Product has been added to the basket

The Chicxulub Discussion Part 2

This page continues the Chicxulub Discussion from Discussion Part 1. Newest contributions at the top.

Chicxulub, Deccan and the K/T mass extinction

From Tim Reston

Sir, Several contributors to the Chicxulub debate have mentioned the near-synchronicity of the K/T boundary not only with Chicxulub but also with the Deccan volcanism, a subject we discussed in a recent paper (REF1). Here I would like to outline briefly the key arguments in that paper and the implications they have for the discussion about Chicxulub, multiple impacts and the K/T boundary.

The K/T boundary is not the only great mass extinction where 'impact signals' (e.g. shocked quartz, microspherules) have been reported at the same time as major cratonic flood basalts (CFBs) (REFS 1 and 2). Impact signals have also been reported, and impacts inferred as the killer blow (see REF 1 and citations therein), at the three other most recent Great Mass Extinctions - the Permo-Triassic, the end Jurassic and the late Devonian mass extinction. These correspond respectively to the eruption of the Siberian Traps CFB (REF 3), the Central Atlantic Magmatic Province CFB (CAMP - REF 4), and the Pripyat-Dniepr-Donets CFB (PDD - REF 5). The odds of a purely co-incidental major impact (large enough to cause a great mass extinction, i.e. perhaps bigger than the Chicxulub impactor) and CFB eruption (REF 1) is not too bad at 1 in 8; but the odds of two, three or four such co-incidences are 1 in 60, 1 in 450 and 1 in 3500 respectively. Unless Earth is exceptionally unlucky, it appears that either the 'impact signals' are not associated with a major impact (which would leave the CFBs as the primary killer) or the 'impact signals' and CFBs are somehow causally related.

Impacts cannot cause CFBs

The idea that a major impact could cause a CFB (REF 6) does not stand up to scrutiny. Any flash melting due to a negative pressure pulse would not explain the chemistry of CFBs and would in any case be followed by a flash freezing event seconds later (REF 1). The depth of crater excavation is insufficient to induce melting except above exceptionally hot mantle, (e.g. a mantle plume - REF 1), requiring again improbable co-incidence levels. Even if an impact could initiate some form of melting anomaly, it is hard to see how this could prove long-lived to generate for instance the Reunion track that can be followed back over 65 million years to the Deccan Traps. If an impact could somehow initiate a mantle plume, the effects of this plume would not be seen for many million years. Finally, a major K/T boundary impact (i.e. slightly later than Chicxulub it would appear) cannot have caused the Deccan Traps and the Reunion plume as the iridium anomaly layer and the K/T boundary actually occur with in the Trap sequence (REF 7) and as the Traps are preceded by alkali magmatism caused by the impact of a plume at the base of the lithosphere.

Do CFBs look like impacts?

If we can rule out the idea that a major impact caused the CFBs, is it somehow possible that the CFBs generated the signals of an impact? Our recent paper (REF 1) explores this option at length, proposing that a lithospheric-scale carbon-rich gas explosion (associated with CFB eruption) is consistent with the likely concentration of these elements in a plume head and their behaviour within the lithosphere. The paper suggests how such a major gas release event might occur (much more work is required to explore this further) during rifting following plume impact (and hence within the Trap sequence rather than at the beginning of it) and considers the energetics of such an explosion. Preliminary calculations show that such an explosion could release sufficient energy (~ 10 to the power 21 J) to eject large volumes of rock into a sub-orbital trajectory at supersonic velocity. The impact when such a mass jet returns to the surface might cause a series of craters, but we doubt if such a jet would remain coherent enough to create a single very large impact crater. The initial explosion, the subsequent mass-jet impact or perhaps most likely the rapid closure of the eruption pipe could all generate shock waves (hence shocked quartz).

The implications of our hypothesis for the timing of the events recorded in the deposits studied by Keller and others are that the Chicxulub impact may have generated the microspherules observed in the region, but had nothing to do with the K/T Ir anomaly and presumably little to do with the mass extinction. Furthermore, although there is good reason to suppose that impacts can occur in clusters, it seems special pleading to suggest that the one major impact discovered (but apparently not big enough and too early to cause a mass extinction) did not cause a Ir anomaly, whereas an otherwise unrecognised larger impact produced both the mass extinction and a dramatic Ir anomaly but no known impact deposits. Chicxulub was found because of the known occurrence of impact deposits in Haiti and elsewhere: where is the evidence for a second large impact of K/T age?

Iridium may not be related to impact

Instead, the K/T Ir anomaly, and perhaps other less pronounced anomalies, may be related not to an impact, but to high Ir-concentrations released by a gas-explosion associated with the Deccan CFB. The most concentrated current Ir emissions (comparable to the concentration at the K/T boundary) are the gases emanating from the active Reunion volcano (REF 8), that is the same source as produced the Deccan Traps. Could a lithospheric-scale gas-release explosion from the Deccan have released sufficient Ir (and in the correct proportion) into the atmosphere to produce a worldwide Ir anomaly? Does the absence of such large Ir anomalies at other mass extinctions reflect differences in the chemistry of gases erupted, related to differences in the chemistry of the source regions of the associated plumes (the core-mantle boundary?)? Again - a subject for further work.

