Dracovenator by request: the age of the Elliot

Randy and Bill asked an apparently simple question that will require an involved answer. What age is the Elliot Formation? What follows is a detailed account of my thoughts that will probably get to technical for those without a geological bent.
Still here? Great I’ll try to make it worth your while. As I mentioned in the last post the Elliot Formation occurs near the top of the vast sedimentary pile that fills the famous Karoo Basin of South Africa. It underlies the Clarens Formation and overlies the Molteno Formation. Dating the formation is difficult as there is little to constrain it, there are no known ash or lava beds within it that have been radiometrically dated, and the formation lacks any marine microfossils or palynomorphs (spores and pollen) that could be used to date it. The upper constraint on its age comes from the Drakensburg lavas that overlie the Clarens Formation. These represent a rapid pulse of volcanism that is precisely dated to 183 million years ago, which places it at the Pliensbachian-Toarcian boundary of the Early Jurassic in the timescale of Gradstein et al. (2004). Thus the Elliot Formation cannot be any younger the Toarcian-Pliensbachian boundary. At the other end, the constraint on the maximum age of the Elliot Formation is much woollier. The Molteno Formation is said to have palynomorphs of Carnian age, but this is a thick unit and it is not known how much younger the upper parts of the Molteno Formation gets. Furthermore it is now know that terrestrial ‘Carnian’ deposits (e.g. the famous Ischigualasto Formation of South America) actually correlate with the latest Carnian to early Norian stages of the marine sequence. A Carnian/Norian age for the Molteno Formation is also supported by the vertebrate fauna of the Pebbly Arkose Formation of Zimbabwe which is presumed to be a lateral equivalent of the Molteno Formation. This fauna consists of a rhynchosaur and a primitive dinosaur that resembles Saturnalia (which comes from the ‘Ischigualastian’ of Brazil). Thus the Elliot Formation is unlikely to be any older than the Early Norian (about 225 million years old). However this is a huge spread of time, about 42 million years in fact, can we narrow it down any further? Before we examine this question we need to dispel a common model for the deposition of the Elliot Formation which is often portrayed in the literature. This is the model of continuous deposition. This figure (taken from Holzforster 2007) is typical.

In this diagram the Elliot Formation is portrayed as filling the entire block of time between the Molteno and Clarens Formation. However this is not realistic. Deposition in the Karoo Basin was controlled by subsidence in response to tectonic activity in the Cape Fold Belt. Tectonic activity is rarely, if ever, continuously sustained over tens of millions of years. Further the sediments themselves show evidence being deposited in discontinuous pulses. The lower and upper members of the Elliot Formation represent two such pulses. Although no angular unconformity or extensive erosional surface separates these two members there is evidence that there was a non-depositional gap between the deposition of these two units. Firstly we have the wholesale changeover in fauna between the two members. There is no known vertebrate species or genus that occurs in both units. There are also lithological differences that indicate that the style of rivers crossing the floodplain had changed, from sinuous, deep permanent streams to shallow emphemeral braided streams. The latter seems to be coupled with a more arid climate which is also betrayed by other indicators of aridity such as the development of extensive calcareous paleosols, deeply muckcracked overbank horizons, and evidence of floodplain denudation during flash-flood events. Other features such as palaeocurrent indicators suggest that the overall direction of drainage also changed as did the source of the sediment. These features could have suddenly ‘switched’ but a time gap allowing these features to change more gradually seems far more likely.
So if we accept two pulses of deposition what age constraints can we put on them? There is little doubt now that the first pulse was Late Triassic in age. I’m currently preparing a paper describing rauisuchians from the Elliot Formation. They have been reported before but this will be the first time the identification will be based on diagnostic derived characteristics. If we accept that rauisuchians went extinct at the end of the Triassic then these occurrences certainly place the lower Elliot in the Triassic. But where in the Triassic? Geology can help us a little here. According to the model of Catuneanu et al. 1998 and Bordy et al. 2004 the deposition in the part of the Karoo basin where the Elliot Formation crops out occurred during an offloading phase when the Cape Fold mountains were shedding sediment after a mountain building event. During the offloading phase the more distal part of the basin (where the Elliot Formation lies) sags after bulging upward, creating accommodation space for the sediment to collect in.


