Saturday, April 28, 2012

Glauconite, mosasaurs, and asteroids: does the taphonomy of a bonebed in New Jersey really indicate preservation of a catastrophic assemblage?

During the course of my master's thesis research, I attempted to find and read every paper ever written on the taphonomy of marine vertebrates so that I could not only get a good idea of what ideas had stood the test of time and which ones hadn't (or for that matter, were profoundly embarassing even to read), but also so that I could identify which areas in marine taphonomy that we had no answers or even hypotheses for yet. For example: we know extremely little about the everyday processes on the continental shelf that affect fossil preservation, from an actualistic perspective. I found that we have two areas where we do in fact have quite a bit of actualistic data: the beach (where we can walk around and poke dead stuff) and the offshore environment (i.e. outer shelf through abyssal plain - where we can send down a submersible to go and poke dead stuff). Everything in between - from ~10 through ~100 meters water depth - we have very little data, except for a couple of failed whale fall experiments. There is now quite a lot of data generated through whale fall studies for the deep sea, and plenty of actualistic taphonomists have walked around and watched dead vertebrates decay, disarticulate, and have their skeletons dispersed across the seashore. The likely reason we don't know much about the shelf is because in the intervening environments, currents and sediment transport processes likely cover/uncover/transport carcasses or bones too quickly to make a submersible visit really feasible. Using submersibles is damned expensive - and even scuba diving is expensive as well (let alone dangerous), and quite frankly, there probably aren't enough people in marine vertebrate taphonomy who care enough to try and get those sorts of funds. It is quite a bit easier to look at the fossil record in continental shelf settings and go from there and what we know about shelf sedimentation - which is exactly what my master's thesis focused on.

Even though I finished my thesis about a year ago now, I'm still quite keen to learn more and read more articles - and there have certainly been a flurry of articles recently, which I'll have to cover later on. Not only because I am quite obsessed with the subject of taphonomy, but also because I still haven't published my thesis yet, and will need to update it with citations to some of these new studies. If all goes well, I'll be presenting at SVP this year on my thesis, and hopefully will be putting a shorter version of it together for submission to "Geology". That all being said - I was quite interested to see this paper come up in a search on Georef: "Obasi et al. 2011. Glauconite composition and morphology, shocked quartz, and the origin of the Cretaceous? Main Fossiliferous Layer (MFL) in southern New Jersey, U.S.A. Journal of Sedimentary Research 81:479-494. Reading the abstract, the authors had apparently used a combination of glauconitic mineralogy, mineralogy of shocked quartz, and taphonomic evidence to identify a marine vertebrate bonebed in New Jersey as representing a catastrophic assemblage - and not just any catastrophic assemblage, but at the K/Pg boundary (for the unitiated - formerly K/T, meaning Cretaceous-Tertiary boundary) and representing a mass death assemblage from the end Cretaceous extinction.

Map of K/Pg strata in New Jersey. From Obasi et al. (2011).

Now, keep this in mind as you read: I do not study catastrophic extinctions, and I generally do not read literature on the K/Pg extinction unless (as in this case) it deals with taphonomy. Otherwise, given my background- the entirety of my reading focuses on the Cenozoic; but with taphonomy, there have been marine vertebrates dying and being moved around by sediment and other phenomena for nearly half a billion
years. Taphonomy has, in the past, been used as a means to an end to try and substantiate catastrophic hypotheses, and has even been used in weird ways by young earth creationists. So when I read a taphonomy article that does not seem sober, I immediately become skeptical. My last point - I have read a lot of other articles about catastrophic assemblages and their taphonomy, and very few ever really make a great case (especially in marine rocks). So it's not that I don't believe in catastrophic assemblages per se - it's that I haven't read very many good articles on them. They certainly do exist, but there is a substantial volume of evidence you need to provide in order to fully eliminate alternative hypotheses. In the past, people have identified many criteria for identifying catastrophic assemblages - are the fossils all preserved on the same bedding plane? Is there no evidence of post-mortem destruction or transport of bones? Can the fossil assemblage reasonably attributed to one event of short duration? Are the fossils well-preserved and reasonably complete (i.e. articulated or associated)? Is the assemblage biased, and can it reasonably be interpreted as a death assemblage? These are all questions that must be asked. I've probably forgotten a few, and there are certainly others that other researchers have demanded as requirements.

Stratigraphic column of the late Cretaceous and early Paleogene of New Jersey. From Obasi et al. (2011).

