Tara Smiley

Ecological diversity of Turkana Basin mammalian faunas in the Miocene: Updates from the Turkana Miocene Project

This talk will present recent work from the Turkana Miocene Project investigating ecological dynamics of mammalian communities in the context of tectonic, climate, and environmental change over the Late Oligocene to Late Miocene in Turkana Basin, Kenya.

FULL TRANSCRIPT 

Thank you so much to the conference organizers, Lawrence, Fred, and Alicia. This has been truly just such a special treat to be here this week with all of you. I've learned so much and I'm really appreciative for the opportunity. I'm excited today to present results, very early stage results from myself and a number of co-authors with the Turkana Miocene Project. Now I'm a relatively new faculty member here at Stony Brook University in the Department of Ecology and Evolution and part of the Turkana Basin Institute. And one of the many gifts the Turkana Basin Institute has given me was the early opportunity to meet Isaiah Nengo. And on my first trip to TBI had a lovely dinner with him and we swapped lots of exciting excitement about the Miocene and from there I was able to be folded into this really exciting project, the Turkana Miocene Project, which you've already heard mentioned of several times. 

But what I'm really excited about for this project is this integration of data approaches and kind of people thinking very cross disciplinarily from earth to climate and life. And one of the best ways to do this integration is in the field. So, it's been really exciting for Turkana Miocene Project to get everyone out in the field together and TBI has played a fundamental role in facilitating that. So, as a paleoecologist and biogeographer, I'm a part of the life team and the life team goals are wide and varied anywhere from the amazing descriptions of the faunal assemblages at many of the sites that you've heard about this week, the isotopic characterization of those fossil communities that Kevin Uno and Dan Green presented and many other goals. But all of these goals kind of fall under this kind of primary driving motivation of understanding the role of landscape and climate change in shaping mammalian ecology and evolution. 

And in particular the interactions between landscape and climate and shaping novel habitats, heterogeneous habitats that generate these opportunities for adaptation and mammalian diversification that shape the paleoecology of fossil communities. Something I'm really interested in and that we work on in my lab is also understanding the deep time relationship between the development of topographic complexity in places like the East African rift zone and biodiversity hotspots. And then of course, how can we study and contextualize key events and primate evolution through the Miocene. The work I'm going to present today is very early stages, but along with my co-authors who, sorry, I should have mentioned there are many people in the life team. I've highlighted my co-authors here with their photos. What we're going to talk about today is kind of our early understanding of species richness patterns through the Miocene and how we're working with this record when we know there's a lot of spatial and temporal variation in fossil sampling.

We can then start to look at the potential correlates, be it climate tectonics or paleo environment of those species richness patterns. This is very much work in progress I'm not going to answer this question at all, but hopefully I can demonstrate how TMP is thinking about answering this question. And then at the very end I'll present briefly present some results looking at is ecological diversity sampled consistently across space and time. From there, can we tell anything about changes in ecological richness and structure through time and in relation to species diversity. This work is in large part possible thanks to a large data compilation done by John Rowan at the University of Albany. And what he's done is put together a long list of taxonomic occurrences for East Africa and for each of those species also classified them into different body size dietary and locomotive categories. 

So, here's a map of all of the sites for which John has compiled information for today's talk, we're just going to be looking at this kind of an Oligocene through Miocene records. So, any of the sites that are in the cooler colors, the grade, and turquoise colors. This dataset can be supplemented by information from the now database is something that we can thank Mikael and Indre and many others for this major resource. And then the Turkana Miocene Project sites are providing really invaluable updates to these faunal lists. And so, I'd like to thank Ellen Miller, Patricia Princehouse, Gabrielle Russo, and Natasha Vitek, as well as many of the taxonomic experts who have helped provide identifications for these faunal lists. So now when we put this together, what can we say about species diversity in the East African region? So here I'm plotting each of these points is one of those sites just kind a raw count of how many species are present at that site. 

And it's very tempting to look at this and ask how might this diversity relate to changes in regional climate, regional tectonics? Some of the kind of key points I'm highlighting here in terms of those landscape and climate change factors. But as we know from Kay's vast body of research and many others, is that there are many other factors that can contribute to variation in species richness across these sites. The first of which is these sites represent different paleoenvironments and taphonomic conditions. Very nice example of that comes from Buluk and Napudet two sites you've already heard about. Buluk is this mature fluvial system. It's sampled really this incredible diversity of large body mammals come from Buluk, whereas Napudet is this volcanically dominated landscape. It's ashfall tuffs proximal to volcanic source, these frequent volcanic interruptions. So, these two sites, just as an example, represent really different depositional environments and really different paleoenvironments. 

So not only could they have hosted different fossil communities, but the ways in which they've preserved those fossil communities may be very different. So, we have an overprint of past paleoenvironments and the taphonomy. We also know that a lot of these sites have been sampled variably over, so some of them are at these historically very well sampled sites and some which are very new like Torpenawi that we heard from Natasha earlier this week. We also know from geologists that there's a lot of variation in the availability of the rock record and what parts of that rock record hold fossils across space and time. And we know that there's variation in our sampling methods. I think a really nice demonstration of that comes from Mikael and Indre's talk and the cello Kyalo Manthi effect on the rodent sampling curve. And then we also are well aware that these sites are under active investigation. 