The reciporical flexural model for basin development, as applied to the deposition of the lower Elliot Formation. Part of a figure from Bordy et al. 2004.

Structural and metamorphic geologists date the end of the mountain building event that is believed to have immediately preceded the deposition of the lower Elliot Formation to 215 million years, give or take 3 million years. This puts the lower Elliot in the mid to late Norian, maybe even extendig into the Rhaetian (depending how long the offloading phase lasted after the mountain building finished) as Randy suggested. Whatever its age I’m willing to bet that the lower Elliot Formation is more or less equivalent to the vertebrate-bearing horizons of the Los Colorados Formation based on similarities of their sauropodomorph faunas. These similarities include Lessemsaurus and Antetonitrus which are very, very similar (though a few telling differences do keep them as separate taxa, e.g. the proportions of metatarsal I). Eucnemesaurus (ex Aliwalia) is also extremely similar to one of the taxa lurking under the label ‘Riojasaurus’.
OK so that’s the lower Elliot, what about the upper Elliot? Here we don’t have ahelping hand from geology. A pulse of mountain building after 215 million years has not been detected (indicating it was a minor event).
Faunally the upper Elliot Formation contains taxa that seem to indicate an Early Jurassic age (e.g. the crocodyliform Protosuchus and a diversity of ornithischians) but this is rather weak reasoning. Nevertheless I think an age younger than the Hettangian one usually assigned to the unit can be supported on the following grounds:
1. The same biozone (the Massospondylus Range Zone)can be found from the beginning of the upper Elliot through to the top of the Clarens. Fossils are extremely rare near the top of the Clarens but Billy De Klerk of the Albany museum has collected a couple of good Massospondylus skeletons, that if I recall correctly, come from near the top of the Clarens. Basically the fauna of the Clarens is a depleted subset of what you find in the upper Elliot (easily explained by the limited sample from the Clarens). The only possible faunal change is that the little trithelodontid cynodont, Pachygenelus, might be replaced by a different trithelodontid, Diarthrognathus, in the Clarens. This suggests to me that we aren't dealing with a large span of time. Dinosaur genera seem to turn over every five million years or so, thus we are probably dealing with a duration in this vicinity for the entire upper Elliot to Clarens sequence.
2. The deposition of the Clarens is terminated by a sudden and precisely dated volcanic event - the eruption of the Drakensburg lavas which is dated to 183 million years.
3. There is no evidence of a time gap between the Clarens Formation and the volcanic eruptions. Indeed there is evidence that one followed the other without a hiatus. For instance in some localities lava flows can be seen to have filled the interdunal spaces in the Clarens dune desert. Clarens deposition seems to have continued after the eruption of the first few flows in some places. A great example can be seen on the road between Barkly East and Rhodes in the Eastern Cape. Lastly there is evidence that the volcanism was begining during the deposition of the Clarens - as is shown by the Clarens filled crater reported by Holzforster (cited in my last post). It would be great to get a date for the pyroclastic flows that form the basal layers of this crater but sadly no-one seems to have done it yet - any young geochronologist looking for a project?
4. So if we accept that the end of the Clarens can be precisely dated to 183 million years AND we only allow a duration of 5 million years or so for the Massospondylus Range Zone then the begining of the upper Elliot Formatio dates to the Early Pliensbachian (about 188 million years). Perhaps we could allow a few million years to have elapsed right at the end of the Clarens, before the volcanic lava flows began spilling out over the basin, and allow a rather long duration for the Massospondylus and its cohabitants but that would still only take us down into the late part of the Sinemurian.

There you have it. My guess is that the lower Elliot is mid-late Norian, there is a hiatus of about 15 million years before the upper Elliot and Clarens which probably span most of the Pliensbachian. How can this be tested? Detrital zircons would be a wonderfull source of data - once again any young geochronologists out there looking for a project?

references

Bordy EM, Hancox PJ and Rubidge BS (2004) Basin development during the deposition of the Elliot Formation (Late Triassic - Early Jurassic), Karoo Supergroup, South Africa. South African Journal of Geology, 107: 395-410.

