Oort cloud perturbations as a source of hyperbolic Earth impactors

The observation of interstellar objects 1I/’Oumuamua and 2I/Borisov suggests the existence of a larger population of smaller projectiles that impact our planet with unbound orbits. We analyze an asteroidal grazing me-teor (FH1) recorded by the Finnish Fireball Network on October 23, 2022. FH1 displayed a likely hyperbolic orbit lying on the ecliptic plane with an estimated velocity excess of ∼ 0.7 km s − 1 at impact. FH1 may either


Introduction
In 2017, the Pan-STARRS1 telescope observed for the first time the reflected sunlight from a metric (∼100 m) interstellar interloper, 1I/'Oumuamua (Meech et al., 2017).Two years later, the second discovery of a large object (0.4-1 km) not gravitationally bound to the Sun, comet 2I/Borisov, was announced (Guzik et al., 2020).The discoverer of 2I/Borisov himself estimated that a spherical volume of 50 au radius may have 50 bodies of more than 50 meters in diameter (Borisov and Shustov, 2021).The Pan-STARRS survey's detection of 1I/'Oumuamua allows the calculation of a number density of 0.1 au −3 , corresponding to 10 4 similar objects within Neptune's orbit and an influx of 3 objects per day (Jewitt et al., 2017).By the expected power laws of object size distribution, a much more abundant population of smaller interstellar objects is expected to cross our solar system, which may eventually collide with the Earth.A review of interstellar objects and interlopers can be found in Jewitt and Seligman (2023) and Seligman and Moro-Martín (2023).
In the early 20th century, the field of meteor science was predominantly focused on determining whether most meteors originated from interstellar or interplanetary sources (Hughes, 1982).However, it was not until the 1950s that the optical observations of fireballs generated by meteoroids exhibiting hyperbolic were reported (Opik, 1950;Almond et al., 1951Almond et al., , 1952)), in addition to subsequent meteor radar echoes detection of interstellar micrometeoroid impacts (Weryk and Brown, 2004;Froncisz et al., 2020) and interstellar dust incoming flux measurements (Meisel et al., 2002a,b).Multiple automated meteor networks have detected numerous hyperbolic Earth impactors, most of which are pointed out as the result of the instrument and method limitations ( Štohl, 1970;Hajduková, 2008;Musci et al., 2012).Hajdukova et al. (2020) reported that, of the total number of recorded events, 12.5% for CAMS, 11.9% for SonotaCO, and 5.4% for EDMOND were apparently hyperbolics.These events are clearly associated with low-quality detection and low angular elongation, so these large datasets cannot be used to discern hyperbolic impactors properly, and truly interstellar projectiles could remain hidden within the error bars.The identification of meteors with extra-solar provenance is a significant challenge and statements about the interstellar origin of IM1 and IM2 cannot be conclusive if the uncertainties of the data are not provided (Vaubaillon, 2022).Recent studies have even suggested that IM1 could be consistent with a common chondritic impactor assuming a lower atmospheric entry velocity (Brown and Borovička, 2023).Peña-Asensio et al. (2022) and (Brown and Borovička, 2023) identified hyperbolic fireballs recorded by the United States Government (USG) satellite sensors, representing ∼1% of total meter-sized impactors, events that are potentially meteorite-droppers.In contrast, there is no evidence of any recovered meteorite with a different composition from that of our solar nebula 1 .This fact opens several hypotheses: (1) CNEOS hyperbolic fireballs are spurious data; (2) There is a viable way for nearby stellar systems to be isotopically homogenous so extra-solar objects do not have distinctive non-chondritic elemental and isotopic compositions.The interstellar mate-rial exchange would be enough to smooth out any differences in the initial inventory of elements; (3) Incoming interstellar objects are biased towards low-strength properties and do not survive either the interstellar medium or the ablation process during the atmospheric entry; (4) There is an efficient mechanism by which objects that belong to our solar nebula acquire hyperbolic orbits.
In this work, we present evidence supporting the latter hypothesis, assuming that the former remains unverified, a matter still awaiting clarification.We show that apparent interstellar meteors may actually be the result of accelerated projectile impacts due to gravitational perturbations induced by massive objects (stars, free-floating brown dwarfs, rogue planets, sub-stellar or sub-Jovian mass perturbers, primordial black holes...) shaping or visiting the outer part of our solar system.In particular, we analyze a likely hyperbolic asteroid-like grazing meteor recorded in Finland in 2022 exhibiting no deceleration, which could be associated with Scholz passage.Additionally, we discuss the IM2 hyperbolic fireballs of the CNEOS database, which may belong either to the Oort cloud or to a hypothetical Oort-like Scholz's cloud if its velocity is overestimated by 22%.All events exhibit non-cometary compositions and probably are not of extra-solar provenance, which has profound implications for solar system formation models.In case they were truly interstellar in origin, the bias towards a high-strength composition of the incoming interstellar population would be reinforced.