Now, to get to the paper. The authors studied the Main Fossiliferous Layer (MFL), which is a bonebed in the lowermost Hornerstown Formation, which is a 10cm thick concentration of invertebrate and vertebrate skeletal elements, including fish, sharks, mosasaurs, and marine birds. These fossils co-occur with Paleocene microfossils. Previous studies (cited within Obasi et al. 2011) have identified the MFL as a major sedimentary hiatus or unconformity (and even as a sequence boundary), and the fossils from the MFL as being reworked from Cretaceous strata into the Paleocene. There has not been a consensus with the placement of the K/PG boundary within the local section - either at the base of the Hornerstown, within, or above the MFL (referenes within Obasi et al. 2011). The Hornerstown Formation and underlying New Egypt and Navesink Formations are very rich in a mineral called glauconite, and the sediment can probably be characterized as greensand (more on this later). Interestingly, within the MFL, are bivalves infilled with sediment from underlying strata, gray rip-up mud clasts, shocked quartz grains (within a single burrow below the MFL). The MFL and associated underlying and overlying strata are also intensely bioturbated. Lastly - vertebrate fossils from the MFL occur as isolated bones and teeth which are occasionally abraded, as well as partially articulated and associated specimens.

Seeing as the author's primary contribution is in the form of mineralogical analysis of glauconite, I'll discuss that first. Glauconite is an authigenic mineral that forms as small sand-grain size pellets, often around fecal pellets or foraminifera, and other tiny organic elements. I am no mineralogist, but if there is one thing I have consistently read in nearly every publication regarding glauconite - it forms during periods of very low sedimentation or during sedimentary hiatuses, often in concert with phosphogenesis (formation of phosphate nodules), and it generally forms on the continental shelf in upwelling regimes. Glauconite often occurs in marine bonebeds, hiatal surfaces, transgressive surfaces of erosion, and sequence boundaries. Using geochemistry and exhamining the morphology of glauconite grains, Obasi et al. (2011) determined that the "maturity" of glauconite grains (i.e. features indicating how well-formed and long the glauconite was able to form for) increased up section through these strata that contain the MFL. Maturity - FYI - is a term applied to mineral grains in sedimentology (and probably other fields in geology); for example (the most widely known use of 'maturity'), the relative maturity of a sandstone is identified based on what proportion of the rock is composed by quartz (the most durable and stable type of grain in sandstone) or not, as opposed to less-stable rock fragments and feldspar grains. Less mature sandstone includes more unstable fragments and grains and less quartz, indicating it is 'fresher' and has not been subjected to as much weathering. The same concept applies to glauconite - mature glauconite has had a long residence time on the seafloor, and vice versa. Lastly, the authors make the case that the sedimentation rate - while already low to form glauconite at all - must have decreased even further through time during "Hornerstown" time.

The trenches, stratigraphic sections, glauconite (middle two rows) and shocked quartz (lower right) from Obasi et al. (2011). Note the green color of the sediment in A and B.

The authors then discuss the formation of the MFL and what processes could have led to its genesis. They begin by trying to evaluate whether or not the MFL represents a transgressive lag deposit. A transgressive lag, for those of you not versed in the arcane ways of sequence stratigraphy - forms during a period of sea level rise; during this time sediment becomes trapped within estuaries (which are backed-up rivers common during periods of sea level rise, whereas deltas are more common during periods of sea level fall), and sedimentation slows down on the continental shelf. The fairweather and storm currents that normally happen on the shelf then erode and rework older sediment instead of transporting around and depositing new sediment that would normally be shed off the continental margin. This erosion can rework fossils from multiple stratigraphic levels into a single bonebed/shellbed (but does not always happen). It is worth noting that transgressive lags are typically identified based on their position within a stratigraphic sequence, in addition to features indicating erosion or lack of deposition. Obasi et al. (2011) argue that because vertebrate fossils do not appear to be reworked based upon analysis of rare earth element concentrations in bones (published in Staron et al. 2001), the presence of partially articulated marine reptiles, the lack of primary sedimentary structures related to sedimentary reworking, and the fact that the MFL does not co-occur with the slowest sedimentation rate (as determined from glauconite maturity). They do acknowledge, however, that there are reworked invertebrate fossils within the MFL, supporting the identification of it as a transgressive lag. They also argue that the middle-outer shelf setting (inferred from glauconite presence) is too deep for a transgressive lag to form and state "Transgressive lags form as the surf
zone and wave base migrate inland during transgression".

