Longhorns on the Mullinax Ranch where the drilling for the K-T and Chicxulub impact ejecta took place.

In March 2005, we drilled the Cretaceous-Tertiary (K-T) boundary sequences in two localities along the Brazos River, Falls County, Texas. The project was organized by Professor Gerta Keller (PI), Geosciences Department, Princeton University, and funded by NSF Grant 0447171. Drilling was done by DOSECC and Logging by Schlumberger.

The onsite scientific team consisted of Gerta Keller and Jerry Baum with Tom Yancey aiding in core removal, splitting of cores, photographing and description.  Thierry Adatte, University of Neuchatel, Switzerland, participated in pre-drilling fieldwork and post-drilling core description, sampling and analyses. Additional team members include Sigal Abramovich, Ben Gurion University, Israel, Michael Prauss, Berlin, Germany, Doris Stueben and Zsolt Berner, Karlsruhe University, Germany, Silvia Gardin and Bruno Galbrun, Paris, France. 
 

Fig. 1. The onsite scientific team: Jerry Baum, Gerta Keller and Tom Yancey. Thierry Adatte, University of Neuchatel, Switzerland, participated in pre-drilling planning, fieldwork and post-drilling core description, sampling and analyses.

       


Drilling location
The cores were drilled along the Brazos River in Falls County, Texas (Fig. 2). In this locality are numerous small exposures of late Cretaceous and early Paleogene rocks, which have been previously studied by many workers over the past 20 years. The outcrops have limited exposures with the best ones along the Brazos River and in the river bed itself. The latter is only accessible during the dry season when the riverbed is dry, which hasn’t happened in eight years. We drilled two localities about 3 km apart (Mullinax-1 and Mullinax-2 & 3). Mullinax-1 was drilled at the same location as the old core (KT1) drilled by Thor Hansen in the middle l980s (and recently published on by Schulte et al., 2006). Mullinax-2 and 3 are overlapping cores drilled near the Darting Minnow Creek waterfall. In between these locations is the Cottonmouth Creek with the old core KT3 of Hansen (published on by Keller, l989) and various outcrops along the creek banks.
 


Fig. 2. New and old drilling localities and outcrops along the Brazos River, Falls County, Texas.

         

Drilling Objectives
  1. Age of Chicxulub impact: The Brazos River area was chosen to test the current controversy over the age of the Chicxulub impact: whether this impact was the K-T killer as commonly believed or predated the K-T boundary by 300 k.y., as suggested by the new Chicxulub crater core Yaxcopoil-1 and outcrops in NE Mexico (Keller et al., 2003, 2004).
  2. Biotic effects of the Chicxulub impact: It is commonly believed that the Chicxulub impact caused the K-T mass extinction. But mounting evidence from NE and Central Mexico and the Chicxulub impact rater core Yaxcopoil-1, indicates that this impact predates the K-T by as much as 300,000 years. The biotic effects of this impact can be tested by analyzing the sediments above and below the impact ejecta layer in the Brazos cores and outcrops. The biotic effects of this shallow water area can then be compared with the deep water sequences of NE Mexico.
Rational for Choosing Texas
The Brazos River area was chosen because it provides a simple and inexpensive test of the previous results from Mexico and the Chicxulub impact crater core Yaxcopoil-1 by simply drilling a couple of 100 feet deep holes across the K-T boundary. In addition, his particular area was chosen for: a) it’s complete stratigraphic sequences, which are comparable to the K-T boundary stratotype section at El Kef, Tunisia; b) the location about 1000 km from the impact crater (and 25 miles from the Bush ranch), c) excellent microfossil preservation, and d) the presence of both the K-T boundary and the “event deposit” with Chicxulub impact ejecta spherules in widely separated stratigraphic intervals. These attributes mark the Brazos River area as the most important K-T impact locality outside Mexico and critical to resolving the current controversy regarding the age of the Chicxulub impact and its potential kill-effect.Drilling OperationDrilling of three 75-100 feet deep holes was done by DOSECC (the same company that drilled the Chicxulub impact crater hole Yaxcopoil-1) with a CS-500 rigand was completed in four days of nearly round-the-clock drilling (Fig. 3). On site, the cores were visually examined, labeled and recorded. The cores were then transported Texas A&M, where they were split, the lithology described and photographed for the record. In the field, A Schlumberger kindly contributed the logging of the holes.

 
       


Fig. 3. Drilling by DOSECC of cores Mullinax-2 and Mullinax-3 near the Darting Minnow Creek of Falls County, Texas.

