Product has been added to the basket

Old Evidence, Partial Interpretation - Part 1

KellerGerta_small.jpgChicxulub Debate Round II: Keller, Adatte, Stinnesbeck respond to Smit.

This is the Keller et al. response to Jan Smit's contribution (his second) entitled "Still smoking".


In our first riposte to Smit we presented detailed empirical evidence on bioturbation (within Smit's tsunami event), multiple impact spherule layers within late Maastrichtian Mendez marls of northeastern Mexico, and late Maastrichtian planktic foraminifera overlying the impact breccia in the new Chixculub core Yaxcopoil-1. This evidence supports a multiple impact scenario with the Chicxulub impact predating the K/T boundary by about 300Ka. But the evidence also disproves the Smit et al. Chicxulub K/T impact-tsunami hypothesis.

Jan Smit responds to our detailed evidence not by presenting much evidence in support of his impact-tsunami hypothesis, but by questioning the integrity of our data and researchers, and by protestations, such as 'I strongly question', 'I strongly contest' or 'I don't believe'. 'I question the existence of so called 'strong evidence' he states. Scientific debates are based on evidence, not by denying the existence of evidence that doesn't fit one's hypothesis. Beliefs belong in the realm of religion.

How strong can an argument for the K/T impact-tsunami hypothesis be when it is largely based on beliefs, spurious interpretations based on very little and selective data, and by ignoring the large body of contrary evidence? Why has Smit not provided any new evidence in support of his impact-tsunami hypothesis? Most of his illustrations, interpretations and sparse data date back to the early l990s. An enormous amount of new data has been collected since then by our team and others. Against this database he argues that his old preliminary interpretations are still valid ? primarily by ignoring or denying the existence of the new evidence.

Here we point out the major flaws in Smit's latest riposte, which essentially consists of re-packaging old interpretations from the early l990s. We address the issues he has raised, although most of these have already been covered in the earlier debate round I, as well as in a riposte on CCNET November 1, 2003. Smit has raised an additional issue regarding the mineralogy of the siliciclastic (his 'tsunami') deposits, which is best answered by Thierry Adatte, the mineralogist on our team, in a separate submission. Mineralogical data also unequivocally show that these deposits are not tsunami events.

Here we address the major problems in Smit's arguments under the following headings (box). This argument is split over four separate web pages.


(A) Multiple spherule layers

  • The myth of impact-generated slumps
  • If not slumps then fluidization due to Chicxulub impact?
  • Fluidized injection of limestone?
  • Bedded Mendez marls and multiple spherule layers
  • Hemipelagic deposition between spherule layers
    • Hemipelagic Mendez marls
    • Hemipelagic limestone in spherule unit 1
    • Marl clasts = limestone layer?!
  • Correlation of multiple spherule layers

(B) Siliciclastic rocks and the myth of impact-tsunamis

  • Channelized siliciclastic deposits
  • Grain size and Tsunamis
  • Burrows and Tsunamis
    • Burrows at top of unit 3
    • Burrows within unit 3
    • J-shaped burrows in units 1 and 3
  • Currents and Tsunamis

(C) Chicxulub Yaxcopoil-1

  • Foraminifera are not dolomite crystals
  • There must be sand in a sandstone
  • Glauconite from impact glass?!

(A) Multiple Spherule layers

1. The Myth of impact-generated Slumps

Smit states that he saw additional spherule layers in the Sierrita outcrop in the early l990s but concluded that these were slumped. The difference between Smit and our group is that he knew the answers based on that one outcrop. When we saw that outcrop with the same ambiguous spherule layers (Keller et al., l994a), we embarked on a long-term investigation of the entire region to investigate their nature and origin. We documented over 40 outcrops with multiple spherule layers in stratified sediments and found only rarely small-scale slumps (e.g. Mesa Juan Perez) in an area spanning over 60km2. We concluded that there is no support for impact-generated regional slumps and that there are up to four spherule layers interbedded in Mendez marls, with all but the lowermost layer reworked. These results are documented in five masters and three Ph.D. dissertations as well as summarized in Keller et al. (2002, 2003a, b). What is on the table for discussion here is this comprehensive evidence versus Smit's field investigations of isolated outcrops (Smit et al., l992, l996).
Fig.1 Figure.1. Rancho Nuevo outcrop, showing undisturbed horizontal stratification of Mendez marls with at least two spherule layers below the siliciclastic deposit.