The authors then discuss the possibility of the MFL representing a condensed deposit. A condensed deposit is formed by low to zero net sedimentation, as opposed to erosion, and fossil assemblages may appear fairly similar, and may not be associated with a sharp erosional surface. They admit that the intense bioturbation supports this interpretation, as does the preservation of fossil vertebrates, but the steady increase in glauconite maturity does not-  in theory, it should peak at the time of lowest sedimentation rate, which should be the MFL under this hypothesis. Finally, they discuss the reasons why they interpret the MFL as being a thanatocenosis formed by the K/Pg bolid impact. Although the authors have not found any evidence for the famed iridium anomaly in their study area (Obasi et al. 2011:491), they argue that the presence of shocked quartz grains in a single burrow below the MFL. Although not found elsewhere in the section, the shocked quartz appears to have been reworked form the Hornerstown Fm. down into the burrow, which extends down into the Navesink Fm. (Shocked quartz is often used as evidence of a bolide impact, as it is a high-pressure low-temperature type of deformation that happens to quartz). They argue that the gray rip up clasts are mud rip-ups from the tsunami wave from the bolide impact, and the shocked quartz is evidence of the bolide impact as well coinciding with bonebed formation. In total, they argue that the MFL formed as marine vertebrate carcasses littered their remains across the seafloor after being killed by the event, leaving some isolated bones and some associated specimens.

There are a lot of serious problems with this study, and I'll try to deal with them quickly so that this post doesn't get any longer. First and foremost, transgressive lags can occur in quite deep environments, and certainly glauconite-producing middle and some outer shelf environments, and glauconite has been identified at sequence boundaries and other large-scale sequence stratigraphic surfaces before. Transgressive lags are not confined to the 'surf zone', and storm waves can rework sediment at the depths which glauconite forms at. Secondly - you cannot make the argument that the lack of primary sedimentary structures (erosional surfaces, cross-bedding, etc.) indicates anything if you also note that the strata are massive and bioturbated (except perhaps that the strata are structureless because they are bioturbated!). Bioturbation is not an alternative mode of sedimentation - it is a process which erases certain physical information and overprints it with biogenic information.

What about the glauconite and taphonomy? We already know that A) glauconite forms during periods of slow sedimentation, B) there are clearly reworked invertebrate fossils, C) some of the vertebrate fossils are abraded, D) the vertebrates are probably not reworked, E) the vertebrates are not concentrated onto a single resolvable bedding plane as a number of clearly complete skeletons. Taken in full, the evidence does not really look that good for a catastrophic assemblage. Relatively little of the negative evidence cited by Obasi et al. (2011) for physically-controlled modes of formation are negative. Instead, nearly all of the evidence cited would positively support a physical control on bonebed formation. My alternative and less hyperbolic interpretation would go something like this: 1) slowdown in sedimentation results in glauconite formation; 2) vertebrates concentrated into 10cm interval due to current reworking of sediment at sea floor surface and transport and abrasion of some bones and disarticulation of skeletons (other elements shed from carcasses over protracted period of time); 3) slow sedimentation allows pervasive bioturbation, deleting the physical sedimentary structures from the strata; 4) K/Pg impact happened sometime before or after, or possibly during. We just don't know when given the evidence, and I don't think that Obasi et al. (2011) really have sufficient evidence to point at any one part of the strat column and say "here!".

Glauconite in the Santa Margarita Sandstone at Limantour Beach, Point Reyes. This is glauconite from a transgressive surface of erosion in California, which apparently should not exist.

Interestingly, Liebig et al. (2007) reported on the taphonomy of a mass assemblage of false killer whales, and found that not too long after the mass death, the skeletal remains were extremely variable in their taphonomic mode and would not be identifiable as a mass death assemblage if it were in the fossil record. Does this mean that an assemblage with variable taphonomy like that can be identified as catastrophic? No - it means in cases like that, we cannot tell, one way or the other. Clearly, there has been some residence time of the fossils on the seafloor - not all are articulated, and some are even abraded, meaning they have been transported around on the seafloor. We also know that there are other cases of many articulated skeletons concentrated onto what appears to be a single bedding plane (Peters et al. 2009) - in the case I'm referring to, a series of basilosaurid skeletons in Eocene strata at Wadi Al Hitan, Egypt-  but sequence stratigraphy identified it as a maximum marine flooding surface. An MMFS is a period of low sedimentation, and in this case it was muddy deposition, and essentially a condensed deposit with a concentration of well-preserved whales on what otherwise would look like a mass death assemblage. So - even in a case where all of the questions I asked above might be addressed positively - you still may not be able to tell the difference. The taphonomic support for the catastrophic interpretation of Obasi et al. (2011) just doesn't really hold up very well.

