Firestone (2009) addresses a number of controversial points arising from his proposal that the megafaunal extinctions of North America at the beginning of the Younger Dryas (YD) cooling were the result of an extraterrestrial accretion event. He and his coauthors (Firestone et al., 2007) have amassed an impressive array of diagnostic evidence including hexagonal nanodiamonds (Kennett et al., 2009) that make involvement of an extraterrestrial event apparently inescapable. The situation is reminiscent of that following publication of the Alvarez hypothesis for the K-T extinction (Alvarez et al., 1980): material with exotic compositions or characteristics has been discovered within a thin, widespread boundary layer—here, at least, on a continental scale at the base of the YD "black mat"—that represents an abrupt biological transition (Haynes, 2008); the layer also includes soot indicative of massive forest fires; and no crater as yet has been discovered.

Debate on the YD impact hypothesis has just begun, and criticism that it does not offer a complete explanation at this time is invalid. Some details of Firestone’s YD impact scenario might prove insupportable, but the impact hypothesis certainly carries equal weight to those causes previously suggested for the YD extinctions, including overhunting by humans, climate change, or hyperdisease. One distinct advantage of the YD impact hypothesis is that it will be more easily tested.

It has been almost the three decades since the Alvarez group proposed that the Cretaceous- Tertiary (K-T) mass extinction was caused by a large-body impact (Alvarez et al., 1980), and a significant amount of scientific effort continues to be expended in investigating causes for the global mass extinctions that punctuate Phanerozoic time. Initially, an obvious corollary to the Alvarez hypothesis was that all mass extinctions were caused by the immense amount of energy released by large-body impacts. Presently, however, K-T extinction at 65 Ma is the only major biologic transition generally considered to be directly associated with a large impact event. Although indicators of impact cratering have also been found at other extinction boundaries (Glikson, 2009), this evidence has been deemed inconsequential in comparison to that globally associated with the Chicxulub crater on the Yucatán Peninsula in Mexico. Moreover, the stronger correlation at global extinction boundaries seems to be with flood basalt volcanism, including recent evidence placing the second, main phase of Deccan volcanism within a narrow interval at the K-T boundary (Chenet et al., 2007). In addition, the greatest Permian-Triassic (P-T) extinction at 251 Ma is now linked primarily to eruption of the Siberian Traps (Svensen et al., 2009). It would appear that the plausible relationship between large-body impacts and global mass extinctions is less compelling than once thought. The record of impacts on Earth, however, is still being deciphered.

It seems likely that the near-surface catastrophic release of energy associated with large impacts would be especially effective in causing rapid devastation to the biosphere. If these colossally energetic events are not generally associated with extinction boundaries, then how else do they manifest themselves in the geologic record? Terrestrial examples are limited: although almost 200 impact structures have been identified so far on Earth, only a handful of strewn fields and distal ejecta layers are known. Perhaps due to our greater familiarity with craters on the solid planets and Earth’s continents (none are known from oceanic crust), our expectations of what distal ejecta layers should look like in the stratigraphic record has been skewed. Thus one potentially important link between impacts and extinctions has mostly been overlooked, and that is the effects of large impacts into the deep oceans (Hagstrum, 2005). Actually, this is the most common type of terrestrial impact as between two thirds to three quarters of all Phanerozoic impacts on Earth have been oceanic. The Earth is a particularly complex system including an atmosphere and hydrosphere, and recognizing how the vast quantities of impact energy are distributed within this system is at the crux of the problem.

Impacts into the deep oceans differ substantially from those on continents, generating megatsunamis that, as efficient transmitters of energy over great distances, can have global effects. Numerical simulations have shown that a Chicxulub-scale impactor (10 km diameter; 30 km/sec speed) landing in a 5-km-deep ocean would initially generate waves with amplitudes of ~4 km and minimum global run-up heights of 300 to 400 m (Ahrens and O’Keefe, 1983). Past extinctions are mainly understood in terms of what happened to life in the oceans, as the terrestrial fossil record is poorly preserved. Furthermore, the highest concentrations of life within the oceans are along the continental margins where devastation by megatsunamis would be greatest. Aside from mechanical damage due to the waves themselves, perhaps even more detrimental effects would result from the input of vast amounts of continental soil and sediment, adversely affecting stationary and filter feeding organisms. Marine life would be even more vulnerable to megatsunami with the continents in a supercontinental configuration, with one large ocean and a single unobstructed coastline, which was the case during the P-T extinction.