 

          
       

Fig. 4. Gerta Keller and Jerry Baum inspecting drill cores of Mullinax-3.

           

Study Approach
The cores and outcrops are being studied by a team of paleontololgists, sedimentologists, mineralogists and geochemists. A sampling session was convened at Princeton University in May 2005 attended by the international project team including scientists from the Germany, France, Switzerland and Israel. The cores were sampled for biostratigraphy and quantitative faunal studies, stable isotopes of bulk rock, organic carbon and individual foraminiferal species, and palemagnetic analysis. Trace element contents are analyzed by ICPMS to evaluate contributions from the Chicxulub impact, possible volcanism in the area and the influx of biolimiting trace minerals and their effect on faunal and floral assemblages. Samples were also taken for bulk rock and clay mineralogy, TOC and granulometric analyses. Geochemical analysis of impact glass spherules and altered glass are performed to evaluate the origin of the ejecta material.

Preliminary Results from DrillingEvent Deposit with Chicxulub impact Spherules:The event deposit (also called tsunami deposit by some workers) is the most prominent feature of the Brazos sequences and has variously been described as impact tsunami deposit (Bourgeois et al., l988), Chicxulub impact tsunami deposit (Schulte et al., 2006), or storm deposits associated with a sea level lowstand (Keller, l989; Yancey, l996; Gale, 2006). The event deposit is highly variable in the Brazos outcrops and ranges from absent to 2 m thick.  Individual features may vary from deposit to deposit as a result of variable erosion. A sea level lowstand and erosion surface marks the base of the event deposit. The first deposition above it is usually a boulder conglomerate, which however is not present in the Mullinax-1 core. Here, the lower part of the event deposit consists of a sandstone rich in altered Chicxulub impact spherules, glauconite, phosphate, broken shells, reworked foraminifera and clasts from the underlying sediments. Three upward fining spherule-rich units are present. The lower two units grade into fine sandstones at the top; the third unit is truncated by erosion. Above the spherule-rich units is a hummocky cross-bedded sandstone, which is strongly burrowed in vertical and horizontal directions and infilled with dark mudstone, which must have originated from once overlying sediments that were subsequently eroded (Fig. 5).  Burrows are truncated by erosion at the top. Overlying this erosion surface are laminated clayey siltstone and mudstone layers. Above these, fine, silty mudstone shows upward fining grain size and marks the end of the event deposit with a gradual transition to normal sedimentation (Fig. 5).             

Age of Event Deposit: The event deposit is within the late Maastrichtian planktic foraminiferal zone CF1 (Plummerita hantkeninoides), which spans the last 300,000 years of the Maastrichtian. The base of this zone is about 1 m below the event deposit, suggesting that deposition occurred in the lower part of the zone CF1 interval.

     


Fig. 5. Lithology, description and interpretation of the event deposit in core Mullinax-1 along the Brazos River of Falls County, Texas. Note the three upward fining glauconitic spherule-rich coarse sandstone units topped by fine sandstone layers and the truncated burrows in the hummocky cross-bedded sandstone. These are diagnostic features of multiple storm deposits.

 


Interpretation: The event deposit shows a complex depositional history of multiple storm (e.g., tempestites) deposits, burrowing, and erosion. Deposition occurred in an incised valley and began during the initial transgression after a sea level lowstand that marks the erosion surface at the base of the event deposit (Fig. 5). The three upward fining spherule-rich layers may reflect tempestites, eroding spherule-rich deposits exposed in nearshore areas by the low sea level and transporting and depositing them offshore in incised valleys. Deposition of the event deposit may have occurred over a considerable time interval as suggested by the presence of burrows, which are truncated by erosion. Gale (2006) observed 3 to 5 burrowed and truncated layers within the event deposit in the Brazos River bed exposures and interpreted these as storm events.
   


Fig. 6. Illustration of sea level changes and depositional events along the Brazos River, Texas. The event deposit coincides with a very low sea level, which caused repeated erosion of sediments and transportation and redeposition primarily in incised valleys.

 

Event Deposit to K-T Boundary:
Between the event deposit and the K-T boundary are upward fining silty mudstones (first 15-20 cm), which mark the end of the event deposit. This is followed by 80 cm monotonous laminated shales and mudstone with shells and burrows infilled with pyrite (Fig. 7). This 80 cm interval contains exclusively late Maastrichtian microfossils, late Maastrichtian carbon-13 signals, and the same clay mineral and bulk rock contents as in the late Maastrichtian below the event deposit. Planktic foraminiferal species show the same relative species abundances as below the event deposit.
     