Edit: I realize that I ran out of steam a bit and neglected to mention some of the most serious reasons why the results of Obasi et al. (2011) do not hold up, taphonomically speaking. For one - the premise that a fossil assemblage deposited in glauconite can be interpreted in any way except that it is extremely time averaged to some degree or another (and thus preserving a physical depositional signal - and any purely biological/catastrophic signal being deleted permanently from the rock record) - is absurd. Glauconite is de facto geologic evidence (i.e. independent of the taphonomic evidence, which in the case of the MFL is just as strong) of time averaging and slow sedimentation. Secondly - what taphonomic characteristics need to be positively addressed in order to accurately demonstrate a catastrophic death assemblage in the marine realm? I guess I will start by asking (and answering) a third question - how many previously published marine tetrapod assemblages can realistically be interpreted to represent a catastrophic death assemblage? I can really only think of one well-published example: the death assemblage of Shonisaurus published by Hogler (1992). There may be a couple of others preserved under similar circumstances - but in this case, all of the Shonisaurus skeletons are articulated, in a small geographic area on a single bedding plane, not preserved in a concentration that appears to be caused hydraulically or by changes in sedimentary budget, etc. Virtually none of these variables are present in the assemblage from the MFL in New Jersey (Obasi et al., 2011), and even many of these have been argued to be of variable utility by taphonomists. 

Lastly, I get the impression that the identification of this exact horizon as the MFL rests solely on the shocked quartz grains in a single burrow. This would correlate the death assemblage with the K/Pg boundary, and voila! There's your cause of death. Once again, here's a perfect example of where correlation and causation may not necessarily be linked, and in this case - there isn't sufficient evidence to link them.


Hogler, J. A. 1992. Taphonomy and Paleoecology of Shonisaurus popularis (Reptilia:
Ichthyosauria). Palaios 7:108-117.

Liebig, P. M., K. W. Flessa, and T. A. Taylor. 2007. Taphonomic variation desite
catastrophic mortality: analysis of a mass stranding of false killer whales (Pseudorca
crassidens), Gulf of California, Mexico. Palaios 22:384-391.

Obasi, C.C., Terry, D.O., Myer, G.H., and D.E. Grandstaff. 2011. Glauconite composition and morphology, shocked quartz, and the origin of the Cretaceous(?) Main Fossiliferous Layer (MFL) in southern New Jersey, U.S.A. Journal of Sedimetnary Research 81:479-494.

Peters, S. E., M. S. M. Antar, I. S. Zalmout, and P. D. Gingerich. 2009. Sequence
stratigraphic control on preservation of late Eocene whales and other vertebrates at Wadi
Al-Hitan, Egypt. Palaios 24:290-302.


Nick Higgs said...

Great post. Really enjoyed reading it!

Robert Boessenecker said...

Thanks, Nick! I thought that the rather limited evidence the authors were using to make a rather controversial claim needed to be covered on here. I will certainly be including more taphonomy on here in the future.

Schenck said...

This is a great analysis. I take it you don't find their (really Staron et al 2011)'s REE anylsis convinving?

It does strike as a rather strange conclusion of theirs that the MFL and everything around it is very slowly accumulating, with sed rates progressively dropping (rather than suddenly) AND that there's this big Meteorite induced Tsunami tearing through the area killing everything (including birds). It's kinda weird, most of the paper reads as strongly supporting the MFL being something like a sequence boundary.

Robert Boessenecker said...

All that the REE analysis indicates that the bones weren't reworked from underlying strata - conversely, that should NOT be taken to mean that the vertebrate component of that assemblage was buried instantaneously, whether separately or all together. REE analysis has no bearing on whether or not bones arrive on the seafloor and sit there, exposed, accruing taphonomic damage over long periods of time. In a situation like that, there's no mechanism to result in a different REE signature between the bone and the sediment. REE analysis is great for testing reworking, but is useless for determining whether or not bones were exposed during a sedimentary hiatus (as the sedimentological and taphonomic data point towards).

I think you're precisely right in saying that the evidence provided suggests that it is a time-rich sequence stratigraphically significant surface, which by itself sheds doubt on interpreting any sort of catastrophic scenario.