Methodology
For the meteor science performed in this work, we use our verified Python pipeline 3D-FireTOC (Peña-Asensio et al., 2021a,b) which: performs the meteor positional reduction from the stellar astrometry accounting for asymmetric radial lens distortions (Borovicka et al., 1995) and atmospheric refraction by a revised Bennett's model (Wilson, 2018), employs the plane intersection method to reconstruct the atmospheric trajectory (Ceplecha, 1987), and computes the heliocentric orbit using the N-body orbital dynamics integrator REBOUND and REBOUNDx packages considering the gravitational harmonics (J2, J4) of the Earth and the Moon (Rein and Spiegel, 2015;Tamayo et al., 2020).Uncertainties are calculated by generating 1,000 clones from the astrometry error fits assuming a normal distribution.
For mass estimation and event classification, it is necessary to calibrate the light curve.Using the visual magnitude of the same reference stars as in astrometry, we perform aperture photometry by subtracting the local background of each one.In this way, a logarithmic fit is conducted to relate pixel values with magnitudes.We correct the atmospheric extinction and calculate the absolute magnitude of the meteor (as observed at 100 km at the zenith).
The pre-atmospheric velocity is a critical quantity for orbit estimation and cannot be directly measured by optical devices.It is necessary to derive it from the distance traveled, for which a smooth function fit of the observed points is usually performed.For this purpose, we apply the function proposed by Whipple and Jacchia (1957) that allows the velocity to be obtained straightforwardly from its derivative.However, for high-altitude grazing meteors, this model does not perform properly as it can not represent the velocity end.What is expected for the atmospheric entry of a meteoroid with these characteristics is a non-appreciable deceleration.For that reason, we assume that, within the error margins of the measurements, the pre-atmospheric and terminal velocities are virtually the same as a first approximation.Nevertheless, as we do not adjust the trajectory for the influence of gravity, we opt to analyze the initial third of the observed data points, conducting a linear fit with the mean value as the most likely velocity.The standard deviation of the fit serves as a measure of velocity uncertainty.As the entire trajectory can be used for velocity estimation without applying a deceleration model, this results in a smaller margin of error than expected for regular meteor velocity estimation from optical observations (Egal et al., 2017).
Assuming the radiated energy by the meteor is proportional to the loss of kinetic energy in the form of mass loss (Ceplecha, 1966), which is only theoretically valid for atmospheric flight with no deceleration (Gritsevich and Koschny, 2011), the initial meteoroid mass can be computed from where τ is the luminous efficiency, v is the velocity, t is the time, I = I 0 10 −0.4M is the radiated energy, I 0 = 1, 300 W is the zero-magnitude radiant power for high-speed meteors (Weryk and Brown, 2013), and M the absolute magnitude.The luminous efficiency in percent is taken from Ceplecha and McCrosky (1976): log τ = −1.51+ log v when v ≥ 27 km s −1 .However, Borovička et al. (2022) note that contemporary luminous efficiency models lead to ∼7 times less mass for velocities above 27 km s −1 , so our initial meteoroid mass may be overestimated by one order of magnitude.For example, using the Revelle and Ceplecha (2001) updated model where ln τ = −1.53+ ln v for v ≥ 25.372 km s −1 , larger average luminous efficiency of are achieved.
Following Ceplecha and McCrosky (1976), meteors can be classified according to the so-called P E criterion: where ρ e is the air density at terminal height, v ∞ is the pre-atmospheric velocity, and cos z R is the apparent radiant zenith distance.
A more physical, dimensionless form of this criterion exists (Moreno-Ibáñez et al., 2020), however, as the analyzed event presents challenges in uniquely determining their atmospheric flight parameters, we turn to the original PE criterion form.From the classical third-order system describing the meteor body deceleration, numerous efforts have been made to define a height-velocity relation (Kulakov and Stulov, 1992;Gritsevich and Stulov, 2006;Gritsevich, 2007Gritsevich, , 2008Gritsevich, , 2009;;Turchak and Gritsevich, 2014;Lyytinen and Gritsevich, 2016;Sansom et al., 2019a;Boaca et al., 2022;Peña-Asensio et al., 2023).Following these works, the dynamics of a meteor can be characterized from the analytical solution using two dimensionless parameters, namely the ballistic coefficient α and the mass loss parameter β.It is possible to express α in terms of the meteoroid bulk density ρ m , the pre-atmospheric shape factor A 0 , the drag coefficient c d , the atmospheric density at the sea level ρ 0 , the height of the homogeneous atmosphere h 0 = 7.16 km, the meteoroid mass m, and the slope of the trajectory with the horizon γ: Assuming that the ablation of the body due to its rotation is uniform over the entire surface of the meteoroid (Bouquet et al., 2014), the mass loss parameter can be calculated directly from the ablation coefficient σ and the entry velocity: We selected a uniformly distributed range of values for the ablation coefficient between 0.014 s 2 km −2 and 0.042 s 2 km −2 suitable for a rocky body based on both the classical single-body ablation and contemporary mass-loss models (Ceplecha et al., 1998;Vida et al., 2018).
Due to the high altitudes and velocities of the atmospheric flight with an enhanced mass loss under the condition of minimal deceleration, standard dynamical fits, such as α-β, may not perform properly as they are often organized as a function of velocity and are not primarily intended for highheight grazers, that is, for non-decelerating flights.Nevertheless, we can use the asymptotic form of the solution obtained to describe meteor trajectories at large values of the mass loss parameter given the formal fulfillment of the condition ln(2αβ) < h e /h 0 < ∞, where h e is the end (terminal) height (Stulov, 1997(Stulov, , 1998;;Gritsevich and Popelenskaya, 2008;Stulov, 2004;Moreno-Ibáñez et al., 2015).To account for the possible change in velocity at the end of the luminous trajectory, we use the latest modification of this solution (Moreno-Ibáñez et al., 2015;Gritsevich et al., 2016;Moreno-Ibáñez et al., 2017): Using FireOwl analysis software (Visuri et al., 2020;Visuri and Gritsevich, 2021), a Finnish Fireball Network tool that performs numerical integration of the meteoroid trajectory (Moilanen et al., 2021;Kyrylenko et al., 2023), we recompute and contrast all the results.Finally, we check the dynamic association with some meteoroid stream or parent body by means of the well-known D D orbital dissimilarity criterion proposed by Drummond (1981).
