Statistical significance of meteorite-asteroid pairs using geocentric parameters

TL;DR

The study tests whether meteorite falls or meteorite-dropping fireballs cluster with near-Earth asteroids when similarity is evaluated in four geocentric coordinates $U$, theta, phi, and lambda_sun using the $D_N$ metric. It integrates 46 recovered meteorites, 535 candidate falls, and radiants for over 30,000 NEAs from six radiant methods, employing KDE-based and uniform-$\lambda_{sun}$ sporadic-background models and large-scale Monte Carlo simulations to derive $3\sigma$ envelopes. The results show no statistically significant excess of similarity that would indicate recent meteorite–NEA streams; any such streams, if present, would constitute less than about $0.1\%$ of falls. The study therefore sets a stringent upper bound on detectable streaming and suggests that concurrent NEA activity does not produce observable meteorite streams, though broader physical processes may influence meteorite properties without forming coherent orbital groups.

Abstract

Aims. We will test the statistical significance of meteorite-dropping fireballs and NEA clustering using the DN similarity function based on four geocentric quantities (U, theta, phi, and lambda). Methods. We calculated the cumulative similarity found between 46 meteorite falls, 535 potential meteorite-dropping fireballs, 20,516 NEAs maintained by NEODyS-2, along with 34,836 NEAs maintained by NASA/JPL HORIZONS. Statistical significance was estimated by either: (1) Kernel Density Estimation-based method to estimate the sporadic background distribution to draw random samples, or (2) applying a uniform random solar longitude (lambda). Each comparison to the synthetic sporadic population is repeated to estimate the 3-sigma region for which the cumulative similarity distribution is consistent with random association levels. Results. The observed DN cumulative similarity distribution between 46 instrumentally observed meteorite falls, 535 potential meteorite-dropping fireballs, and over 30k NEAs radiants (estimated using 6 different radiant methods) reveals no statistically significant excess of similarity between the populations consistent with streams. Conclusions. Based on nearly 600 fireball observations, there is no statistically significant clustering between meteorite falls and NEAs using geocentric impact parameters. If some meteorites arrive in streams, they make up less than ~0.1% of all falls. Recent asteroid or meteoroid physical processes could still explain features found in meteorites, but this activity is not producing distinguishable orbital streams or pairs.

Figures (4)

• Figure 1: Geocentric parameter distribution of 535 potential meteorite-dropping fireballs (blue) versus the KDE-estimated PDF (red) using a Gaussian kernel with a bandwidth of [0.2564, 0.2564, 0.2489, 0.2564, 0.1619, 0.2333, 0.2564]. This PDF, which approximates the corresponding observed sporadic population, was used to estimate the degree of random association in the population through Monte Carlo simulations.
• Figure 2: Geocentric parameter distribution of estimated NEA radiants (blue) versus the KDE-estimated PDF (red) using a Gaussian kernel: NEODyS-2 (a), Method A (b), Method B (c), Method H (d), Method Q (e), and Method W (f). These PDFs were used to estimate the degree of random association in the population through Monte Carlo simulations.
• Figure 3: CSDs of the 46 recovered meteorite falls and NEA theoretical impact radiants: NEODyS-2 (a), Method A (b), Method B (c), Method H (d), Method Q (e), and Method W (f). The red curve, labelled 'Actual Data', represents the cumulative similarity between every possible radiant--fall pair combination for the 46 falls and the theoretical NEA radiants of a particular method. The NEA radiants were estimated using six different methods, one for each panel. Two sporadic association regions are estimated. The blue region encompasses the $3\sigma$ confidence region for random associations based on the KDE-drawn samples. The red region represents the $3\sigma$ confidence region for random associations based on a uniformly random $\lambda_\odot$. In panel (a), the 'observed' and 'random' curves reported by carbognani2023identifying are also included for comparison purposes.
• Figure 4: Same as Fig. 3 but for 535 potential meteorite falls $>$1 g observed by GFO, FRIPON, or FRIPON sensors and NEA theoretical impact radiants.