Fig. 7. Interval from the event deposit to the K-T boundary at Mullinax-1. Note that there is 80 cm of normal sedimentation with all the characteristic of the late Maastrichtian above the event deposit. This indicates that the event deposit predates the K-T boundary.

        

Interpretation: The 80 cm ofstratigraphic separation of the K-T boundary and event deposit (including the upward fining interval), the return to pre-event depositonal characteristics, including carbon isotope values and similar late Maastrichtian planktic foraminiferal assemblages provides unequivocal evidence that the Chicxulub impact predates the K-T boundary mass extinction.

The K-T Boundary
The boundary is identified based on the characteristic K-T carbon-13 shift and the coincident faunal and floral turnovers from the Maastrichtian to the Danian. These two characteristics identify the K-T boundary mass extinction globally. In Mullinax-1 and all other Brazos area sections, there is no significant lithological change at the K-T boundary and no evidence of significant environmental changes.

The K-T boundary at Brazos does not show the abrupt mass extinction pattern known from the deep sea where all large complex tropical species (2/3 of the assemblages) disappeared at or before the K-T boundary. At Brazos, species diversity is very low even below the event deposit (~28 species as compared with ~60 in deep water environments) and decreases gradually to <20 species by K-T time. This gradual decrease and low diversity reflects the high biotic stresses of very shallow water environments and the exclusion of deeper dwelling species. At the time of the sea level lowstand at the base of the event deposit, water depth at Brazos was no more than 30-50 m and may have been even less than 20 m in some areas. Species which could not adapt to this shallow environment gradually disappeared (red lines mark extinctions, Fig. 8). Thus, it is the shallowing environment, rather than an impact event, that caused the gradual decrease in species diversity.

 The smaller and environmentally more tolerant survivor species (blue in Fig. 8) show little adverse effects until the K-T boundary, when they rapidly decrease and disappear, except for the disaster opportunist Guembelitria species, which thrived. This pattern is consistent with the deep sea.  Evolving Tertiary species also show the same pattern as in the deep sea. Thus, the shallow water survivors and evolving Tertiary species of the Brazos area show the same biotic response to the K-T event as in sequences globally. 

However, there is no significant assemblage change associated with the event deposit, except that a few deeper dwelling species disappeared by the time of the sea level lowstand.

   


Fig. 8. Relative abundance of planktic foraminiferal species across the event deposit and the K-T Boundary at Brazos core Mullinax-1. Note that there is no mass extinction at the event deposit. A gradual decrease in species diversity begins well below the event deposit and culminates at the K-T boundary.

 

  

Preliminary Results from Fieldwork
More evidence of the pre-K-T age of the Chicxulub impact was found in nearby outcrops of the Cottonmouth Creek and Darting Minnow Creek tributaries of the Brazos River. Here the event deposits are exposed beneath gentle waterfalls that drape over the resistant sandstones. The event deposits are highly variable in thickness and lateral extent and all thin out and disappear over tens of meters. The thickest deposit is found at the Darting Minnow Creek where about 1.8 m is exposed (Fig. 9). At this locality, the K-T boundary is located above the waterfall, but difficult to reach in the creek bed.

   



Figure 9. Darting Minnow Creek event deposit with Chicxulub impact spherules near the base (boots of G. Keller are on Chicxulub spherule deposits). Thin-bedded sandstones overlie the glauconite, phosphate and spherule-rich deposits. The K-T boundary is above the event deposit, but difficult to access in the Creek bed.

   

In the Cottonmouth Creek, the event deposit is relatively thin at about 30 cm. As in the Darting Minnow Creek, a waterfall drapes over the resistant sandstone (Fig. 10). At this locality we discovered 3 cm thick yellow clay layer about 45 cm below the base of the event deposit.  This yellow clay consists of 100% Mg-smectite, which is a cheto type smectite that is characteristic of altered Chicxulub impact glass. This is also shown by the geochemistry of relic glass fragments, which reveal a nearly identical composition as Chicxulub crater impact glass. Thus, the yellow clay layer may represent the original Chicxulub ejecta spherule layer, which is now almost completely altered to cheto smectite. This clay layer was deposited at the base of zone CF1, about 300,000 years before the K-T boundary mass extinction. The reworked spherules at the base of the event deposit may have been eroded from nearshore areas during the low sea level and transported via the incised valleys where they now rest.
   