The luminous phase of FH1 started at an altitude of 126.55±0.03km (24.3104±0.0008• E, 63.6677±0.0002• N), traveling a distance of 409.47±0.09  2 shows the 3D scaled atmospheric flight reconstruction.The geocentric radiant, namely the corrected meteor anti-apex, is calculated outside of the gravitational influence field of the Earth and the Moon (at 10 times the Earth's Hill sphere), being the right ascension α R = 117.160±0.009• and the declination δ R = 19.444±0.020• .The best convergence angle between the observations is ∼50 • (for Sastamala and Vaala stations), where plane intersections with angles smaller than 5 • are excluded (only for Sastamala and Tampere stations).Table 5 in the Appendix shows the position vectors of the initial and final points of FH1's luminous path in the Earth-centered Earth-fixed coordinate system, as recorded by each of the four stations.
Figure 3 shows the apparent point-to-point velocities, together with the fitted (73.7±0.6 km s −1 ) and the parabolic velocity threshold for this specific atmospheric trajectory (∼73 km s −1 ).For 0.2 second intervals, Nyrola detection has 71% of all instant velocity measurements above this threshold, while for Vaala it is 64%.Nyrola and Vaala stations record at 25 fps, Tampere at 2.5, and Sastamala is an image of 5 seconds of exposure.Note that the apparent dispersion of the point-to-point velocities depends on the time interval selected and, paradoxically, the smaller the interval, the greater the dispersion, but the more accurate the final result.In the appendix, Tables 6 and 7 offer the detected positions of FH1 for each frame, represented in a horizontal coordinate system comprising azimuth and elevation.
We are aware that the plane intersection method may not be appropriate for long-duration grazing events (Borovicka and Ceplecha, 1992;Sansom et al., 2019b;Shober et al., 2020).Still, for this work, the approach described above is sufficient given the high velocity, brief duration (5.56 s), negligible atmospheric drag, and no dark flight calculations are required as no terminal mass is expected.Nonetheless, utilizing the fundamental equations of motion, we calculate the descent of an object due to gravitational acceleration under these conditions to be 147 m.From the inbound velocity, we estimate the heliocentric osculating orbital elements at impact to be the following: semi-major axis a = −8±5 au, eccentricity e = 1.07±0.06,inclination i = 177.18±0.04• , the longitude of the ascending node Ω = 30.10390±0.00010• , argument of periapsis ω = 16.1±0.8• , and true anomaly f = 343.9±0.8 • .The orbit shows no close encounters with any planets.Figure 4 illustrates the obtained heliocentric hyperbolic orbit.
Figure 5 shows the meteoroid absolute magnitude for every frame from Nyrola and Vaala stations, which are in good agreement with each other.FH1 curve light has a mean luminous efficiency of τ = 2.278 ± 0.018 %, a peak brightness of M = −3.0 ± 0.5, and yields a photometric initial mass of m = 1, 312 ± 54 g.Using the Revelle and Ceplecha (2001) updated model yields an average luminous efficiency of 15.96±0.13%.Note that the meteoroid underwent a smooth and gradual ablation without any flares or catastrophic disruption, resulting in the absence of saturated pixels in all recordings.
For FH1 meteor we obtain P E = −4.173±0.009.Ceplecha and McCrosky (1976)  (1998) assigned to ordinary chondrites.Note that the photometric mass used for P E classification must be computed using luminous efficiency from Ceplecha and McCrosky (1976) in Eq. 2. The obtained value is in good agreement with our estimated luminous efficiency as Revelle and Ceplecha (2001) found that asteroidal fireballs should be around 5.57% and 1.35% for carbonaceous chondrite-like fireballs.This is consistent also with similar works Subasinghe et al. (2017); Drolshagen et al. (2021a,b).In any case, the FH1 meteoroid has a consistency equal to or greater than ordinary chondrites, tending towards high-strength materials.Conservatively, taking an asteroid-like bulk density of 3,700 kg m −3 , we compute an initial meteoroid diameter of 8.75±0.12cm, which in contrast may be 4.59±0.06cm in diameter based on modern luminous efficiency models.Given the inferred meteoroid size and bulk density, an asteroidal origin seems likely (Blum et al., 2006;Trigo-Rodríguez and Llorca, 2006), although it could also be compatible with rocky pebbles ejected by cometary disintegration during inner solar system trips (Trigo-Rodríguez and Blum, 2022).From Eq. 3, we obtain a value of α of 444±4 or 849±8 depending on the initial photometric mass estimates previously calculated, and β = 25 ± 7, far away from being a meteorite-dropper event (Gritsevich et al., 2012;Sansom et al., 2019a;Boaca et al., 2022).Eq. 5 yields a velocity decrease over 5 m s −1 and 10 m s −1 , which is below the resolution of the measurements and within the uncertainty margin estimated for the velocity along the flight.
We corroborate with the FireOwl pipeline that an asteroid-like meteoroid with no catastrophic disruption and the estimated characteristics would behave in agreement with the observations.However, differences between compositions are almost marginal as the projectile experiences a low air drag during its ∼5.56 seconds of flight.Therefore, on this occasion, the dynamic models cannot provide conclusive results concerning the meteoroid density.The two candidates are the Taurids swarm and the comet P/2015 A3, with a D D =0.160 and D D =0.177, respectively.These values are well above the typically accepted threshold (Galligan, 2001), so this event is definitely not associated with any known parent body or meteoroid stream.
In summary, FH1 was a non-cometary centimeter-sized meteoroid in an inbound retrograde likely hyperbolic orbit lying almost on the ecliptic plane.It exhibits no close encounter with any known planet and a velocity excess at impact of ∼0.7 km s −1 with respect to the barycentre of the solar system.Photometry of the meteor phase yields an asteroid-like (or higher) bulk density.FH1 is the first likely hyperbolic event detected by the FFN since the beginning of the year 2004, with over 2,000 manually analyzed meteors.All computed parameters can be found in Table 2.

