Towards a definition of a meteor cluster: Detection of meteor clusters from meteor orbit databases

Abstract

As of today, there is no official definition of a meteor cluster. It is usually identified as a large number of meteors sharing a similar radiant and velocity, all occurring within a few seconds. Only eight clusters have been reported so far, from single-camera or camera network observations. We aim to provide an overview of meteor clusters to help define what constitutes a cluster by potentially adding more to the already identified ones and determining their common parameters. A search for new clusters is performed in publicly available International Astronomical Union meteor databases with the DBSCAN algorithm. Then, a statistical significance method is applied to derive the most promising cluster candidates. However, the method still lacks a way to debias the atmospheric area surveyed by the cameras due to a lack of publicly available data. A set of 16 statistically significant potential clusters is identified, involving 4 to 7 fragments. The 90th percentile includes a duration of 8 seconds, a velocity difference of 2.2 km/s, and a radiant spread of nearly 4 degrees. The velocity difference may arise from the method used for orbit computation. Meteor clusters might be more frequent than currently reported. However, we recommend that future meteor orbit databases also include a way to estimate the surveyed area by the cameras involved in the detection. This would strengthen the veracity of the 16 identified cluster candidates and ultimately allow scientists to fully debias the number of clusters, and hence derive the physical lifetime expectancy of meteoroids, which is often overlooked due to the focus on collisional lifetime estimates only. We also recommend that any future cluster observation report includes the expected number of random occurrences and consider the event to be real if this value is below 0.1.