A final comment. Impacts do occur - Chicxulub may be the bad luck scenario where a (moderately) large impact occurred at the same time as a CFB. However, the correspondence of the four most recent Great Mass Extinctions with large CFBs (REFS 1 and 2) is overwhelming evidence that such CFBs are associated with a major killing mechanism (perhaps the gas release explosion we propose, perhaps something else). The popular view, that an impact is always responsible, cannot be correct.


J. Phipps Morgan, T. J. Reston and C. R. Ranero. Contemporaneous mass extinctions, continental flood basalts, and 'impact signals': are mantle plume-induced lithospheric gas explosions the causal link?, Earth Planet. Sci. Lett., 217, 263-284, 2004

V. Courtillot, Evolutionary Catastrophes, Cambridge Univ. Press, 2002.

P.R. Renne, Z. Zichao, M.A. Richards, M.T. Black and A.R. Basu, Synchrony and causal relations between Permian-Triassic boundary crises and Siberian flood volcanism, Science 269, 1413-1416, 1995.

A. Marzoli, P.R. Renne, E.M. Piccirillo, M. Ernesto, G. Bellieni and A.

De Min, Extensive 200-Million-Year-old continental flood basalts of the Central Atlantic Magmatic Province, Science 284, 616-618, 1999.

M. Wilson and Z.M. Lyashkevich, Magmatism and the geodynamics of rifting of the Pripyat-Dnieper-Donets rift, East European Platform, Tectonophysics 268, 65-81, 1996.

A.P. Jones, G.D. Price, N.J. Price, P.S. DeCarli and R.A. Clegg, Impact induced melting and the development of large igneous provinces, Earth Planet. Sci. Lett. 202, 551-561, 2002.

N. Bhandari, P.N. Shukla, Z.G. Ghevariya and S.M. Sundaram, Impact did not trigger Deccan volcanism: evidence from Anjar K/T boundary intertrappen sediments, Geophys. Res. Lett. 22, 433-436, 1995.

N.J. Evans and C.F. Chai, The distribution and geochemistry of platinum-group elements as event markers in the Phanerozoic, Palaeogeography, Palaeoclimatology, Palaeoecology 132, 373-390, 1997.

Smitten with failed impact-tsunami theory

Gerta Keller's last word.

Please note: There will be two more contributions before we close the Great Chicxulub Debate - one from Keller and another from Smit, wrapping up their experiences of the debate and pointing the way forward.

Smit's final riposte to the Chicxulub Debate (Jan 5, 2004) concludes that 'the K/T boundary impact and the Chicxulub impact solidly remain one and the same'. To draw this conclusion he states that 'all these conclusions and evidence (referring to Keller and others - published references and presented in these web debate pages) are interpretations, not facts'. He further states, 'the large body of contrary evidence - to the K/T impact-tsunami hypothesis' (as presented by Keller and others) 'contains very little substance as well'.

Smit's statements sum up the major problem why the K/T impact mass extinction theory has continued for more than 20 years with no significant progress. It is the infatuation factor with a very popular theory to the point where the theory becomes fact and the real facts that don't fit the theory are ignored, dismissed, or labeled 'interpretation'. It has also become a game of showmanship where denial, innuendo (and sometimes character assassinations) can take the place of real scientific investigations, and where the goal is to 'outsmart your opponent'.

Unfortunately, science loses in this game. Science progresses by testing hypotheses through empirical investigations, which either confirm or refute the particular hypothesis. I believe Smit has turned this whole process on its head by refuting the investigations and resultant empirical data and treating the hypothesis as fact.

More than 12 years ago Smit et al. (l992) proposed the impact-tsunami theory to explain the siliciclastic deposits separating the K/T boundary from the underlying impact spherule ejecta deposits. At the time, the data were limited to just one outcrop (El Mimbral) followed by a few other now classic outcrops, such as El Penon. At the outset, the impact-tsunami theory seemed to fit the field investigations, at least on a superficial level. But detailed investigations of more than 40 outcrops over the past 10 years revealed critical evidence that invalidate this theory for the northeastern Mexico region for which it was developed (Adatte et al., l996; Keller et al., 1997; 2002, 2003; Ekdale and Stinnesbeck, l998).

We summarized the evidence in this web debate (see parts I and II, as well as Adatte) and replied three times to Smit's ripostes. Smit's final riposte has no new information or substantive arguments. Instead, he simply rebuts or dismisses each and every piece of evidence we presented as 'interpretation', as 'insignificant', as 'non-existent', or counters by showing his own illustrations of what (he believes) we show in our photos. Invariably, his illustrations support his assertions, but make no sense and bear no semblance to the data we present and show in our photos. For example, he has done this repeatedly in the case of the burrows of units 1, 2 and 3; he repeatedly claims that all photos of j-shaped burrows are mirror images of one and the same burrow; he repeatedly shows photos of crystals and claims that these are what Keller mistakes as foraminifera in the Chicxulub core, etc.

Here we summarize the critical data we presented in the Chicxulub debate. For the details and illustrations the reader is referred to the preceding web discussion pages.

1. Bioturbation

Fossil burrows are present within the three lithological units that supposedly form the impact-tsunami deposits. Their presence effectively rules out deposition over a period of hours to days by a tsunami.