Catuneanu O, Hancox PJ, Rubidge BS (1998) Reciporical flexural behaviour and contrasting stratigraphies: a new basin development model for the Karoo retroarc foreland system, South Africa. Basin Research 10: 417-439

Holzforster F (2007) Lithology and depositional environments of the Lower Jurassic Clarens Formation in the eastern Cape, South Africa.South African Journal of Geology, 110: 543-560.

Thoughts from the field: Welcome to Lake Drumbo

I want to float some ideas based on some observations I've made in the field this year. So what have I been up to? Well firstly I've been poking my nose further south than I usually do. Most of the sites I've been working over the past half a decade are in the north end of the main Karoo Basin of South Africa. Here the dinosaur bearing Elliot Formation is relatively thin, probably because it lies over the solid Kaapval Craton, a big ancient continental block. This block probably prevented the basin floor from sagging too deeply during mountain building events off to the south (Tectonics is a VERY important factor controlling of deposition in the Karoo Basin). But down south you get off the Craton and the Elliot Formatin becomes thicker...over four times thicker. So with some money from Germany, and a collaborative team from Germany (led by Ollie Rauhut) and Great Britain, we began looking down south (mostly in the Eastern Cape Province). Despite the great thickness of sediment good outcrop is hard to find as the rainfall is quite high and vegetation rather thick.
Another rucurrent problem is figuring out where you are in the stratigraphy, these rocks don't come with labels!
So first a quick primer of the stratigraphy of the dinosaur bearing part of the Karoo.
The top of the sedimentary pile is usually referred to as the 'Stormberg Group' although this is not an officially recognised stratigraphic unit. The Stormberg Group consists of three formations: the Molteno, the Elliot and the Clarens. The Molteno is a series of coarse grained to conglomeratic sandstones, with fine grained siltstones, mudstones and coals in the south. It is rich in plant fossils (and some insects) but has not produced a single piece of bone - although there a some dinosaur tracks reputed to have come from this formation. The Elliot Formation marks the begining of the dinosaur body fossil record in South Africa. It is divided into two units: the upper and lower. At times it can be very difficult to determine in which unit you are in. The Elliot Formation is predominately made of red overbank muds and silts deposited on a humid (lower) to semi-arid (upper) floodplain. It would appear that the Elliot Formation covers quite a time range, with the lower member almost certainly being of Late Triassic age while the fauna of the upper member is very much Early Jurassic in aspect. The Clarens Formation consists of pale cream to white massive cliff-forming sandstones that are aeolian (wind blown) in origin. It records further aridification and the onset of a dune desert. Strangely the fauna doesn't seem to change much, if at all between the upper Elliot and the Clarens. You get pretty much the same taxa in the Clarens, just fewer specimens.
Now here is a parorama of the main valley wall on Upper Drumbo near Barkly East.

We excavated two dinosaur skeletons (Nelly and Charlie) from near the bottom of the Valley. You can see the cap of massive sanstone at the top of the peak in the centre (Castle Rock. This is clearly Clarens Formation. But where is the top of the Elliot Formation and where in the Elliot Formation do our dinosaurs come from?

At first glance the change in slope at the top of the valley seems to mirror the outcrop pattern of the boundary between the lower and upper Elliot Formations.

A colourised diagram of the outcrop of te Elliot and Clarens Formations. Note the change in slope between the two members of the the Elliot Formation. Modified from Bordy et al. 2004.
You can see this even better from a photograph taken from a higher vantage point looking across at Castle Rock, rather than up at it from the valley floor.

In this photo only the thick sandstone bench at the top of the valley can be seen. The smooth ramp-like slope leading up to Castle Rock is obvious. The mountains in the distance are made of the 2km thick pile of basalt that was extruded toward the end of the Liassic. You can read more about them here.
If this was the case then the upper Elliot Formation would form the smoother upper slopes leading to Castle Rock and our dinosaurs would be from the lower Elliot (hence Triassic). BUT one of them is almost certainly a Massospondylus (typical of the upper Elliot) and the sediments that enclose them are also typically upper Elliot in aspect. Perhaps everything from the base of Castle Rock to the valley floor is upper Elliot. Thats a verticle height of 210 m. The upper Elliot reaches thicknesses of 150 m down in the south, but 210m is a bit much. A few observations point me towards what I think is the answer. Firstly you will notice a thin bed of narrowly banded siltstone just below the big sandstones of the valley rim.