Discussion
The distinctive attributes of FH1, primarily its remarkably high eccentricity, could conceivably prompt conjectures regarding its interstellar origin.Notably, such speculations have been posited recently in the context of certain hyperbolic fireball events cataloged in the CNEOS database.However, a critical observation emerges when analyzing orbital inclination, which appears as a key indicator that urges to exercise caution before leaping to interstellar suppositions.We need first to investigate more plausible scenarios, including the possibility that these intriguing projectiles are either indigenous to our solar system, subject to measurement inaccuracies, or potentially subjected to gravitational accelerations.
In this section, due to the similarity with FH1, we discuss the possible interstellar origin of some CNEOS fireballs considering their uncertainties from events detected independently by ground-based stations.Additionally, we put forth the hypothesis that hyperbolic Earth impactors may be celestial bodies native to our solar nebula, which have been perturbed by close encounters with massive objects.More precisely, we propose that IM2's trajectory aligns exceptionally well in time and direction with the Scholz system fly-by when considering an overestimated velocity.

CNEOS 'interstellar' fireballs
As of October 2023, the CNEOS public database includes ∼956 fireballs starting from 1988.Among them, 6 events have hyperbolic orbits (see Table 3).These interstellar candidates have orbital inclinations lower than 25 • (with an average of 12±9 • ).As interstellar interlopers may originate from any part of the sky, the expected inclination should be an isotropic probability density function, which follows a sinusoidal distribution (Engelhardt et al., 2017) and, therefore, is uniform in cos i.This implies that the random likelihood that n orbital inclinations fulfill Consequently, the orbits of 1I/'Oumuamua and 2I/Borisov had a likelihood of being lower than | −58 • | (the largest inclination which is 1I/Omumuamu's) of ∼22%.By comparison, the likelihood of detecting six interstellar objects with inclination orbits smaller than 25 • is ∼0.00007%.This is without considering that all 6 events are in prograde orbits, which should be expected in the 50% of extra-solar visitors and would further reduce the likelihood.Therefore, multiple options can be inferred: there are shortcoming data in the CNEOS database, these hyperbolic fireballs belonged to our solar system, or they came from sources with a directional bias.
As the error bars are not provided by USG sensors, it is necessary to narrow down the uncertainties and determine the frequency of spurious data in the database.Devillepoix et al. (2019) reported that CNEOS fireball radiants are off for most events, sometimes by only a couple of degrees but other times as much as 90 • .They compared the radiants of 9 events recorded simultaneously by ground-based stations and found that the velocity vector of 4 of them was incorrectly measured by the USG sensors: Buzzard Coulee Table 4: Comparison of 17 fireballs detected by USG sensors and published on CNEOS website with independent ground-based analysis.The geocentric radiant in right ascension and declination, the entry velocity, the radiant position angle deviation, and the velocity deviation are shown.The papers used as a reference are listed in the last column.Some of the geocentric parameters have been calculated from the apparent atmospheric data.From the comparison of independently analyzed events presented in Table 4, it is found that the CNEOS radiants have a mean deviation of 21.3±43.2• in right ascension and 4.8±6.8• in declination, and a mean entry velocity deviation of 11.1±15.6%.It can be deduced from the standard deviations that the errors do not follow a normal distribution, as these distributions strongly depend on the relative geometry of the sensor and the fireball.
Eliminating the fireballs that appear as outliers based on the radiant position and velocity errors (2008TC3, Buzzard Coulee, Kalabity, Romania, Sariçiçek, 2019 MO), deviations become Gaussian where the radiant would be reduced to 1.9 • in right ascension and 1.7 • in declination, and a 3.6% deviation in velocity, which is the scenario we assume for the study of the CNEOS fireballs.Therefore, this assumption remains valid solely under the condition that the hyperbolic CNEOS events are part of the same distribution as the events characterized by elliptical orbits in Table 4. Hence, if these events are indeed outliers, the presented results here should be ignored.Consequently, 65% of CNEOS events provide measurements accurate enough for a rough estimation of heliocentric orbits.However, this also implies that the results for two out of the six hyperbolic fireballs will be inaccurate, rendering these estimated errors inapplicable.Note that there is a lack of a significant correlation between the velocity error and the actual velocity value.When a linear fit is performed, it yields a coefficient of determination as low as 0.027.In principle, we should not assume that the faster hyperbolic ones will necessarily exhibit larger errors These hyperbolic events, 2009-04-10 18:42:45, 2014-01-08 17:05:34 (IM1), 2017-03-09 04:16:37 (IM2), and 2021-05-06 05:54:27, appear to be somehow clusterized.The geocentric radiants of these 4 fireballs are suspiciously distributed around the Gemini constellation with an average radiant distance to the constellation center point of 34.2 • .We check the likelihood that 4 out of 6 randomly selected events from the CNEOS database (the 255 fireballs analyzed in Peña-Asensio et al. ( 2022)) have a lower mean distance value to Gemini (10,000 draws).We find that these 4 hyperbolic events represent 1.7σ with respect to the mean, which denotes a probability of 6.9% having occurred by chance (see Figure 6).From a completely isotropic radiant distribution, the probability of having obtained a smaller distance for 4/6 events is 4.3%.However, this does not strictly imply they are associated with each other as the anisotropic radiant distribution could be explained as well by both observational bias and solar system induced dynamics.