These fossil burrows have been widely documented from the alternating sand-silt-shale layers of unit 3 where they are abundant and diverse in several discrete layers (e.g. Chondrites, Ophiomorpha, Planolites, Zoophycos). Chondrites and Ophiomorpha burrows are truncated within unit 3 by overlying sand beds El Penon (Chondrites) and Rancho Canales (Ophiomorpha) respectively. In the massive sandstone of unit 2, burrows (j-shaped spherule in-filled) are rare, but present near the base where they are truncated by rapid deposition of sand. Similar j-shaped spherule in-filled burrows truncated at the top by erosion are observed in the sandy limestone layer (SLL) that is present in the spherule unit 1. No burrows are observed within the spherules layers above and below the SLL (Ekdale and Stinnesbeck, l998; Keller et al., l997; 2002, 2003, this web discussion). These burrowed horizons represent repeated colonization of the ocean floor during deposition of unit 3, and in unit 2 and 1 also indicate invertebrates lived on the ocean floor repeatedly and disappeared at times of rapid sediment influx. Sediment deposition therefore must have occurred over an extended time interval that far exceeds a tsunami event.

Our evidence of bioturbation has generally been ignored by Smit. But faced with it in this debate he has tried hard to discredit the findings by saying that they simply represent root traces, cracks, flute casts, mud structures, rusty scratches, and even wasp nests. More specifically, he refutes the presence of the j-shaped spherule-infilled burrows in units 1 and 2 as flame structures or mere rusty scratches and shows a photo of something that looks like a rusty scratch (Smit, Fig. 1c), which he says is the only thing he observed. He seems oblivious to the fact that the j-shaped burrows are up to 8cm long and 2cm wide and clearly in-filled with spherules. They cannot be mistaken for rusty scratches. He has also argued that the j-shaped burrows we illustrate from units 1 and 2 are mirror images of one and the same burrow, even though we show the photos are clearly in their respective locations (the SLL and base of unit 2, riposte II, fig. 18).

We illustrated two burrows, in both color and black and white (as reproduced by Smit, Fig.1) one each from units 1 and 2, which Smit mistakenly assumed to represent four burrows. He further claims that one burrow was mislabeled as unit 3; if that is the case it is a typographic error. Despite our evidence, despite our explaining his error, he repeats that all images represent one and the same burrow, one photo from a hand specimen and the counter image from the field. It seems mind boggling to me that he can continue this absurd argument with such certainty. Could it be it because these burrows negate his tsunami theory?

Smit equally fervently argues that the multiple burrowed horizons in unit 3 simply represent burrowing downwards after the tsunami deposition. His 'evidence' is his interpretation in a block diagram showing Ophiomorpha burrowing down to 1m and branching out. While it is true that Ophiomorpha can burrow down to such depths, it does not mean that all Ophiomorpha burrows can be interpreted as originating at the top of unit 3. For example, we have shown that at Rancho Canales the Ophiomorpha burrows are oblique, not bifurcating, clearly different from the 'vertical bundles of tubes - that spread out horizontally' as shown by Smit for El Penon, and clearly truncated within unit 3. The organisms thus lived on the ocean floor during unit 3 sediment deposition. Moreover, there is no question that the small centimeter-long burrows of Chondrites within various fine-grained layers of unit 3 represent in situ burrowing during deposition of unit 3. The sole purpose for denial of these facts seems to be the desire to fit data to the impact-tsunami hypothesis.

We suggested that examining the microfossils within the burrows as a simple test to determine whether burrowing occurred from the Tertiary into the underlying strata of unit 3. We conducted such tests and found only late Maastrichtian microfossils within the burrows. Smit claims that no Tertiary microfossils would be present because they did not evolve for the first few thousand years after the mass extinction. At El Mimbral, early Danian microfossils are present in the cm above the red layer that contains the K/T iridium anomaly (Keller et al., l994). In the most expanded K/T boundary section at El Kef, the first Danian species are present within the basal 3cm of the boundary clay. Hence, burrowing by Ophiomorpha or any other large invertebrate would have carried plenty of microfossils into the burrows downward. None is observed.

2. Zeolite layers indicate volcanic influx inconsistent with tsunami hypothesis

Two distinct layers enriched in zeolites (clinoptilolite-heulandite) are recognized near the base and top of unit 3 in all sections examined (see Adatte et al., l996; Adatte this debate, Fig. 4). Additional zeolite-enriched layers associated with smectite are also observed in unit 1, as well as in the underlying late Maastrichtian Mendez marls and the early Tertiary shales of the Velasco Formation. These different zeolite enriched layers are correlatable from section to section over a distance of more than 300 km. An in situ-diagenetic origin of these zeolites is unlikely because of their geographic distribution and excellent corretability in different lithologies, such as sands, silts, shales and marls (Fig.4). These layers are therefore detritical in origin and indicate discrete periods of volcanoclastic influx. Their widespread presence within units 1 and 3 is further evidence that deposition occurred over an extended time period that is inconsistent with the impact-tsunami hypothesis.

Smit argues that the zeolite layers represent reworked volcanic material from the bentonites in the Mendez marls and are therefore not inconsistent with tsunami deposition. How does a tsunami wave selectively remove a bentonite layer and re-deposit it as discrete layer? The high energy waves of a tsunami would rule out such discrete redeposition.

