Watch For Falling Rocks

Extraterrestrial Impact as a Cause of Species Extinction

Lecture © Jennifer Duffy
BIOL 606 Session, University of Alberta, February 16, 2000

Extinction is the death of a taxon without issue, that is, the disappearance, from the fossil or zoological record, of every member within a taxonomic group, without evidence for direct descendants. There are various extinction types, the distinction among which hinges mainly on arbitrary cut-offs for the intensity of the event; background extinctions involve the death of up to 5% of the existing species and occur at a frequent rate of 0.25 species/million years, while pulsed extinctions kill greater than 5% of the biota, and include mass extinctions (of which there are only five throughout geological history) which cause the death of at least 60% of the species in one event (Raup, 1992).

The causes of extinction, and even the existence of some extinctions, have been hotly debated since the beginning of science. However, with Cuvier's discovery of ancient proboscidean remains (Rudwick, 1997), it became clear that readily visible organisms had disappeared completely. This, coupled with future finds of extinct groups such as trilobites and ammonites, led to a general acceptance of the existence of extinction - on a small or background scale at least. For mass extinctions, however, the debate continued, and at the close of the 19th century, two rival perspectives emerged which were based on divergent views of the completeness of the fossil record.

Gradualism, which was championed by Lyell and Darwin, holds that natural causes such as sea level changes and interspecific competition drive extinction on a background scale, and that mass or simultaneous extinction events are just artifacts of an incomplete fossil record. Catastrophism, on the other hand accepts that background extinctions could be caused by biotic and small scale earthly factors, but maintains that cataclysmic and often extraterrestrial events are required for mass extinctions. This debate continues today, but in 1980, the idea of catastrophism was bolstered by the Alvarez paper which provided physical evidence, in the form of the Iridium anomaly at K-T boundary sites, that could be indicative of a causal link between extinctions and impacts.

From this idea sprang numerous theories linking extinction and impacts, including the idea of periodicity (Raup and Sepkoski, 1984), in which bolides impact the earth every 26 My causing many of the pulsed extinctions of the past 250 My. After much debate and no physical proof, this theory awaits vindication. Instead, Raup (1991) turned to a predictability model for extinctions, called the Kill Curve, which argues that extinctions of a particular magnitude will occur within a certain length of time. He combined this with data on the age and size of impact craters on the earth to derive the Impact-Kill Curve (Raup, 1992). This predicts extinction intensity from crater size and the average likelihood that specific extinctions (and therefore, specific crater sizes) will occur in a given length of time.

Recent inquiries (Morrison, 1992) into the nature of impacts maintain that bolides have both the destructive power and relative abundance, both in the past and future, to be considered possible threats to life on earth. Whether they play a role in all of the pulsed extinctions throughout geologic time, as always, requires further testing (including specific data on crater age), but the possibility does exist.


Alvarez, LW, Alvarez, W, Asaro, F and HV. Michel 1980. Extraterrestrial Cause for the Cretaceous-Tertiary Extinction. Science 208:1095-1108.

Morrison, D. 1992. NASA International Near-Earth Object Detection Workshop. JPL, Pasadena. 85p.

Raup, DM. 1991. A Kill Curve for Phanerozoic Marine Species. Paleobiology 17:37-48.

Raup, DM. 1992. Large Body Impact and Extinction in the Phanerozoic. Paleobiology 18:80-88.

Raup, DM. and JJ. Sepkoski. 1984. Periodicity of Extinctions in the Geologic Past. P.N.A.S. 81:801-801.

Rudwick, MJS. 1997. Georges Cuvier, Fossil Bones, and Geological Catastrophes. University of Chicago Press, Chicago. 308p.


Rapporteur: Rich Palmer

Because the lecture focused on extinction in the fossil record, rather than modern causes of extinction, the discussion focused mainly on whether bolide impacts were a likely cause of apparent mass extinctions in the fossil record. Since the data for all such analyses are based on apparent disappearance of taxa, one fundamental question is purely methodological: how does one confirm that a taxon has actually gone extinct? Clearly, proving extinction of a taxon is impossible, since a 'negative' statement can never be proven. Also, we know of cases where major taxa supposedly went extinct many millions of years ago (monoplacophoran molluscs, and coelacanth fishes) but were later to be discovered alive in the deep sea.

One problem with Raup's approach was also noted: he used a Poisson distribution for bolide size, but at the same time proposed periodic (regular) cycles . . . both of these propositions cannot be true. Concerns were also raised about the falsifiability of Raup's 'kill curve' (% extinction vs. crater size) predictions: a) too few large-scale extinctions are known, b) we may never know the actual craters associated with some of these extinctions, and c) the 'error bars' for the kill curve are plus/minus a factor of two, so the precision of the prediction isn't high.

Several questions prompted interesting discussion:

1) Would bolide size affect impact probability?

Not significantly, since the probabilities are already so low, the diameter difference would likely have only a negligible effect.

2) What kinds of problems are introduced by examining patterns of extinction of higher taxa (families & genera) versus species?

Clearly, using higher taxa introduces some error, since some higher taxa have many species whereas others may have only a few. Presumably, though, the same general patterns would be comparable for all taxonomic levels. Nonetheless, it might be interesting to know whether species-rich families were more or less prone to extinction than species-poor ones.

3) Several issues came up regarding the End-Cretaceous extinction.

- Among mammals, extinctions were much more frequent among taxa of large, compared to small, body size.

- Among terrestrial plants, rather little extinction was observed. Possible reasons for lower extinction probability in plants, at least as far as bolide impacts are concerned, were:

- seed banks could provide a reserve that could easily replace extensive loss of adult plants- impact events would also likely suppress herbivores, so plants would be able to re-establish themselves quickly.