetary space) enters the Earth atmosphere, it hits the air molecules. The energy involved in this process is so high that atoms are excited. The disexcitement process involves the emission of light, (often) detectable by the naked eye as a meteor (a.k.a. a “shooting star”). Usual meteoroid initial geocentric velocity is comprised between 11 km/s (caused by Earth gravity) and ~72 km/s (the heliocentric hyperbolic limit). Meteors usually start at an altitude of ~100 (± 20) km. The end point depends on the size and velocity of the initial meteoroid. Meteorite falls are associated with bright meteors (fireballs). As a rule of thumb, if the initial meteoroid is large and slow enough, and is still visible at “low” altitude (typically < 30 km), the possibility of a meteorite fall is real. However, a fireball does not necessarily lead to meteorite fall. Any “fast” meteoroid (typically > 40 km/s) has nearly zero chance to survive the atmosphere entry (unless it is extremely large, like a few decametres, in which case it creates a crater during a violent impact), simply because the shock with the atmosphere is too violent, making the ablation and fragmentation process extremely efficient. In short: the fastest is the meteoroid, the less likely it is to survive.

During the Earth atmosphere entry, the meteoroid ablates and looses mass and velocity. The meteor light curve (variation of light as a function of time) is a way to derive its tensile strength: any spike in the light curve is an indication of a fragmentation event. The meteoroid tensile strength is computed by assuming the dynamic pressure is responsible for the fragmentation, and is computed with the equation: P = ρ V2, with ρ the atmospheric density and V the meteoroid velocity. Remember that the meteoroid decelerates during its atmospheric entry, as a natural consequence of its interaction with air molecules. Therefore, the velocity considered here is not the initial velocity, but the velocity at the time of the light curve spike, i.e. after the meteoroid has decelerated because of friction with the air.

The meteor light curve also allows the scientists to derive the “photometric mass”, by assuming the fraction of energy released in radiation (the so-called luminous efficiency τ). Estimates of τ vary from different studies and ranges from 0.01% to 10%. Most recent estimates spread from 0.1% to a few % at most. In other words, there are still lots of debates and works to do regarding this parameter. Last but not least, the measured deceleration allows the scientists to derive the “dynamic mass”, assuming the shape and the density of the meteoroid. Usual shapes are either sphere or brick. Usual density range from 0.1 to ~8x1E+03 kg/m3 (iron ones).

Meteor Observations

Current and widely spread method to observe the meteors is to set up at least 2 cameras, pointing towards a similar portion of the atmosphere, able to detect the same meteor from two different sites. Doing so, and after thorough and careful calculations, the 3D-trajectory can be computed. From this trajectory, provided the data are time-resolved, the velocity can be derived.


The derivation of a velocity that makes sense is far from being easy. The reasons are numerous. First, any position measurement has errors. If one computes the velocity from 2 id="1b"consecutive positions, the spread in the velocity will necessarily be extremely high. The reason is simply because v = d/t, with d being the distance between 2 positions: any position measurement error will induce a high velocity error. Another reason is that the geometry is different from one camera to the other. Suppose the meteor appear above camera 1: this camera will detect the meteor before camera 2 because camera 2 is farther away from the meteor. Suppose now that the meteor is travelling towards camera 2: instead of detecting a moving object, the camera will detect a constant position object, making the derivation of the 3D trajectory much less accurate, with consequences on the derived velocity. Regarding reduction tools, recent studies (see next paragraph) have shown that the very method used to compute the velocity will provide more or less accurate results. To make it simple: any method developed before 2016 is very likely to underestimate the real velocity uncertainty. This is no surprise that many meteor orbits are published with a 1-sigma error bar for the semi-major axis (or eccentricity). Last but not least, remember that the meteoroid velocity decreases along its path, because of the friction with the atmosphere molecules.