3.Sandy limestone layer in unit 1 inconsistent with tsunami hypothesis

A 10 - 20cm thick sandy limestone layer (SLL) is present within the spherule unit 1 in most outcrops spanning an area of 300km (Keller et al., l997; Adatte et al., l996). This SLL contains some spherules at the base and top, but not generally within. The SLL is burrowed as observed by the presence of a j-shaped spherule-infilled burrow which is truncated at the top. Whole rock and clay-mineral compositions differ for the SLL and the cemented spherule-rich layer above and below. (1) The spherule-rich intercalations are primarily composed of calcite (up to 60%), decreased phyllosilicates, quartz and plagioclase; intercalations of Mendez marls have the same composition. (2) The thick SLL differs from these sediments by showing lower calcite, but higher quartz, plagioclase, chlorite and illite. This suggests distinctly different detrital influxes during deposition of the SLL and the spherule rich layers. It marks a change in the depositional environment from the spherule layer above and below to sandy limestone deposition with burrowing organisms on the ocean floor.

Smit argues that the SLL does not represent hemipelagic deposition because it 'is not even continuous over more than 10m'. The SLL is present in unit 1 of most outcrops over an area spanning 300km and hence can be regionally correlated. The siliciclastic deposits (units 1 to 3) were generally deposited in submarine channels. Exposures of non-channelized sequences are rare, but when they do occur, they also show a 10-20 cm thick sandy limestone layer (e.g. La Sierrita). It seems that Smit's argument rests on semantics, and essentially comes down to his interpretation - i.e., to call it a high-energy sandstone consistent with tsunami deposition. Even so, what is this sandstone, and the burrowing, doing within a spherule ejecta deposit that is supposed to have been deposited within hours?

4. Multiple Spherule layers in Mendez marls

The Mendez marl Formation below the spherule unit 1 was not investigated until a few years ago (Stinnesbeck et al., 2001, Keller et al., 2002, 2003). This investigation revealed the presence of up to four additional spherule layers interbedded in 10m to 12m of Mendez marls. Over 40 sections have been analyzed through detailed field investigations and laboratory analyses.

Biostratigraphy indicates that deposition of all spherule layers occurred within the Plummerita hantkeninoides zone (CF1), which spans the last 300 ky of the Maastrichtian. The lowermost spherule layer consistently is near the base of this zone. In most outcrops the 2m to 4m between the spherule layers consist of undisturbed marls. The multiple spherule layers can be correlated. This has been demonstrated particularly for the El Penon and Loma Cerca sections, which are 25 km apart and show very similar stratigraphic positions for the spherule layers interbedded in the top 10-12 m of the Mendez Formation. We interpret the stratigraphically lowermost spherule layer as the oldest layer and the original spherule ejecta deposit with an age of deposition about 300Ka prior to the K/T boundary. All other spherule layers, including the spherules of unit 1 are probably reworked from the original deposit at various times during the latest Maastrichtian in association with sea level changes.

Much has been made of some small (<10 m) isolated slumps in the Mesa Juan Perez area, which was documented by our team (Schulte et al., 2003). Smit has ceased on this local small slump to interprete all multiple spherule layers as slump deposits. In support he produced a photo of a presumed slump next to perfectly layered Mendez marls at Rancho Nuevo, which was disputed by Markus Harting (this debate). Without any data of his own or any other evidence, he concluded that all spherule layers could have been deposited as slumps over a period of 10 years (why 10 years?); that none of the spherule layers can be correlated over more than 10 m (why 10 m?), ignoring the data we present for excellent correlation over 25 km.

5. Maastrichtian Foraminifera above suevite in Yaxcopoil-1

In the new Yax-1 core drilled in the Chicxulub crater, a 50 cm thick laminated micritic limestone with minor small scale (2cm) oblique structures near the base and four thin green glauconite layers disconformably overlie the suevite breccia and underlie the K/T boundary. Late Maastrichtian planktic foraminifera characteristic of the Plummerita hantkeninoides zone CF1 have been observed and documented by us in the laminated intervals (see fig. 21, 22 of Keller reply II). Half a dozen foraminifer experts have confirmed these images as bona fide Cretaceous planktic foraminifera. Their presence in sediments above the suevite breccia and below the K/T boundary effectively rules out a K/T age for the Chicxulub impact.

In addition, magnetostratigraphy shows this interval to have been deposited in Chron 29R prior to the K/T boundary, and stable isotope data indicate normal late Cretaceous values. Sediment analysis shows that the green layers are of glauconitic origin and represent in situ formation over a prolonged time interval. This data is consistent with the earlier observations of a pre-K/T age of the oldest impact spherule layer in NE Mexico discussed above.

Smit's response to this evidence is flat denial of its existence. He claims that these images are dolomite rhombs (We have already replied to this rather strange claim in Keller et al. riposte II). He states that Arz has observed some Albian foraminifera within these sediments, but that they are not the same (there is no confirmation by Arz). This denial seems ludicrous, especially when dolomite rhombs are so totally different from foraminiferal images. It seems to serve only one purpose - to save the K/T impact theory.

Smit proposes to test whether there are foraminifera present by preparing polished thin sections. We have done so a long time ago along with ultra-thin thin sections. We have hundreds of images from these thin sections and some of the images were already reproduced in these pages. Other microfossil specialists have confirmed them as foraminifera. According to Smit even Arz has identified foraminifera in these sediments. Smit's call for an 'impartial moderator to perform this test' skirts the issue and seems to serve no other purpose than to obfuscate and delay recognition that Chicxulub predates the K/T boundary mass extinction. A better approach to solve the disagreement is for other foraminiferal specialists to examine these sediments, and to examine the same interval in other UNAM cores taken in the Chicxulub crater. Earlier studies of PEMEX cores have already indicated that there is Late Maastrichtian sediments above the impact breccia (see Ward et al., l995).