I've see a similar bed at the very base of the Clarens Formation in a number of locations (but not all). These are probably a series of playa lake deposits (about the only fossils you will find in them are little mussel shrimps or conchostracans) and they form a pretty good marker for the end of the Elliot.the smooth slope above the valley is not made of red overbank fines like the Elliot Formation. Further evidence that the upper slopes are actually within the Clarens Formation can be found if you actually climb up to them to take a look. The slope isn't made of red fluviatile mudstones, instead you will find thinly laminated pale creamy-grey shales. I've seen this lithology before - in small localised lenses of the Clarens Formation. They are interpreted as small interdune emphemaral pond, or playa lake deposits. However the thickness here (about 80 m) is, as far as I can find out, unprecedented. It seems to me that far from being a dune desert, this part of the world was host to a pretty large long lasting lake. Even more cool is that the same big package of shale can be found in near Blikana and at Rhodes. This may indicate a lake about 100 km across. Strange that I can find no mention of such an obvious feature in the literature. So what lived in or or around it (assuming my interpretation is correct? Bugger all I'm afraid. You won't find even scattered fish bones or scales. My guess is that it was very shallow, hypersaline and prone to frequent drying out. Much like Lake Eyre in the Tirari Desert of South Australia. Indeed given that it is surrounded by a dune desert and the northern section of the Lake is about the same size as the lake I'm suggesting, Lake Eyre makes a pretty good analogue.

Lake Eyre

However there IS a report of a large Lake in the Clarens Formation off to the South West of this deposit. However this lake is an altogether different kind of thing. Holzforster (2007) reports on lacustrine deposits in the Clarens Formation in the Exterem South West end of the outcrop of the Formation. However here the Clarens is incised down into the Elliot Formation, and the lake deposits are underlain be thick pyroclastic deposits. In other words this south-western lake was developed in a volcanic crater. Sadly no fossils have come from here either, - its a pity because a late Early Jurassic Liaoning-style lagerstatte would be sooo cool.

references

Bordy EM, Hancox PJ and Rubidge BS (2004) Basin development during the deposition of the Elliot Formation (Late Triassic - Early Jurassic), Karoo Supergroup, South Africa. South African Journal of Geology, 107: 395-410.

Holzforster F (2007) Lithology and depositional environments of the Lower Jurassic Clarens Formation in the eastern Cape, South Africa.South African Journal of Geology, 110: 543-560.

Field trip photos

I'm currently writing up a report on the lightening quick field trip I took at the end of November for this blog but am getting delayed by all sorts of real-world duties. So for now I'll just show-case some of the pictures.



Matthew Yates assists Charleton Dube and Sifelani Jirah in excavating a dinosaur tail.



The tail of 'Charlie' the sauropodomorph (yes another one!) emerges



Upper Drumbo, type locality of Dracovenator regenti, and site of 'Charlie'. Casle Rock is in the background but where is the contact between the Elliot Formation and the Clarens Formation?.