Smit states that he and his colleagues never found any additional spherule layers below the one at the base of the siliciclastic deposit unit 1 which he considers the Chicxulub ejecta fallout. Nevertheless, he argues that if multiple spherule layers are present, they all must be due to slumps. How does he know? What evidence does he have for such a sweeping statement? His only evidence is a local small slump documented by us in the Mesa Juan Perez area. Such slumps are very rare. Most sections show normal undisturbed horizontal bedding as shown in Figure 1 for Rancho Nuevo.

Smit argues that the Rancho Nuevo section is an example of undisturbed horizontal stratification of Mendez marls below the siliciclastic deposit without multiple spherule layers. But he is wrong. Markus Harting (read this on the Debate Page) observed a second spherule layer in Mendez marls 70cm below the siliciclastic deposit and further work is in progress to investigate the underlying bedded marls for additional spherule layers.

2. If not slumps, then impact-generated fluidization?

Smit implicitly seems to concede that the slump interpretation has no legs to stand on. So he proposes that if slumps did not cause the multiple spherule layers, then it was fluidization as a result of the Chicxulub impact. As an example of this he shows the Rancho Nuevo outcrop, which he claims shows a 'diapir-like' squeezed-upward sequence next to the siliciclastic deposit (see Fig. 2). Apart from an outcrop photo, Smit provides no evidence for this interpretation. One would expect to see some evidence of the mixed and folded sediments, of mixed microfossils as compared with the undisturbed strata below the siliciclastic deposit, sedimentological and mineralogical analyses.
Fig.2 Figure.2. The Rancho Nuevo outcrop of Smit with his interpretation of 'diapir-like' squeezed-upward Mendez marl. Investigations by Harting (see submission to the Discussion Page) found Smit's 'diapir-like' feature to be an artifact of the exposed angle of the channelized deposit.

But even if such evidence were available, it is physically impossible that the same impact-shaking event that created the channel deposit simultaneously squeezed-up sediments from below while leaving the beds right next to it (and underneath the concurrently depositing siliciclastic units) intact.

Markus Harting conducted detailed field and laboratory analyses of this Rancho Nuevo outcrop and found no evidence of fluidized sediments, or 'diapir-like' squeezing upward of Mendez strata (see his submission to this debate). He observed that Smit's 'diapir-like' structure is simply an artifact of the exposed angle of the channelized siliciclastic deposit. Hence, rather than fluidized sediments it is a channel cut into Mendez marls.

Smit's second argument for fluidization fares little better. He reasons that since the bedding in the Mendez marls often disappears several metres below the siliciclastic deposits, this is the result of 'fluidization due to ground shaking as a result of the Chicxulub impact.' Again, he provides no evidence for this sweeping interpretation, which can easily be tested by careful field and laboratory analyses (e.g. sedimentological and microfossil analyses).

Our team has conducted such investigations of over 40 sequences and found no evidence of fluidized sediments, nor did we find the bedding in the Mendez marls to disappear. Where bedding does disappear, we found this generally to be due to surface weathering, downslope transport and accumulated slope debris. A little digging below the surface will usually uncover the bedding planes. Since all of these outcrops are on relatively steep slopes capped by the siliciclastic deposits, only the upper steeper parts with constant erosion expose fresh bedding.

3. Fluidized injection of limestone?

Smit still holds to his early l990s interpretation that the sandy limestone layer within the spherule layer of unit 1 (just below the siliciclastic deposit unit 2) is somehow squeezed into this unit by impact-generated fluidization. On the same figure he labels the underlying bedded Mendez marls as 'fluidized Mendez'. We find no evidence of either fluidized limestone or fluidized Mendez marls at El Mimbral. Smit's illustration shows no such evidence. This seems another ad hoc interpretation similar to Rancho Nuevo.