However, the issue of the age of the Chicxulub impact does not solely rest on the presence of these planktic foraminifera; there is also the magnetostratigraphy and the stable isotopes. Moreover, the sedimentology itself does not support backwash and crater infill for this interval.

6. Normal marine sedimentation or Backwash and crater infill?

We have shown evidence that the critical 50cm interval was deposited in a low energy but variable environment which was interrupted repeatedly by long periods of very slow deposition during which glauconite formed (see Keller et al., riposte II).

Smit interprets this interval as high-energy backwash and crater infill consistent with a post-impact tsunami event.

The prosecution rests?

We herewith conclude our part in this debate. No purpose is served by continuing to re-hash the same issues over and over again. The purpose of the Debate was to present the facts and interpretations to the public. We never expected to convince Jan Smit that his K/T impact-tsunami theory failed and should be retired. Others will make that decision for him. Our aim was to bring the varied evidence that doesn't fit the K/T impact-tsunami theory into the open, to let open-minded scientists and interested non-scientists see what support there is for each side and to allow them to draw their own educated conclusions. It is unfortunate that there has been absolutely no input into this debate from the K/T impact community that over the years has so strongly supported the impact-tsunami theory. Why this deafening silence? Why was there no voice in support for Jan Smit? The controversy is not over. There is more evidence in the pipeline and slowly but surely the true history of the dinosaur extinction will unravel itself.


Adatte, T., Stinnesbeck, W., and Keller, G., l996. Lithostratigraphic and mineralogical correlations of near-K/T boundary clastic sediments in northeastern Mexico: Implications for mega-tsunami or sea level changes? Geol. Soc. Am. Special Paper 307, 197-210.

Ekdale, A.A. and Stinnesbeck, W., l998. Ichnology of Cretaceous-Tertiary (K/T) boundary beds in northeastern Mexico. Palaios 13, 593-602

Keller, G., Stinnesbeck, W. and Lopez Oliva, J.G., l994. Age, deposition and biotic effects of the Cretaceous/Tertiary boundary event at Mimbral, NE Mexico. Palaios, 9, 144-157.

Keller, G., Lopez-Oliva, J.G., Stinnesbeck, W., and Adatte, T., 1997. Age, stratigraphy and deposition of near K/T siliciclastic deposits in Mexico: Relation to bolide impact? Geological Society of America Bulletin 109, 410-428.

Keller, G., Adatte, T., Stinnesbeck, W., Affolter, M., Schilli, L., and Lopez-Oliva, J.G., 2002. Multiple spherule layers in the late Maastrichtian of northeastern Mexico.Geol. Soc. Amer., Special Publication 356, 145-161.

Keller, G., Stinnesbeck, W., Adatte, T., and Stueben , D., 2003a. Multiple impacts across the Cretaceous-Tertiary boundary. Earth Science Reviews 62, 327-363.

Keller, G., Stinnesbeck, W., Adatte, T., and Holland, B., Stueben, D., Harting, M., C. de Leon and J. de la Cruz, 2003b. Spherule deposits in Cretaceous/Tertiary boundary sediments in Belize and Guatemala. J. Geol. Society of London, 160, 783-795.

Schulte, P., Stinnesbeck, W., Stueben, D., Kramar, U. Berner, Z., Keller, G., Adatte, T., 2003. Fe-rich and K-rich mafic spherules from slumped and channelized Chicxulub ejecta deposits in the northern La Sierrita area, NE Mexico. Int. J. Earth Sci. 92, 114-142.

Smit, J., Montanari, A., Swinburne, N.H.M., Alvarez, W., Hildebrand, A., Margolis, S.,

Claeys, P., Lowrie, W., and Asaro, F., l992. Tektite bearling deep water clastic unit at the Cretaceous-Tertiary boundary in northeastern Mexico. Geology, v. 20, 99-103.

Smit, J., Roep, T.B., Alvarez, W., Montanari, A., Claeys, P., Grajales-Nishimura, J.M. and Bermúdez, J., 1996. Coarse-grained, clastic sandstone complex at the K/T boundary around the Gulf of Mexico: Deposition by tsunami waves induced by the Chicxulub impact. Geological Society of America Special Paper 307,151-182.

Stinnesbeck, W., Barbarin, J.M., Keller, G., Lopez-Oliva, J.G., Pivnik, D.A., Lyons, J.B., Officer, C.B., Adatte, T., Graup,G., Rocchia, R., and Robin, E., l993. Deposition of channel deposits near the Cretaceous-Tertiary boundary in northeastern Mexico: Catastrophic or 'normal' sedimentary deposits? Geology 21, 797-800.

Stinnesbeck, W., Keller, G., Adatte, T., Lopez-Oliva, J.G., and N. MacLeod, l996.Cretaceous-Tertiary boundary clastic deposits in northeastern Mexico: impact tsunami or sea level lowstand? In MacLeod N and Keller, G., (eds),Cretaceous- Tertiary Mass Extinctions. W.W. Norton & Company, New York, 471-518.

Stinnesbeck, W., Schulte, P., Lindenmaier, f., Adatte, T., Affolter, M., Schilli, L., Keller, G., Stueben, D., Berner, Z., Kramer, U. and J.G. Lopez-Oliva, 2001. Late Maastrichtian age of spherule deposits in northeastern Mexico: Implication for Chicxulub scenario. Canadian Journal of Earth Sciences 38, 229-238.

Dolomites in juxtaposition, or forams?