The massive objects fly-by hypothesis
Not all gravitationally unbound objects from our solar system are necessarily interstellar interlopers.There are different mechanisms capable of accelerating objects native to our solar nebula into hyperbolic orbits.Some of them are the secular perturbations induced by the Galactic disk or fly-by impulsive interaction of massive extra-solar bodies (Fouchard et al., 2011;Królikowska and Dybczyński, 2017).Furthermore, close encounters with the Sun or giant planets may result in an inbound excess velocity, although these are not frequent enough to explain the observed hyperbolic meteor orbits (Hajduková et al., 2014(Hajduková et al., , 2019)).Mercury has also been identified as a possible efficient producer of hyperbolic projectiles to Earth (Wiegert, 2014).Other exotic hypotheses suggest unseen stellar companions to the Sun (Davis et al., 1984) or unknown planets (Socas-Navarro, 2023) as a source of hyperbolic Earth impactors.
When the idea of the Oort cloud was introduced, i.e. a very distant region with long-period comets, it was also pointed out the existence of a mechanism to shorten their perihelia, for example, inbound hyperbolic injection produced by passing stars (Oort, 1950).Indeed, recent studies suggest that stellar close encounters send accelerated bodies into the planetary zone (Dybczyński and Królikowska, 2022).Higuchi and Kokubo (2020) showed that celestial bodies of sub-stellar mass (down to approximately 0.2 Jupiter masses) possess the ability to divert Oort cloud comets into hyperbolic trajectories characterized by small eccentricity but large perihelion distance.Other stellar systems may have also their own Oort-like clouds which could induce an influx of extrasolar objects through the planetary region when approaching the Sun (Stern, 1987).The most recent stellar fly-by to our solar system was the low-mass binary star WISE J072003.20-084651.2, also known as Scholz's star (hereafter Scholz), which crossed the outer layers of the Oort cloud at 52 +23 −14 kau about 70 +15 −10 kya ago (Scholz, 2014;Burgasser et al., 2015;Mamajek et al., 2015).de la Fuente Marcos et al. ( 2018) analyzed hyperbolic small bodies of the data provided by JPL's solar system Dynamics Group Small-Body and the Minor Planets Center (MPC) databases.They found strong anisotropies on the geocentric radiant distribution with a statistically significant overdensity of high-speed radiants towards the constellation of Gemini, which appears to be consistent in terms of time and location with the Scholz fly-by.Precisely the geocentric radiant of FH1 falls in this constellation, as well as 4 of the 6 hyperbolic fireballs from the CNEOS database that are close to Gemini or the recent Scholz motion direction.
We test the compatibility of this hypothesis by integrating backward in time the FH1 grazer and the 6 CNEOS interstellar candidates for 109,000 years to account for the estimated upper time limit of the Scholz close encounter (85,000 years).To this end, we use an orbital integrator based on a leapfrog scheme with different time steps to properly resolve the Earthinduced zenith attraction prior to the impact (Socas-Navarro, 2019).We account for the gravitational influence of the Sun, Earth, Moon, Mars, Jupiter, Saturn, Uranus, and Neptune by querying ephemerids to JPL HORIZONS system.Figure 7 shows the apparent motion of the objects starting from their geocentric radiants for the considered time.None of the events experienced a close encounter with planets, only 2021-05-06 05:54:27 fireball passed ∼5 years ago at ∼3 times Hill radius from Uranus.
New results definitively dismiss the possibility that Scholz may have penetrated the dynamically active inner Oort cloud region (<20 kAU), but support the notion that it would have passed through the outer Oort cloud where objects can have stable orbits (Dupuy et al., 2019; de la Fuente Marcos and de la Fuente Marcos, 2022).Given the apparently better constrained time (∼80 kya) and distance (∼68 kau) for the Scholz close encounter, and its current separation from the Sun (6.80 pc), it can be computed a linear velocity with respect to the solar system barycenter of v * = 82.4± 0.3 km s −1 , which is a valid approximation for 100 kya within 2.5% accuracy (Mamajek et al., 2015).Considering the fly-by occurring at high velocity and the low mass of the Scholz system (M * = 165 ± 7 M Jup ), it appears plausible that small bodies may have been injected towards the Earth.
The meteoroid FH1 was 36±18 kya ago at ∼67 kau, and the IM2 object reached the same distance 14±2 kya ago.The closest encounter found in the simulation of FH1 with the Scholz trajectory was at 39 kau, while IM2 passed at 131 kau.In spite of almost intersecting trajectories, the excess velocity of IM2 at impact (-8 km s −1 ) causes it not to be compatible in time with the Scholz passage.Looking at Table 4, it can be seen that ∼18% of the events have velocity errors around 30% or more of the nominal value.If the IM2 measurement had an uncertainty of 22% it would be perfectly compatible in time with the Scholz fly-by.Velocities proximal to the parabolic limit for FH1 are consistent, both temporally and directionally, with the passage of Scholz.This consistency necessitates only a 1% reduction in Scholz's nominal velocity, a value well within the estimated range of uncertainty.Figure 8 presents the geometric configuration of the encounter involving FH1, IM2, and Scholz.(Meech et al., 2017), and 2I/ Borisov (de León et al., 2020) are shown.The 6 dated events correspond to the hyperbolic fireballs in the CNEOS database and include the new mean deviation found for the 17 fireballs compared.All CNEOS non-hyperbolic events are also depicted, together with a center point of the Gemini constellation.Markers represent the radiant position at impact or at the current time in the case of Scholz.The ecliptic plane is plotted in yellow.