As noted above, there are no fluidized Mendez marls at El Mimbral or in any other of the several dozen sections we investigated. As further evidence of fluidization, Smit describes a folded bentonite in figure d, but it is not apparent on the figure. If there were such a bentonite layer, even if folded, it would be more evidence of normal deposition, than fluidization. At least two bentonite layers are within the siliciclastic unit 3 at Mimbral and these can be correlated throughout the region as shown by Adatte et al. (l996). Such correlatable bentonite layers are further evidence for long-term deposition and against Smit et al's impact-tsunami interpretation.
Figure 3. El Mimbral a larger view of the same outcrop shown in Smit's figure d, showing the continuity of the sandy limestone layer (SLL) within the spherule unit 1. Smit interprets this sandy limestone bed to be squeezed into the spherule layer by impact shaking.

In Smit's figure d, the sandy limestone layer is not labeled, but clearly forms a hard resistant bed, which ends curving up near the end of the channel. For Smit this is evidence of fluidization due to impact shaking. A more expanded view (several metres) of this outcrop bed in Figure 3 shows that the sandy limestone layer is hard and resistant and its curving part is due to small faults near the channel edge. Moreover, the same sandy limestone layer can be traced in outcrops over 300km (Keller et al., l997). In a subsequent argument Smit claims that this sandy limestone layer doesn't exist, and is really just Mendez marl clasts; we will discuss this claim later.

It is inconceivable that this sandy limestone layer (SLL) is due to fluidized impact-injection for many reasons.
  1. This SLL is widespread and can be traced in outcrops over 300km. How could a random fluidizing process produce such a widespread correlatable layer?
  2. How could the fluidized injection be squeezed into the same stratigraphic position within the spherule deposit of unit 1 over an area spanning 300km?
  3. How could the fluidized process produce a normal sandy limestone bed?
  4. An erosional surface marks the top of this sandy limestone bed, which would not be preserved in a fluidized squeezed layer.
  5. There are J-shaped and spherule infilled burrows truncated at the top by erosion in this sandy limestone layer at El Penon (Figure 4). No burrows would be preserved in a fluidized squeezed layer.
  6. Where is the fluidized limestone supposed to come from? There is no such unit in the Mendez marls below or above.
  7. The whole rock and clay mineralogy of the SLL is distinctly different from Mendez marls, or the spherule layer.
  8. There are no spherules in the limestone layer, except at the contacts above and below. If it was squeezed into the spherule layer like toothpaste, why was it not mixed with the spherules? Why is it not broken up?
  9. If it is impact-fluidization, it should have been deposited simultaneously with the spherule layer. Is this a question of which comes first the chicken or the egg?

For all these reasons it defies physics and common sense to interpret this sandy limestone layer as toothpaste-like fluidized and impact injected precisely between two spherule layers that were supposedly deposited at the same time plus or minus a few hours according to Smit's impact-tsunami scenario. Smit has no explanation of where the limestone came from, how this apparently random injection process targeted only the spherule layers over an area of 300km, how the burrows and erosion surface happened to be there? The only reasonable interpretation is the obvious one - namely that this SLL represents normal pelagic sedimentation, including burrowing, and that the spherule layers above and below represent two rapid depositional events (reworking and transport).

4. Bedded Mendez marls & multiple spherule layers

Smit claims that he and his collaborators found no additional spherule layers in sections with bedded Mendez marls below the siliciclastic deposit. In fact, they are easy to find in many outcrops as shown by our team, which documented them in over 40 outcrops. But finding them requires a little digging to get beneath the rubble. Princeton's undergraduates have discovered them in several new sections during their field trips.

Figure 4. A trench at El Penon, which was dug 2-3 feet deep to get to fresh bedrock and through the surface rubble. Three spherule layers were discovered in the10m of Mendez marls below the siliciclastic deposit.