From Mike Simmons* Posted 7.01.04

Sir, As a completely impartial observer I have enjoyed the debate on the timing of the K/T related impact, especially the recent rebuttal by Smit to the arguments of Keller published in Geoscientist (and copied in the discussion website). Herewith a couple of observations from a sometime micropalaeontologist.

One of Keller's key arguments is that Maastrichtian foraminifera occur in pelagic micrites above the impact material in the Yaxcopoil-1 core. Smit refutes this with illustration of facies (his Figure 5) in his recent Geoscientist article, noting that these are sandstones, not micrites. From his illustrations I would be inclined to disagree, this look like dolomitic micrites with silt-grade clasts (of what exactly is impossible to tell). However the top-right photograph contains some suspicious foraminifera-like blobs.

Keller, in response, has illustrated what she believes are Maastrichtian planktonic foraminifera from this unit. Although the photographs are rather small, I am dubious that they are what she suggests they are. They could well be dolomite rhombs in juxtaposition. So, in summary I tend to accept Keller's view that the post-impact micrites are indeed micrites, but cannot confirm her interpretation of Maastrichtian foraminifera.

I am experienced in studying Maastrichtian planktonic foraminifera in thin-section (see for example Simmons, M.D., Williams, C.L. & Hancock, J.M. 1996. Planktonic foraminifera across the Campanian/Maastrichtian boundary at Tercis, south-west France. Newsletters in Stratigraphy, 34, 65-80) and would be happy to adjudicate.

*Neftex Petroleum Consultants Ltd

Not a single shred... January 5, 2004

Not a shred of evidence holds up that Chicxulub is older than the K/T boundary, replies Jan Smit 

Long-term deposition of near K-T clastic deposits in NE Mexico: Lithostratigraphic and Mineralogic evidence? No, strong arguments for rapid deposition

says Jan Smit - received 24.12.03

Sir, I wholeheartedly agree with Thierry Adattes presentation (below) of the data and mineralogy, and most of his conclusions. Except in his interpretaton that these beds represent necessarily long-term deposition.

The three units of the clastic beds are indeed correlatable over 300km, as already published in Smit et al. (1996). This feature by itself argues for a depositional mechanism that operates over this distance. Such mechanism is compatible with tsunami deposition. However, also other depositional mechanisms are often suggested in connection with the clastic beds, like sealevel changes, or turbidity current (flysch) deposits. However, the latter two are amply known from the geological record, but to my knowledge such persistent subdivision in three distinct units has not been described from other known occurrences. This points to a rather unique depositional mechanism for the clastic beds.

To repeat Adatte here:
  • He argues for rapid deposition of Unit 1.
  • He argues for rapid deposition of Unit 2.
  • He argues for rapid deposition of Unit 3.
The sequential differences in composition of the three units are excellenty explained by progressive erosion and transport triggered by tsunami waves, including the influx of volcanic detritus that is amply available in the rising Sierra Madre in the west at that time, as witnessed by the frequent bentonites in the Mendez and Velasco Formations

In other words, Adatte pleads strongly for rapid deposition of the clastic beds, totally compatible with deposition by tsunami waves or deposition triggered by tsunami waves.


Smit, J., T. B. Roep, et al. (1996). Coarse grained, clastic sandstone complex at the K/T boundary around the Gulf of Mexico: Deposition by tsunami waves induced by the Chicxulub impact? The Cretaceous-Tertiary Event and Other Catastrophes in Earth History. G. Ryder, D. Fastovski and S. Gartner. Boulder, Geol. Soc. of Amer. Sp. Pap. 307: 151-182.

Long-term deposition of near K-T clastic deposits in NE Mexico: Lithostratigraphic and Mineralogic evidence

Reply to Smit from Thierry Adatte - posted 23.12.03*

*Geological Institute, University of Neuchâtel, 11 rue Emile Argand, 2007 Neuchâtel, SWITZERLAND,

1. Lithology and mineralogy of clastic deposits

In northeastern Mexico, clastic deposits are discontinuously present near the top of the Mendez marl Formation and just below the K-T boundary. The clastic deposit is generally subdivided into three distinct units based on their lithology, sedimentology and mineralogy. At the base is a spherule-rich unit 1, which varies from 2 cm to 100 cm thick and frequently contains a sandy limestone layer. (The additional spherule layers present in the Mendez marl Formation below were earlier discussed in Debate Round I and II by Keller Adatte and Stinnesbeck (see also Keller et al., 2002, 2003). The overlying unit 2 consists of laminated sandstone up to 4 m thick, but this unit is absent in some sections. The top unit 3 consists of intercalated sand, silt and shale layers that reach 3 m, but this unit is also of variable thickness and absent in some sections. The variable thicknesses of these units across northeastern Mexico is shown in Figures 1a and 1b and indicates that these clastic deposits are therefore channelized.

Clues as to the origin, nature and tempo of deposition can be obtained from the lithology and mineralogy of these clastic sediments (Adatte et al., 1996). These show a surprisingly constant bulk rock composition based on analyses of 20 sections. Moreover, the three units are lithologically and mineralogically correlatable over 300 km. In all outcrops, the clastic deposit disconformably overlies the gray marls of the Maastrichian Mendez Formation.