Long-period objects can acquire excess velocity from relatively low gravitational perturbations with no need for very close encounters if they are oriented in the appropriate direction at the appropriate time.Moreover, the perturbation may not necessarily have occurred during the time of the maximum approach of Scholz, which took ∼21.5 kya to traverse the Oort cloud.Considering that an object is fixed in reference to the solar system barycenter, it is possible to estimate the time-integrated impulse exerted by a passing star from classical impulse approximation (Rickman et al., 2005): where G is the gravitational constant, b o is the vector from the Oort cloud object to the Scholz closest approach, and b * is the vector from the Sun to the Scholz closest approach.
To elucidate, consider a notional object situated at a radial distance of 39 kau beyond the point of closest approach between the Scholz star system and the Sun, which occurs at 68.7 kau.According to Eq. 6, the interaction with the Scholz could impart a maximum velocity change of approximately 0.136 m s −1 .In the defined encounter geometry within the Oort Cloud, the object would need to achieve a velocity of 0.982 m s −1 to attain a perihelion distance of 1 au.The velocity impulse generated by the Scholz star would represent approximately 14% of the object's perigee velocity.Therefore, such a perturbation has the capacity to significantly modify the orbital parameters of a distantly located object, resulting in a highly eccentric inbound trajectory toward Earth.
Detection of exocomets and warm and cold debris belts around stars suggests the existence of Oort-like outer clouds material in other stellar systems (Marois et al., 2010;Kiefer et al., 2014).Highly eccentric evaporating comets are compatible with the metallic absorption lines observed in debris disk spectra, which may be evidence of exocomet clouds (Beust et al., 1990;Hanse et al., 2018).Hanse et al. (2018) found that 25% to 65% of the mass lost from the Oort cloud is due to objects being either injected into the planetary region or ejected into interstellar space mainly because of stellar encounters.A hypothetical Oort-like Scholz cloud should have the outer edge less distant from its center than the Sun due to its smaller mass.Assuming that both stellar systems have undergone similar processes, we can establish that the binding energies of their outer clouds of material are roughly the same and, therefore, their gravitational potential energies as well: where R is the distance from the objects to the star, m o is the mass of the surrounding objects, ⊙ subscript refers to the Sun, and * to another star.
Accordingly, the outer edge of the Scholz outer cloud should be scaled down as R * = 0.16R ⊙ .
Given that the classical outer edge of the Oort cloud is 200 kau (Dones et al., 2004), it is expected that the Scholz system has an outer cloud edge of 32 kau, close enough to the Sun for any object to be disturbed towards the planetary region during the fly-by, even more so when Scholz Hill sphere at the closest approach was 26 kau.In fact, IM2 clone trajectory traversed the path of the Scholz cloud and a considerable percentage of the clones intersected its Hill sphere path, while some FH1 clones passed at 2 Scholz Hill sphere radius.These possible injections could be facilitated by the joined effect of the Galactic disc, other passing stars interactions, and the presence of massive objects on the outskirts of the Scholz system or of our solar system.For example, the Scholz motion region passes through the zone of high probability for the putative planet 9 and multiple hyperbolic CNEOS fireballs fall around it (Batygin and Brown, 2016;Brown and Batygin, 2021;Socas-Navarro, 2023).Note that although the maximum velocity change in a fly-by is double the relative velocity of the encounter, a very massive nearly static object could redirect another with a zero net velocity change into a hyperbolic orbit due to the new geometry of the motion with respect to the central body.However, this situation would lead to slow unbound orbits, which might explain the modest impact velocity excess for many hyperbolic solar objects.
With successive data releases from Gaia, the space observatory of the European Space Agency, the identification of new stellar close encounters has been increasing.For example, in the first Gaia data release (GDR1) 2 stars (out of 300,000) were found to come within 1 pc of the Sun (Bailer-Jones, 2018), while in GDR2 were 26 stars (out of 7.2 million) passing within the same distance (Bailer-Jones et al., 2018) and 61 stars (out of 33 million) in GDR3 (Bailer-Jones, 2022).Bailer-Jones et al. ( 2018) inferred the present rate of stellar encounters to be 19.7±2.2 per Myr within 1 pc.This implies that the Oort cloud is expected to have experienced ∼2 interloper visits in the last 80 kya.However, only one has been identified during this period of time.
We stress that as uncertainties of CNEOS data are unavailable, no definite conclusions can be established regarding IM2.Nevertheless, considering that there may be complex gravitational interactions, we claim that both FH1 and IM2 are consistent with being gravitationally accelerated impactors originating from the Oort Cloud, likely injected during Scholz's recent fly-by.
Additionally, IM2 could plausibly be an object ejected from the outer cloud of the Scholz binary system.The other CNEOS hyperbolic events (if they are not outliers) possibly have experienced close encounters with massive objects (such as stars, free-floating brown dwarfs, rogue planets, sub-stellar or sub-Jovian mass perturbers, rogue planets, primordial black holes...) when traversing the Oort cloud less time ago than the Scholz passage, which in the case of a star could be supported by the current rate of stellar encounters.