At El Penon spherule layers were discovered interbedded with Mendez marls at 7.7m, 8.5m and 9.5m (see Fig. 5a,b). We consider all but the lowermost spherule layer as reworked. Spherule layers at similar depths within the Mendez marls were discovered at Loma Cerca as shown in Figure 6.
Fig 5a, Fig 5b
Figure 5a, b. Spherule layers at El Penon interbedded in Mendez marls at about 8.5m (a) and 9.5m (b - below) below the siliciclastic deposit.

Smit's example of Rancho Nuevo as an undisturbed bedded Mendez marl without multiple spherule layers is a good illustration of a section where his team could not find these spherule layers, though Harting recently detected them in zone CF1 in careful field and laboratory examinations. This is still further evidence that the additional spherule layers are not slumps, but represent deposition during normal pelagic sedimentation during the last 300Ka of the Maastrichtian (zone CF1).

Of course, there are some sections in which the Mendez marls below the siliciclastic deposits do not contain multiple spherule layers. For example, we didn't find them at El Mulato, though they are present at Rancho Canales only 5km away. Since multiple spherule layers are reworked, one should not expect them to be present everywhere because their presence is controlled by variable erosion and depositional patterns. In addition, there is variable erosion of the Mendez marls below the channelized siliciclastic deposits. We have observed that in many sections all but 1-2m of zone CF1 has been eroded.

5. Hemipelagic sediments between spherule layers.

Smit states 'we have not observed normal hemipelagic sedimentation between spherule layers'? We believe him, especially since he has just argued at length that multiple spherule layers don't exist and that he and his colleagues have not found any in the bedded Mendez marls of any outcrops, even at Rancho Nuevo where Harting documented them, as noted above.
Fig.6 (a) Hemipelagic Mendez marls

Multiple spherule layers interbedded in Mendez marls have been observed in over 40 outcrops in northeastern Mexico, as noted in the previous riposte. The Mendez marls between the spherule layers are clearly bedded, as in Rancho Nuevo, Loma Cerca or El Penon. For example at Loma Cerca spherule layers are observed at 6.5m, 7.5m and 10m below the siliciclastic unit (Fig. 6). Spherule layers 2 and 3 are close together, though separated by an erosional unconformity and each layer contains upward fining spherules.

Figure 6. Lithologic column of Loma Cerca showing the multiple spherule layers separated by up to 6m of normal hemipelagic marl deposition.

We have conducted detailed mineralogical analyses on many outcrops to determine the sedimentary composition for the Mendez marls, spherule units and the siliciclastic deposits (Adatte et al., l996; Stinnesbeck et al., 2001). Here we show the analysis for the Loma Cerca outcrop that illustrates the nature of the hemipelagic Mendez marls and the mineralogically different multiple spherule layers (Figure 7).

Here we show the analysis for the Loma Cerca outcrop that illustrates the nature of the hemipelagic Mendez marls and the mineralogically different multiple spherule layers (Figure 7).
Fig.7 Figure 7. Grain size distribution and bulk mineralogy of the Mendez marls at Loma Cerca support hemipelagic sedimentation interrupted by spherule deposition.

Whole rock compositions (XRD analyses) of the marl below the spherule layers show relatively uniform average compositions (45-50% calcite, 12-18% quartz, 5-9% plagioclase, 25-30% phyllosilicates) that are comparable to those observed in marls between the spherule layers (40-45% calcite, 15% quartz, 5-10% plagioclase and 30-35% phyllosilicates). Similar average Mendez marl whole rock compositions have been observed at El Penon, El Mimbral and many other sections (Adatte et al., 1996).

The high calcite and phyllosilicates contents suggest normal hemipelagic sedimentation under weak hydrodynamic conditions. Quartz and plagioclase are more abundant in the second and third spherules layers, and also in the fourth layer of unit 1, than in the first (and oldest) spherule layer. This reflects the increasing detrital flux in the reworking and re-deposition of spherule layers 2 to 4.

The clay fraction is qualitatively and quantitatively identical in the marls below and between the spherule layers. Fine grain mixed-layers of illite-smectite and chlorite-smectite types (<2µm) are dominant, compared to mica and chlorite.implying normal and quiet hemipelagic sedimentation. The same pattern is observed throughout northeastern Mexico (Adatte et al. 1996).