Mendez marl Formation

Mineralogical analysis of the Mendez marl Formation indicates an average composition of 48% calcite, 30% phyllosilicates, 15% quartz and 8% plagioclase (Figs. 2a-2b). Clay-mineral contents, such as mica (illite), chlorite, irregular mixed-layers chlorite-smectite and illite-smectite, are variable (Fig.3) This is especially the case with chlorite, which varies in the Mendez Formation from 60% (<2µm size fraction) at Lajilla I and Mulato sections to less than 20% in the La Sierrita and El Peñon sections. These mineralogically different sediment layers below the unconformity suggest erosion to variable depths within the Mendez marls.

Unit 1: spherule rich layer

Whole rock and clay-mineral compositions of the cemented spherule-rich layer of unit 1 are very similar in all outcrops and show two trends. 1) Whole rock composition of unit 1 marks repeated intercalations of spherule-rich layers with increased calcite (up to 60%), decreased phyllosilicates, quartz and plagioclase; intercalations of Mendez marls have the same composition. 2) The thick sandy limestone layer (S.L.L.) within unit 1 differs from these sediments by showing lower calcite, but higher quartz, plagioclase, chlorite and illite (e.g. El Peñon, Figs. 2a & 3). This suggests distinctly different detrital influxes during deposition of the S.L.L. and the loosely cemented spherule-rich sediments above and below.

Unit 2: laminated sandstone

Whole rock and clay mineral contents of unit 2 are more regular than in unit 1 and characterized by low calcite (<40%) and phyllosilicates (19%), but higher quartz (31%) and plagioclase (albite >16%, Figs. 2a, 2b). Clay-minerals consist of chlorite, mica (illite) and illite-smectite mixed layers, chlorite-smectite mixed layers being almost absent (e.g. El Peñon, Fig. 3). These mineralogical data indicate a significant increase in detrital influx and more rapid deposition relative to the underlying unit 1 as well as to the overlying unit 3.

Unit 3: Sand-silt-shale layers

Whole rock and clay mineral compositions of unit 3 are highly variable (Fig.2a, 2b). Two distinct clay mineral associations can be identified (Fig.3). 1) The rippled sandy limestone (RSL) at the top of unit 3 and the sand layers within it are characterized by high chlorite (39-41%) and mica (35-42%), and suggest increased detrital influx and probably more rapid deposition similar to unit 2. 2) In contrast, the marly shale layers are enriched in finer chlorite-smectite (11%) and illlite-smectite irregular mixed-layers (up to 70%, <2µm size fraction, These shale-mineral associations are similar to those of the Mendez marls and indicate periods of normal hemipelagic sedimentation during deposition of unit 3.

Granulometric data also reflect the sand-silt-shale layers by the alternating coarser and finer grain-size pattern (see Fig. 13 of Chicxulub Debate Round II, part B: Keller, Adatte, Stinnesbeck). The small size of the clay minerals within the finer grained layers precludes settling through the water column within a few hours after a tsunami wave, as Smit suggests.

Zeolite layers indicate volcanic influx

Two distinct layers enriched in zeolites (clinoptilolite-heulandite) are recognized near the base and top of unit 3 in all sections examined (Fig. 4). Additional zeolite-enriched layers associated with smectite are also observed in unit 1, as well as in the underlying late Maastrichtian Mendez marls and the early Tertiary shales of the Velasco Formation. These different zeolite enriched layers are correlatable from section to section over a distance of more than 300 km. An in situ-diagenetic origin of these zeolites is unlikely because of their geographic distribution and excellent correlatability in different lithologies, such as sands, silts, shales and marls (Fig.4). These layers are therefore detritical in origin and indicate discrete periods of volcanoclastic influx. Their widespread presence within units 1 and 3 is further evidence that deposition occurred over an extended time period that is inconsistent with the impact-tsunami hypothesis.

No evidence for tsunami origin of clastic deposits

The clastic deposits with their lithologically and mineralogically distinct and correlatable units and subunits represent differing flow and depositional regimes with varying detrital influx and rates of sediment deposition. These units, sub-units and even individual zeolite-enriched layers are correlatable on a regional scale and lend no support chaotic deposition as would be expected from tsunami deposition.

Deposition of the spherule-rich layers of unit 1 was relatively rapid with a source primarily from shallow neritic environments, transported into deeper waters (Keller et al., l994). Unit 2 represents erosion and rapid re-deposition of massive sand lenses. The sand, silt and shale layers of unit 3 and their mineralogic variability indicate high detrital influx and rapid deposition (sand and silt layers) alternating with normal hemipelagic deposition (marls-shale layers). Deposition of the rippled sandy limestone at the top of unit 3 occurred slowly enough to permit a thriving benthic community to exist (Ekdale and Stinnesbeck, 1994). Discrete layers enriched in zeolites within unit 3 also indicate discrete volcanoclastic influx. Thus, deposition of unit 3 probably occurred over a longer time period than either unit 1 and unit 2, and under more variable environmental conditions.


The varied detrital influxes of the clastic deposit are likely related to the latest Maastrichtian eustatic sea-level lowstand, which occured during the last 100-200 kyr before the K/T boundary, and the increased terrigenous influx related to the uplift of the Sierra Madre (Laramide Orogeny) accompanied by volcanic activity (Adatte et al., l996). Clastic deposits similar to units 2 and 3 are not unique in the area and have been observed also in the Upper Maastrichitian (Linares Cementery) and Danian Velasco Formation. Only the spherules of unit 1 are unusual and these have been shown to be reworked from shallow waters and redeposited. Morever, up to four spherule layers are present and widespread in in the Mendez marls up to 12 m below the base of the clastic deposit. The spherule unit 1 therefore cannot represent the original Chicxlulub impact ejecta fallout, nor can units 2 and 3 represent impact-generated tsunami deposits.