Conclusions
We have analyzed an unusual grazing meteor (FH1) recorded in October 2022 by the Finnish Fireball Network.The cm-sized meteoroid exhibited an inbound likely hyperbolic orbit and an (at least) asteroidal consistency.Considering that its orbital plane coincides with the ecliptic and is close to the parabolic velocity limit, it seems more likely to be a perturbed Oort cloud object rather than an interstellar interloper.Within the estimated uncertainties, FH1 could be associated with the known most recent stellar encounter with our solar system, i.e., the Scholz system.
4 of the 6 hyperbolic CNEOS fireballs exhibit a statistical oddity in the geocentric radiant distribution around the Gemini constellation, an area with an overdensity of hyperbolic radiants and identified as compatible in time with the Scholz fly-by.We show statistical evidence that these events cannot really pertain to a randomly incoming interstellar population as the likelihood of their low orbital inclinations is extremely improbable compared to the expected one (with a probability of having occurred by chance of 0.00007%).Therefore, these impactors most likely belonged to our solar nebula and have been perturbed by massive objects lying on or intersecting the plane of the ecliptic.These massive objects could also form part of the Oort cloud, although this would limit the excess velocity of the projectiles.
Given the new mean uncertainties estimated in this work for CNEOS detections by benchmarking with 17 independent ground-based observations, the 2017-03-09 04:16:37 (IM2) fireball appears to be consistent with the Scholz close encounter if the velocity was overestimated by 22%, which is within the error range for ∼18% of events compared.In addition, IM2 showed a peak power in its light curve that corresponded to a dynamic pressure (i.e.aerodynamic strength) typical of iron meteorites, about ∼75 MPa, and most likely produced a recoverable metallic meteorite (Peña-Asensio et al., 2022;Siraj and Loeb, 2022b).This would inaugurate a window of opportunity for stellar archaeology sample collection with known trajectory beyond the tiny presolar grains embedded in meteorites.
If these hyperbolic impactors were interstellar visitors, it would have significant implications both for the incoming flux of extra-solar objects and for the characterization of their physical properties, which would be biased toward high-strength compositions and low inclinations.If they were Oort cloud objects, FH1 would be the second cm-sized asteroid-like observed object after Vida et al. (2023) and IM2 the first detected iron-like body from the outermost part of our solar or another stellar system.This would provide further evidence for the massive proto-asteroid belt and Jupiter's "Grand Tack" dynamical instability scenario (Shannon et al., 2019).It would imply that the Oort cloud could currently be populated not only by weak cometary objects but also by ice-free rocky material scattered by giant planets's round trip to inner orbits.
The absence of interstellar meteorites and the low orbital inclinations of the hyperbolic projectiles studied indicate that the population of massive objects forming and crossing the Oort cloud may be larger than previously thought, injecting large meteoroids into the planetary regions.Our results reinforce the idea that passing stars or other massive objects represent a source of hyperbolic Earth impactors that must be examined in detail on a case-by-case basis before claiming the interstellar origin of any object with excess velocity.