The Drakensburg Lavas and the First Great Dinosaur Dying

What you are looking at is a thick pile of basalt, that was extruded onto the Earth’s surface some 183 million years ago, during the latter part of the Pliensbachian stage (or at the Pliensbachian-Toarcian boundary, depending on whose timescale you follow) of the Early Jurassic. They are part of a 2 km thick sheet that is centred on the mountainous nation of Lesotho in Southern Africa. They take their name, the Drakensburg Group, from the Drakensburg Range, a ragged row of peaks said to resemble the back of a dragon that runs along the border of Lesotho and the South African province of Kwazulu-Natal. This large pile of basalt is an erosional remnant of a truly enormous volcanic province. Other parts of what was once a continous sheet of lava extend north to Zimbabwe, Botswana and Zambia, and westward into Namibia. What is even more jaw dropping is that if Gondwana is reassembled, then these southern African lavas (generally called the Karoo flood basalts) are just part of one enormous province that extends into the Southern tip of South America (the Chon Aike Province) and across Antarctica (the Ferrar Province) and into southern Australia. Taken together the total volume of magma that was either extruded onto the surface, or emplaced as intrusions below it, would come to more than two and a half million cubic kilometres (Wignall 2001). This volume actually exceeds the estimated original volume of the famous Deccan Traps of India, which were extruded at the end of the Cretaceous (when most dinosaur lineages famously kicked the bucket). Given that these vast volcanic outpourings seem to be linked with mass extinction events with disturbing regularity it seems odd that the truly enormous Karoo-Ferrar province is not linked to a big extinction event – or is it?
An extinction event amongst marine molluscs (yay! see molluscs have much to teach us!) in the late Pliensbachian has been recognised in Europe and South America and this has been tied to the Karoo-Ferrar eruptions (Hallam 1961, Aberhan and Fürsich 1996). But the general consensus is that this was a weak mass extinction, well below the level of the ‘big five’ mass extinctions.
But how sure can we be? One thing that is clear to dinosaur aficionados is that the early Middle Jurassic has an abysmal record of terrestrial faunas and this may well be masking the effects of a terrestrial mass extinction. Indeed the first stage of the Middle Jurassic Epoch, the Aalenian, is the only Mesozoic stage that does not have its own valid, diagnostic dinosaur taxon (or at least it didn’t a few years ago, maybe there is one now). Another thing that the dinosaur record shows is that prior to the middle Jurassic, dinosaur faunas were rather uniform the world over with a community structure dominated by basal sauropodomorphs (usually a massospondylid) with small coelophysid and larger dilophosaurids representing the theropod contingent and much rarer small basal ornithischians. This type of fauna can be found in Southern Africa (Elliot, Clarens and Forest Sandstone Formations), North America (Kayenta, Navajo and Portland Formations), Antarctica (Hanson Formation) and China (Lower Lufeng Formation). It is interesting that the two dominant components of this faunal association, the basal sauropodomorphs and the coelophysids are basically holdovers from the Triassic. However once the record picks up again higher up in the middle Jurassic things have changed a great deal. Gone are the coelophysoids and basal sauropodomorphs*. In their place we find ceratosaurs and tetanurans filling the large predator niches while eusauropods and eurypods (that is ankylosaurs and stegosaurs) occupy the large herbivore niches. This combination of taxa remained dominant around the world to the end of the Jurassic. So was this turnover a gradual affair? Maybe not, and I have suggested that there was actually a terrestrial mass extinction event that cleared away the coelophysoids and basal sauropodomorphs in my so far unpublished chapter in the upcoming Complete Dinosaur II. If so, it would seem very likely that this event was the same one that killed those poor little clams in the late Pliensbachian. In other words the Drakensburg and associated lavas really were significant for dinosaur evolution. Perhaps without them we may never have got such majestic beasts as Apatosaurus and Brachiosaurus. Quite independently Ronan Allain and Najat Aquesbi came to the same conclusion in their monograph on Tazoudasaurus, which I featured here. Ronan and myself must think alike for this isn’t the first time we’ve come up with the same idea more or less simultaneously. Earlier we both published the connection between the dating of the Karoo-Ferrar volcanics and the age of Vulcanodon at more or less the same time (Allain et al.2004, Yates et al. 2004).
Nevertheless there exists an alternative explanation. A team of French geologists led by Fred Jourdan have suggested that the late Pliensbachian extinction event was really mild because the Karoo-Ferrar basalts were extruded over an extended 8 million year period (Jourdan et al. 2005). Other continental flood basalt provinces show a pattern where 90% or more of their volume is extruded in a brief spell of less than 600 000 years. Jourdan et al. clearly demonstrated that the lavas to the north of South Africa were extruded over a period extending from 182 to 177 million years ago. Does this spell the end of the late Pliensbachian dinosaur extinction hypothesis? Perhaps but I’m not ready to discard this idea just yet. Note that the long duration of eruptions is restricted to regions north of South Africa. The Drakensburg (an erosiaonal remnant of a truly vast area shown by the intrusions that riddle the rest of the Karoo Basin) still yields a tight cluster of dates, while palaeomag indicates the whole pile experienced just one magnetic reversal (Duncan et al. 1997). What we need is a comprehensive sampling of the Antarctic, South American and Australian lavas to see whether they also extruded rapidly at the same time the Drakensburg lavas were extruded.

