Dynamical studies of the asteroid belt reveal it to be an inadequate source of terrestrial impactors of more than a few kilometers in diameter. A more promising source for large impactors is an unstable reservoir of comets orbiting between Jupiter and Neptune. Comets 100–300 km across leak from this reservoir into potentially hazardous orbits on relatively short time scales. With a mass typically 10 3 –10 4 times that of a Chicxulub-sized impactor, the fragmentation of a giant comet yields a highly enhanced impact hazard at all scales, with a prodigious dust influx into the stratosphere over the duration of its breakup, which could be anywhere from a few thousand to a few hundred thousand years. Repeated fireball storms of a few hours' duration, occurring while the comet is fragmenting, may destroy stratospheric ozone and enhance incident ultraviolet light. These storms, as much as large impacts, may be major contributors to biological trauma. Thus, the debris from such comets has the potential to create mass extinctions by way of prolonged stress. Large impact craters are expected to occur in episodes rather than at random, and this is seen in the record of well-dated impact craters of the past 500 m.y. There is a strong correlation between these bombardment episodes and mass extinctions of marine genera.

We use the occurrence of unusual or out-of-season dust storms and dissolved ion data as proxies for dust to propose a calendar-year chronology for a portion of the Greenland Ice Sheet Project 2 (GISP2) ice core during the early sixth century A.D. Our new time scale moves a small sulfate peak to early 537 A.D., which is more consistent with recent findings of a 6 mo to 18 mo time lag between volcanic eruptions and atmospheric fallout of their sulfate aerosols. Our new time scale is consistent with a small volcanic input to the A.D. 536–537 climate downturn. We use the time range of Ni-rich fragments and cosmic spherules to provide an independent test of the chronology. The time range of Ni-rich fragments and cosmic spherules matches historical observations of “dancing stars” starting in the summer of A.D. 533 and lasting until A.D. 539 or 540. These dancing stars have been previously attributed to cosmogenic dust loading of Earth's atmosphere. The time scale cannot be shifted to be either younger or older by 1 yr without destroying the match to historical accounts of dancing stars.

Sn-rich particles, Ni-rich particles, and cosmic spherules are found together at four discrete stratigraphic levels within the 362–360 m depth interval of the Greenland Ice Sheet Project 2 (GISP2) ice core (72.6°N, 38.5°W, elevation: 3203 m). Using a previously derived calendar-year time scale, these particles span a time of increased dust loading of Earth's atmosphere between A.D. 533 and 540. The Sn-rich and Ni-rich particles contain an average of 10–11 wt% C. Their high C contents coupled with local enrichments in the volatile elements I, Zn, Cu, and Xe suggest a cometary source for the dust. The late spring timing of extraterrestrial input best matches the Eta Aquarid meteor shower associated with comet 1P/Halley. An increased flux of cometary dust might explain a modest climate downturn in A.D. 533. Both cometary dust and volcanic sulfate probably contributed to the profound global dimming during A.D. 536 and 537 but may be insufficient sources of fine aerosols. We found tropical marine microfossils and aerosol-sized CaCO 3 particles at the end A.D. 535–start A.D. 536 level that we attribute to a low-latitude explosion in the ocean. This additional source of dust is probably needed to explain the solar dimming during A.D. 536 and 537. Although there has been no extinction documented at A.D. 536, our results are relevant because mass extinctions may also have multiple drivers. Detailed examinations of fine particles at and near extinction horizons can help to determine the relative contributions of cosmic and volcanic drivers to mass extinctions.