Figure 1 :
Figure 1: Blended image of the FH1 videos recorded by Sastamala station with (top) and Nyrola station (bottom).The green illuminated area is an aurora borealis.

Figure 2 :
Figure 2: 3D scaled atmospheric flight reconstruction of the FH1 meteoroid by using the Python software 3D-FireTOC.

Figure 3 :
Figure 3: Apparent velocity points of FH1 derived from Nyrola and Vaala observations computed at intervals of 0.2 seconds, fitted velocity, and the parabolic threshold.The error bars are multiplied by a factor of 10.

Figure 4 :
Figure 4: Osculating heliocentric orbit of the FH1 meteoroid (J2000).The arrow at the bottom right shows the direction to the point of the vernal equinox.

Figure 5 :
Figure 5: Photometry of FH1 from Nyrola and Vaala stations.The mean uncertainty of the magnitude is 0.44 and 0.55 respectively.

Figure 6 :
Figure6: Result of randomly selecting 6 CNEOS events and computing the mean minimum distance to the Gemini constellation center of 4 of them.The calculation is repeated 10,000 times.In cyan color are shown the draws with equal or smaller distances than the 6 CNEOS hyperbolic events.The fit to a Gaussian distribution is shown in red.

Figure 7 :
Figure7: Apparent motion of the objects starting from their geocentric radiants during the backward orbital integration.Scholz, FH1 grazer event, 1I/'Oumuamua(Meech et al., 2017), and 2I/Borisov (de León et al., 2020)  are shown.The 6 dated events correspond to the hyperbolic fireballs in the CNEOS database and include the new mean deviation found for the 17 fireballs compared.All CNEOS non-hyperbolic events are also depicted, together with a center point of the Gemini constellation.Markers represent the radiant position at impact or at the current time in the case of Scholz.The ecliptic plane is plotted in yellow.

Figure 8 :
Figure 8: Integration of clones for FH1, IM2, and Scholz.The diagram illustrates the heliocentric cartesian coordinates evolution during the encounter, including the respective outer clouds of material associated with both the Sun and Scholz.The arrow at the bottom right shows the direction to the point of the vernal equinox.

Table 1 :
Longitude, latitude, and altitude of the station recording the FH1 meteor.
• N).The flight angle with respect to the local horizon (i.e., the slope) was 3.588±0.013• , with an azimuth of 229.915±0.007• (zero being north and positive in clockwise direction).Figure

Table 2 :
Revelle and Ceplecha (2001))ric, physical, and heliocentric orbital (J2000) computed parameters of FH1 grazing meteor.The values with two results correspond to the luminous efficiency models considered:Ceplecha and McCrosky (1976)on the left andRevelle and Ceplecha (2001)on the right.