(b) Hemipelagic limestone in spherule unit 1

A sandy limestone layer (SLL) is present within spherule unit 1 at El Mimbral and many other sections over a distance of about 300km (see Fig. 3, Adatte et al., l996; Keller et al., l997). This SLL contains J-shaped spherule infilled burrows (Fig. 8a), and has erosional upper and lower surfaces. No clasts of any kind have been observed in this limestone and spherules are present, except at the contact with the spherule layers above and below (Fig. 8b).
Figure 8a. Sandy limestone layer within spherule unit 1 at El Penon with J-shaped spherule infilled burrow truncated at the top. The limestone layer indicates hemipelagic sedimentation.

This SLL indicates that spherule deposition within unit 1 occurred in two phases interrupted by long-term hemipelagic sedimentation. Therefore, deposition of unit 1 could not have occurred as part of an impact-tsunami event.
Fig.8b Figure 8b. White sandy limestone layer within spherule unit 1 at El Mimbral.

c) Marl clasts = limestone layer?

Smit argues that the sandy limestone layer at El Mimbral, and presumably in all other localities, is just Mendez marl clasts welded together. To make this argument he shows a photo of large 'welded' marl clasts mixed with spherules - a common occurrence at Mimbral (Fig. 9, Fig. g of Smit). But this photo simply shows that the marly sediments were soft when the reworked spherules from shallow waters were rapidly deposited and mixed into the soft substrate.
Fig.9 Fig. 9. Mendez marl clasts in spherule unit 1 at Mimbral (Smit Fig. g). Smit interprets these marl clasts as being the same as the limestone layer shown in Figs. 3, 8a, 8b.

Smit essentially argues that marl clasts are the same as the contiguous sandy limestone layer, which we show in Figures 8a, 8b and Fig. 3 above. It is hard to take such an argument seriously. In this case, it is especially hard since he also argued that the very same sandy limestone layer was injected by impact-generated fluidization! (see section #3 above). Moreover, in his l992 publication (Smit et al., l992) he argued that this was a well-cemented packstone. The fact is, it is neither. Limestones are not marls, nor are they composed of marl clasts 'welded' together.

Why such contorted interpretations? It appears to be hypothesis-driven rather than driven by the empirical record. If the limestone can be explained as reworked marl clasts, then deposition can be rapid and consistent with the impact-tsunami. But if it is a limestone, then it represents a period of hemipelagic sedimentation separating two reworked spherule deposits, which cannot be related to an impact tsunami.

6. Correlation of multiple spherule layers

Smit claims that based on our Figure 2 (This debate, Round 1) correlation of the multiple spherule layers is not supported between Loma Cerca and Mesa Juan Perez. Of course, we clearly stated that correlation for these particular outcrops is tentative because of limited outcrop exposures. But in the same figure we show the excellent correlation between Loma Cerca and El Penon (see Fig. 10), which are 30km apart. In both sections the lowermost (original) spherule layer is near the base of zone CF1 and even the subsequent reworked layers seem to correlate. This is unlikely by chance, but reflects widespread reworking events.
Fig.10 Figure 10. Correlation of multiple spherule layers in Mendez marls at Loma Cerca and El Penon, which are 30km apart. (Click on image to get a bigger file in a new window)

Smit chose to ignore this correlation and focus his critique on the other sections for which no robust correlation was claimed. Focusing on some auxiliary data while ignoring the main evidence seems to be a recurring pattern in his critiques. But he appears to have no new data to support his assertions.

Reworked spherule layers should not necessarily correlate. If they do, it means that reworking occurred simultaneously at various localities, probably due to a lower sea level leading to erosion from shallow shelf areas and transport into deeper waters. This appears to have been the case as indicated by the presence of shallow water benthic foraminifera and other shallow water debris (wood, leaves), particularly in spherule unit 1 (Smit et al., l992; Stinnesbeck et al., l993). Also, spherule layers do not generally form contiguous sedimentary horizons, but are frequently found in channelized deposits as noted below.
  • Continue reading in Old Evidence, Partial Interpretation - Part 2