Fig. 10. Cottonmouth Creek waterfall over the event deposit with reworked Chicxulub impact spherules. The original Chicxulub ejecta layer was discovered in a yellow clay layer 60 cm below the base of the event deposit. The yellow clay represents a cheto smectite clay consisting of altered Chicxulub impact glass spherules.

 

    


Reworked clasts with Chicxulub spherules
At the base of some event deposits in outcrops is a conglomerate bed with transported clasts and boulders described in Yancey (l996). Some of the rounded clasts contain an amazing history of the Chicxulub impact spherules. Inside some of these clasts, shown in Fig. 11A & B, one can find other clasts with spherules cemented in between these. They indicate a previous history of deposition and lithification, which was then followed at a later time by erosion and re-deposition at the base of the event deposit. In some clasts, spherules are scattered in fine lithified mudstone (Figs. 11C & D). In other clasts, spherules infill cracks within the host sediment prior to lithification, erosion and transport to the base of the event deposit where they now rest (Fig. 11E & F).
 


Fig. 11 A, B: Clasts from the basal conglomerate of the event deposit contain Chicxulub impact spherules. C, D: spherules in mudstone clasts. E, F: spherules within cracks of mudstone clasts. E: cracks rimmed by sparry calcite. Insert shows morphology of crack and total length of ~2 cm. F: clast with cracks infilled with spherules and sparry calcite, then truncated by erosion and followed by normal sedimentation. These clasts reveal a history of Chicxulub ejecta fallout and lithification well prior to exposure to erosion, transport and redeposition at the base of the event deposit.

 

Summary of Major Findings:

  1. The K-T boundary was recovered 80 cm above the upward fining of the event deposit in core Mullinax-1. The boundary was identified based on the characteristic K-T carbon-13 shift and the coincident faunal and floral turnovers from Maastrichtian to Danian assemblages.
  2. The ~100 cm interval between the K-T boundary and the event deposit contains exclusively late Maastrichtian microfossils, late Maastrichtian carbon-13 signals, and the same clay mineral and bulk rock content as in the late Maastrichtian below the event deposit. This indicates that deposited occurred during the latest Maastrichtian and that the event deposit is therefore a late Maastrichtian event.
  3. The event deposit (also commonly called impact-tsunami deposit) consists of discrete and separate depositional events separated by truncated burrows and changes in depositional regimes. This indicates that the event deposit was deposited over an extended time period and was probably caused by multiple (seasonal?) storm events.
  4. The Chicxulub spherules in the event deposit are reworked from an older deposit. This is apparent by the abundance of glauconite, phosphate and clasts from the underlying sediments.
  5. Clasts containing cemented clasts and spherules reveal an amazing history of Chicxulub spherules deposition at an earlier time, with subsequent erosion, lithification, clast formation, transport and redeposition at the base of the event deposit.
  6. A yellow altered glass layer about 45 cm below the event deposit may represent the original Chicxulub impact ejecta layer. This layer consists of smectite (Cheto) with rare relic glass still present with geochemistry consistent with Chicxulub impact glass.
  7. The age of the yellow clay layer is indicated by planktic foraminifera as near the base of zone CF1, or about 300,000 years prior to the  Chicxulub impact.
  8.  These Brazos results are consistent with the earlier results from NE Mexico and the Chicxulub crater core Yaxcopoil-1 (Keller et al., 2003, 2004), which indicate  that the Chicxulub impact predates the K-T boundary by about 300 ky.

Preliminary Conclusions

The new Brazos cores and outcrops reveal critically important K-T boundary sequences that help unravel the mystery of the K-T boundary mass extinction and the role the Chicxulub impact may or may not have played in the demise of the dinosaurs. Together with the many outcrops in NE Mexico and the Chicxulub impact crater core Yaxcopoil-1, there is now very strong evidence that in all three localities the impact ejecta layer predates the K-T boundary mass extinction.

In the Brazos sequences, as in NE Mexico, there is a clear stratigraphic separation between the K-T boundary and the event deposit, as well as between the latter and the underlying original Chicxulub impact ejecta layer. This sequence of events refutes the two popular and generally accepted hypotheses that the Chicxulub impact caused the K-T mass extinction and that the underlying event deposit represents a tsunami deposit generated by the Chicxulub impact.

The new evidence points to a more complex multi-event scenario for the K-T mass extinction with a successive series of volcanism, climate, sea level and impact events over a relatively short time interval (500 ky) eventually leading to the demise of specialized large species at the K-T boundary. No species extinctions were caused by the Chicxulub impact and other biotic effects are still to be determined.