Figures.1a and 1b Biostratigraphic and lithostratigraphic correlations in eleven K-T boundary outcrops in NE Mexico which contain clastic deposits near the K-T boundary. Note that units 1, 2 & 3 and the Sandy Limestone Layer (SLL) are correlatable over a distance of more than 300km.


Fig.2a and 2b Mineralogical correlation of five K-T outcrops in NE Mexico, based on whole rock compositions. Note that each units is characterized by particular mineralogical features which are correlatable. Note the more variable compostion of unit 3 corresponds to the marly shales and fine sandstone layers. In comparisoon, the sandstone of unit 2 is more homogenous. Note also the different mineralogical composition of the SLL compared to the spherule layers above and below.
Fig.3 El Peñon I, phyllosilicate distribution from the Mendez Formation through the clastic deposit (<2µm size-fraction). Note there are two layers enriched in zeolites in unit 3 (red mark). Also note that the shale layers of unit 3 (purple) and the spherule layers of unit 1 spherule are enriched in illite-smectite and chlorite-smectite mixed layers in unit 1 and in the shale layers of unit 3. The small size of these clay minerals precludes settling through the water column within a few hours after a tsunami wave, as suggested by the impact-tsunami hypothesis. Note also the different clay composition of the Sandy Limestone Layer (SSL, green), which is enriched in mica to the detriment of Iillite-smectite and chlorite smectite.
Fig.4. Correlation of zeolite-enriched layers in seven K-T boundary outcrops in NE Mexico. Note that zeolilte enriched layers are correlatable over a distance of more than 300km.


Adatte, T., Stinnesbeck, W., and Keller, G., l996. Lithostratigraphic and mineralogical correlations of near-K/T boundary clastic sediments in northeastern Mexico: Implications for mega-tsunami or sea level changes? Geol. Soc. Am. Special Paper 307, 197-210.

Ekdale A.A. and Stinnesbeck W. l998. Ichnology of Cretaceous/Tertiary boundary beds in northeastern Mexico. Palaios 13: 593-602.

Keller, G., Stinnesbeck, W., Adatte, T., MacLeod, N and Lowe, D.R., l994. Field Guide to Cretaceous-Tertiary boundary section in northeastern Mexico. Houston, Texas, Lunar and Planetary Institute Contribution No. 827, 110 p.

Keller G. et al. 2002. Multiple spherule layers in the late Maastrichtian of northeastern Mexico. Geological Society of America Special Paper 356: 145-161.

Keller G., Stinnesbeck W., Adatte T. and Stueben D. 2003. Multiple impacts across the Cretaceous-Tertiary boundary. Earth-Science Reviews 62, 327-363.

Slumping more evident in the evidence than in the rocks...

Reply to Smit by G. Keller says Inclined beds no support for impact slumping - posted 22.12.03

Sir, Smit (Discussion page 1) desperately hangs onto his impact-tsunami hypothesis by citing ever more isolated examples of local disturbances observed in some outcrops. First it was a small, localized slump in one outcrop at Mesa Juan Perez. Then it was some small local disturbance on the side of the Rancho Nuevo outcrop adjacent to perfectly stratified sediments. And now it has come down to merely inclined strata below the clastic units along the hillsides of Rancho Nuevo and El Penon (Smit Figures 1 and 2, below).

Inclined bedding on hillsides is common in sediments of any age and is generally due to local tectonic disturbance, including uplift, and the rheology of sediments. The late Maastrichtian sediments that make up the hillsides in northeastern Mexico are no exceptions. Any rocks deposited more than 65Ma. ago will have undergone some degree of tectonic disturbance and/or uplift. Hence, one would expect that not all strata remain horizontally layered. Certainly, invoking an impact-tsunami to explain the presence of such a common feature, which is present throughout the geological record, is rather extreme and peculiar.

Smit's Rancho Nuevo figure 1 shows perfectly undisturbed layering of 'sandstone' beds, except that they are inclined below horizontal beds of the clastic deposit, which Smit interprets as tsunami deposits. (In debates round I and II of Keller et al. we have shown that based on mineralogical evidence (see also Adatte this debate) and several horizons of bioturbation, these deposits could not have been deposited over a period of hours to days by a tsunami event, but rather represent long-term deposition.) In Figure 2 Smit shows one of the most poorly preserved of many outcrops in the El Penon area and then argues that the sediments are similarly inclined versus the horizontal overlying clastic unit. No inclined bedding is apparent in these photos, but we take Smit's claim at face value.

Smit's explanation that these inclined beds are due to 'gliding, slumping, fluidization' due to the impact-tsunami event is rather curious. By the same interpretation he has the overlying horizontally layered clastic deposits laid down within hours to days by the same tsunami event. Yet, these clastic 'tsunami' beds erosionally truncate the underlying inclined beds of the Maastrichtian Mendez Formation, as evident throughout the region. In fact, this is a classic textbook example of original deposition followed by erosion and deposition of the horizontally layered clastic sediments. The minor disturbance is linked to tectonism and the rheology of marls, whereas the more resistant sandstones are differently affected. During the late Cretaceous thrusting of the Sierra Madre, such minor tectonic features as the general inclination of the clastic beds and the local disturbance of the underlying marls, are quite common without any help from a tsunami.