Long story short: the derivation of a meteor velocity is an extremely complicated process that required 2 PhDs in the recent years to tackle (see: Egal et al. (2017) and Vida et al. (2018)). Most accurate instruments, dedicated to meteor velocity measurements provide at best, depending on the meteor and camera geometry (see previous paragraph), m/s accuracy velocity. Several km/s accuracy is expected from low-resolution instruments (time and space) or low-resolution method. Remember also that the meteoroid is constantly decelerating during its flight in the atmosphere (see “Basics of meteor phenomena” section).

From Velocity to Orbit

The derivation of a meteoroid orbit from its derived velocity is another complicated problem. This can be done analytically or numerically, and a comparison by Clark & Wiegert (2011) between the 2 methods shows that, in most cases, the analytical approach is correct.

However, when publishing an orbit (usually as Keplerian osculating elements), the epoch for which these elements are provided matters. Indeed, because of gravitational perturbations, any orbit in interplanetary space (slightly but) constantly changes. In particular, for an apparent hyperbolic meteor, its past orbit might simply have been perturbed by Jupiter (known to kick small bodies out of the Solar System): this does not mean it comes from another planetary system (see also e.g. Wiegert (2014)).

Hyperbolic Meteors

Hyperbolic meteors were detected long ago: a quick search shows that such meteors were detected several decades ago. In other words, CNEOS 2014-01-08 is not the first (apparent) interstellar meteoroid.

Even more, hyperbolic meteors are expected from any observation set: given the meteor position and velocity uncertainty (see above), any meteor orbit survey program (for a review see Koten et al. 2019) will produce data containing apparent hyperbolic meteors. This is expected and perfectly normal, and is a natural consequence of measurement errors and error propagation given the method used to compute the velocity. The percentage of apparently hyperbolic meteors in a meteor orbit database is actually an indicator of the real accuracy a network (hardware, method + reduction pipeline) provides.

Another reason is the extremely fine line between a high eccentricity orbit and a hyperbolic orbit: suppose a meteoroid comes from the end of the Solar System (let say from a region close to Pluto). Its velocity as measured as a meteor in the Earth atmosphere will be “high” (typically > 30 km/s, depending on the orbital geometry): this is a simple consequence of the gravitation. Suppose now that an interstellar meteoroid coming from outside the Solar System at nearly zero velocity (at infinity): it will also show a “high” velocity. The difference between the two velocities might be less than 0.1 km/s. Again, given the uncertainties in deriving the velocity, it is expected that apparent hyperbolic meteors are measured from any meteor orbit survey. Remember also that the meteor samples the trajectory over a few dozen km, while the whole orbit might bring the meteoroid several hundred millions of kilometres from the Earth (!). In other words, we sample an extremely tiny portion of the meteoroid orbit, and the slightest error on the velocity makes it appear either bound or unbound (i.e. hyperbolic).

For a full review of the problem of apparently hyperbolic meteors, see e.g. Hajdukova et al. (2020).

Is CNEOS 2014-01-08 peculiar?

Orbit: is CNEOS 2014-01-08 interstellar?

In order to establish the hyperbolic (interstellar) character of CNEOS 2014-01-08, it would be necessary to provide at the very least:

  1. The accuracy of the measurement.
  2. The method used for the computation of the velocity.

The measurements were performed by the so called “U.S. Government sensors”, for which no information is provided, except that there are visible wavelength detectors. Any such satellite likely orbit at an altitude higher than 200 km (otherwise, the orbit decay would make it unstable). Since meteors occur at 100 km of altitude, the physical distance between the sensor and the meteor is at least twice greater than any ground-based network. The space resolution has to be at least equivalent to that of the best meteor camera resolution (which is 1 arcsec, see Vida et al. (2021)). Similarly, the frame rate must be at least comparable to ground meteor cameras, i.e. at least 25 fps. Given the performance of modern optics and sensors, to build a space satellite showing such performances is likely. In order to perform 3D measurement of the position and velocity, at least 2 satellites were used. The overall geometry matters in this case (as seen in the “Velocity” section). Last but not least, the very method used to derive the velocity matters when one wants to compute an accurate meteor velocity.