*There is one recorded Middle Jurassic basal sauropodomorph, Yunnanosaurus youngi, but I would like to see a better stratigraphic control on its age.

references

Aberhan M, Fürsich FT (1997). Diversity analysis of Lower Jurassic bivalves of the Andean Basin and the Pliensbachian-Toarcian mass extinction. Lethaia 29: 181-195

Allain R,Aquesbi N, Dejax J, Meyer CA, Monbaron M, Montenat C, Rechir P, Rochdy M, Russell DA and Taquet P (2004). A basal sauropod dinosaur from the Early Jurassic of Morocco. Comptes Rendus Palevol 3(3):199-208

Duncan RA, Hooper PR, Rehacek J, Marsh JS, Duncan AR (1997) The timing and duration of the Karoo igneous event, southern Gondwana. Journal of Geophysical Research 102 (B8): 18127-18138.

Hallam A (1961). Cyclothems, transgressions and faunal change in the Lias of North West Europe, Transactions of the Edinburgh Geological Society 18: 132–174.

Jourdan F, Féraud G, Bertrand H, Kampunzu AB, Tshoso G, Watkeys MK, Le Gall B (2005). The Karoo Large Igneous Province: brevity, origin and relation with mass extinction questioned by new 40Ar/39Ar age data. Geology 33: 745-748.

Wignall PB (2001) Large Igneous provinces and mass extinctions. Earth Science Reviews 53: 1-33.

Yates AM, Hancox PJ, Rubidge BS (2004). First record of a sauropod dinosaur from the upper Elliot Formation (Early Jurassic) of South Africa. South African Journal of Science 100: 504-506.

Picture of the day: Golden Gate

I have no time for a lengthy post, I'm in the midst of teaching three honours modules and trying to catch up on a backlog of research papers I need to write. I got a big one away just last week...
Anyway this picture shows Golden Gate National Park at Rooidrai (Red Corner) where James Kitching found the Massospondylus embryos that featured in Science a few years back. The large sandstone bluffs are aeolian deposits belonging to the Clarens Formation, while the red friable rocks below are the fluvial mudstones and sandstones of the Elliot Formation. The embryos were found in the spoil heap created by blasting for the road. You can see the heap to the right of the truck ('bakkie'pronounced 'bucky' in South African) disappearing off down the gully.

A trip to the Limpopo Province

As I was growing up no African place name seemed more exotic and remote than Limpopo. Actually the Limpopo Province lies just a few hours north of Johannesburg, yet I had never visited it until this week. This is a pity because as it turns out it is a beautifull place indeed. We were on a quick trip up to the Soutpansberg Range to check out claims of metazoan fossils in the Paleoproterozoic Soutpansberg Group (approx. 1700 million years old).


Sandstones and Conglomerates of the Soutpansberg Group

Such a fossil would exceed the next oldest metazoan fossils by about a billion years, and to say I was sceptical would be an understatement. Nevertheless one has to check these things up, just one day someone may really have found something exceedingly significant - but not this time. Nevertheless although no fossils were found (as expected) the day turned into a lovely natural history outing.
At the very start of the hike we saw a large flock of crested guinea fowl, the much rarer cousin of the ubiquitous helmeted guinea fowl, but I could not get a decent picture (although I did get a new tick for my birdlist). We found evidence for leopards in the form of scratched tree trunks (but no leopards of course!). Here are some other shots I took during our search.


It is well known that everything in Africa has thorns but check out these - you'd swear the thing was predatory!

I also snapped this cute little scarab



A story surrounding this little plant is worth relaying.



This is Khat (Catha edulis), a largely East African and Arabian plant well-known for the amphetamine-like drug it produces. Nevertheless it is growing wild all over the Soutpansberg. The place we were staying at used to be a girls school that was closed down after the students discovered this local plant and Khat-chewing became endemic in the school.
After the Soutpansberg we stopped in at Makapansgat on our way back to Jozi. This is a wonderfull little valley that is famous for its limestone caves, one of which (the Limeminer's Cave) has produced remains of the protohuman Australopithecus africanus



Although there are no obvious australopithecus fossils lying around, you can still see this string of vertebrae, left in the cave roof where an articulated sabre-toothed cat was removed (apparently the specimen is at wits but I haven't seen it).



Lastly to end with the palaeo theme, there wasn't just plio-pleistocene fossils to see. The caves themselves were etched into 2 billion year old dolomites with very nice stromatolite fossils.