Another significant difference between numerically simulated continental and oceanic impacts of equivalent Chicxulub-scale projectiles is that most of the high-speed ejecta for an oceanic impact, from the upper ~6 km of the target, would be water or water vapor (Roddy et al., 1987). The distal ejecta layers from large oceanic impacts, therefore, would most likely appear far less significant in the stratigraphic record, in terms of material thickness, as compared to the K-T impact layer. The flow of shock and seismic energy within the Earth might also be different beneath a continental versus an oceanic transient crater. Shock wave attenuation is expected to be higher in continental crust due to its greater heterogeneity and the presence of tectosilicates, whereas a K-T sized impactor would penetrate well into the upper mantle (< 3 projectile diameters) during a deep ocean impact, possibly transferring a significant fraction of the projectile’s energy deep into Earth’s interior. Boslough et al. (1996) have modeled the axial focusing of seismic energy by Earth from a Chicxulub-scale impact, and found that the distal energy is sharply re-concentrated within the antipodal lithosphere and upper asthenosphere. Upward displacement and through-going fracturing of the antipodal lithosphere could perhaps produce the radial dike swarms and flood basalt volcanism associated with large igneous provinces (LIPs) like the Deccan and Siberian Traps. In addition, it is generally assumed that a Chicxulub-scale impact on oceanic crust would produce local volcanism infilling the seafloor crater, and the apparent antipodal distribution of hotspots might be related to this and the axial focusing mechanisms (Hagstrum, 2005). All of the hotspots antipodal to those associated with younger LIPs on Earth are in the oceans; in addition, the antipodes of only a few hotspots are located within continental crust, consistent with its possible shielding effect on the formation of antipodal hotspot pairs.

Ocean reefs have been a common casualty of past mass extinctions, but the Late Cretaceous rudistid reefs and other macrofauna (e.g. inoceramid clams) disappeared several million years before the K-T boundary (Ward, 2000). This was also a time of abrupt changes in sea level, ocean chemistry, and global clay-mineral deposition (Robert and Chamley, 1990). What appears in the marine stratigraphic record to be a rapid fall and rise in sea level at ~68 Ma could instead result from the brief introduction of energy to the depositional environment by megatsunamis. ⁸⁷Sr/⁸⁶Sr isotopic ratios indicate a sharp increase in the deposition of continental ⁸⁷Sr in the oceans at this time (Martin and Mcdougall, 1991), which might be commensurate with the backwash removal of soils and sediment from the inundated shores. Interestingly, this is also when Deccan volcanism started in India (Chenet et al., 2007). Unfortunately, at ~68 Ma the oceanic crust antipodal to the Réunion hotspot, in what is now the eastern Pacific Ocean, has been subducted beneath western North America. The Guadalupe hotspot off Mexico (19°N, 249°E), however, could have been exactly antipodal to Réunion (21°S, 56°E) at that time within a ~10 mm/yr hotspot drift limit (Hagstrum, 2005).

Having studied K-T stratigraphic sections around the Gulf of Mexico and drill cores from the Chicxulub crater, Keller et al. (2009) have concluded that the Chicxulub impact predated the Ir-rich boundary layer by ~300 kyr. I have also examined the NE Mexico K-T sections and agree that the Chixulub impact appears to have occurred well below the "tsunami" sandstone layers and the overlying Ir horizon, and that the layer of spherules at the base of the sandstones is most likely reworked. Similar sandstone layers occur elsewhere in the section of mostly marls, and are more readily explained as storm surge deposits. The smaller Ukrainian Boltysh impact structure (25 km diameter) has also been dated to ~65 Ma indicating multiple impacts near the K-T boundary. The P-T extinction was also a multiple extinction event, with the end-Guadalupe mass extinction occurring ~5 Myr before the main P-T event. Multiple large impacts over relatively short periods of time might have been related to comet showers into the inner solar system (Hut et al., 1987).

Presently, neither a large-body continental impact nor subaerial flood basalt volcanism has convincingly explained the full extent of global extinctions both on land and in the oceans. Perhaps the two events are antipodally related through the more common, but as yet unrecognized, large oceanic impacts. Large-body impacts have certainly occurred throughout geologic time, and even relatively small and more frequent objects like that related to the 1908 Tunguska airburst have locally had devastating effects. LIPs are often associated with continental rifting, and thus the origin of hotspots, their antipodal distribution, and the breakup of supercontinents remain open questions. We are in the midst of a revolution in the Earth sciences in which the far-reaching effects of catastrophic impact events are being recognized in both the biologic and geologic records; for most of modern geology’s history, until as late as the 1960s, they were not. The Alvarez paper in 1980 catalyzed this revolution, and those of Firestone and his colleagues are recent contributions, which, along with future work, will continue to advance our understanding of large-body impacts on Earth.