Unfortunately, none of the space and time resolution, number and geometry of the detection, nor the very method used to compute the trajectory, velocity and orbit of CNEOS 2014-01-08 are provided. As a consequence, there is no proof that the data are sufficiently accurate to indicate an interstellar trajectory. There is no proof that CNEOS 2014-01-08 is interstellar: its apparently hyperbolic orbit might simply reflect the influence of measurement uncertainty, as expected from any meteor observation, and as reported for the past few decades. The real velocity error can easily be of several km/s. This would challenge the results of a study based on a ~1 km/s accuracy.

Nature: is CNEOS 2014-01-08 of high density?

The light curve of a meteoroid allows one to derive its mass (see section “Introduction”). Spikes in the light curve allow the meteor scientists to derive its tensile strength, thanks to the ρ V2 equation. However, the velocity considered here is NOT the initial velocity: this is the velocity at the time of the light curve spike. Usually, for “large” (dm scale) object, such spike occurs when the meteoroid is “low” in the atmosphere (typically below 50 km of altitude). By then, it has lost quite an amount of velocity, because it is now entering the densest part of the atmosphere. In order to derive a tensile strength, it is therefore necessary to consider the velocity change along the 3D-trajectory.

In order to derive the tensile strength of CNEOS 2014-01-08, it would be necessary to consider the velocity change along its trajectory, and consider its velocity at the time of the light curve spike. Unfortunately, no velocity change curve is provided by CNEOS. In all cases, the velocity of CNEOS 2014-01-08 below 30 km of altitude is necessary smaller than the initial velocity at the top of the atmosphere. As a consequence, there is no proof that CNEOS 2014-01-08 is of high density (or tensile strength).

Regarding the mass: in absence of velocity curve, the dynamic mass (or size) of CNEOS 2014-01-08 cannot be derived. The photometric mass assumes a luminous efficiency τ. Given the latest estimates, a value of τ greater than 5% is unlikely (but more works are needed to constrain this parameter). Usually, comparing the dynamic and photometric mass allows scientists to put boundaries on possible mass and size of the meteoroids. An accuracy of a factor of 2 or 3 in mass estimate is perceived as a good estimate. As a consequence, given the data publicly available for CNEOS 2014-01-08, an accuracy of a factor of ~10 or more in the mass estimate is plausible.

Is it possible to recover pieces of CNEOS 2014-01-08?

According to CNEOS, CNEOS 2014-01-08 fell near the northeast coast of Papua New Guinea, i.e. in the ocean. If the initial velocity derived by CNEOS is correct (44.8 km/s), then the chances of atmospheric survival is small, since the fastest is the meteoroid, the less likely it is to survive the atmosphere entry (see section “Introduction”). If the initial meteoroid experienced 3 fragmentation events above 20 km of altitude (which is likely given the measured light curve) then a significant fraction of mass was lost during each event.

Suppose a least one fragment made it to the bottom of the ocean: the probability to recover it is also very low. The search of meteorites on dry land proves to be extremely challenging: the search area is often very large, and even teams of more than 10 people may fail at finding anything, unless very accurate measurements and searches are performed or we have the best luck ever. The only case a meteorite was found at the bottom of a lake was Chelyabinsk in 2013, because the very location of the impact was perfectly known since it fell during the winter and made a hole in the ice of a frozen lake. Russian scientists did wait for the (warmer) spring to dive and recovered a ~50 cm piece of meteorite.

“Impossible is not French” said Napoleon, but given the lack of proof regarding the accuracy of the observation, the derivation of the trajectory, velocity and tensile strength, it seems extremely unlikely to recover a piece of CNEOS 2014-01-08 at the bottom of the ocean.

For more information about meteors please check our list of IMCCE meteor activities.

Last update Monday 07 November 2022