So people are going out finding new fossils all the time. So we're adding to our understanding of fossil diversity at each of these sites with a lot of careful and critical work. So what can we do then considering all of these potential factors that might be influencing species diversity? Well, we can take Kay's course in driving and hopefully stay on the road and start to work with this fossil record kind of at the regional scale. Try and leverage the whole fossil, the record of East Africa and attempt to think about species richness patterns at these broader spatial scales. And to do this, we can assume that species are present in East Africa between their first and last appearance and the regional fossil record. So that's these in the schematic, here are the fossil occurrences, here's that stratigraphic range we would build from that first and last appearance across the region. 

We can then use preservation models and preservation priors that are based on, we can change the shape of these preservation priors, but also take into account information about the number of fossil occurrences between that first and last appearance date and generate credible intervals. So, kind of expand that stratigraphic range. We know we're never finding the first or last occurrence of a fossil. We can then take that information and move across time and tally up diversity for 1 million year time bins. When we do this, we often have to remove singleton taxa or taxa that are only found in one bin because they can often represent kind of these oversampled sites, or they definitely represent a sampling bias. And what that means is exquisite some of our really exceptional fossils like alesi that you just heard about are kind of dropped from this way of thinking about diversity. 

But what these diversity patterns do is they can textualize these really exceptional finds and then we can use various sub-sampling approaches, which I'll show you a few to test the robustness of the diversity record we generate. So, here's a view of those stratigraphic ranges. For all of the taxa on John's data set the black lines represent the first and last occurrence of each taxon and the fossil record and then the gray dash lines represent those credible intervals. And so, you can see some are really wide that indicates fewer fossil occurrences and higher uncertainty on those ranges and some are quite tight. Now one of the problems with these approaches is they assume even sampling through time and through time space and across different groups. And we know that this is not necessarily true in the East African record. There are significant events that may shape our understanding of when we actually expect to get a first appearance date. 

So, I think this is a really, John Kappelman really talked about the importance of funnel exchange and the later legacy and early Miocene. We really need to take into account information like this when we think about those preservational models and how we can better understand species or lineage durations. So let's try and drive carefully. And now we're going to look at some patterns of species richness through time. And so, in blue here is that the way of thinking about species richness where we just use information directly from the fossil record, the first and last appearance dates, the dash line includes those singleton taxa. So you can see where we have some really, really well sampled million year time bins. And then the black is that view of species richness using confidence intervals. There's a lot of variation in this curve and we can walk through some of that variation and talking a little bit more detail. 

But I think what hopefully jumps out to you is that the sustained diversity high from about 20 to 15 million years ago, so elevated diversity from the early Miocene to middle Miocene, the credible interval approach, smooth some of the volatility and sampling. We would hope that it would do some of that, but of course we know that there are better ways to potentially, it may mask some of the variability that's actually biologically true. Now in these periods of low diversity, what's going on? Obviously we have some of this can definitely be attributed to our lamp post problem. Where do we have lamp posts actually shining light on the fossil record? We know from the work that's been presented here and a vast body of research that the African continent was not species to ate in the Oligocene. The Fayum tells us that the Oligocene was very species rich. 

And so, there are definitely caveats in this record but keeping the focus on the East African record really enables us to test the unique role that climate and tectonics may play. We may find kind of lamp post problems within the record few, there are fewer fossil localities here. And then we also have some edge effects that might be what's happening. We can have them kind of in the middle of the record where we have few fossil localities. And that's certainly happening at the end of the record. And I think if we were to include the Plio-Pleistocene in here, this would be a much more prominent peak. You can see all these singleton taxa that a lot of them have records in the Plio-Pleistocene so they would contribute to diversity kind of beyond what I'm showing you in this window here. So we can just look at the confidence interval diversity curve and test a few more things. 

Well, we can ask is this simply just a function of the number of localities that we have per time bin? And while there is kind of a broad correspondence that peak does line up when we have a lot of localities in the record, these are not significantly correlated with each other. We can do some re-sampling approaches where we remove, randomly remove five sites from the record to ask, is this dominated by just a few sites? And we can even do special case scenarios where we take out the most species rich and the poorest species, the sites with the least number of species and ask how are individual sites potentially contributing this record? What we see when we do that is yes, there is an influence in the late Miocene there are a few sites, the upper and lower Nawata that are really contributing to this peak here, but this middle Miocene peak, even when we remove sites repeatedly, we always get this sustained diversity high. 

So this gives us a certain amount of confidence that uneven sampling across space and time is maybe not the only factor, maybe not even the dominant factor shaping this curve and species richness. So what might be causing it, we can look back to that site level diversity. And what should jump out to you right away is that the Turkana sites in turquoise are a major contributor to this peak. They're a big part of this record. What we can also notice is that none of them are right up here around 70 species they're all about half of that. So what that might indicate to us is that we have high spatial turnover in these taxonomic assemblages. So maybe you can build high regional diversity by having high alpha diversity at a few sites, high alpha diversity, high within site diversity across all of your sites. Or you can build it just with moderate diversity at any single site, but a lot of turnover from site to site. 

So is that what we're observing in this regional diversity peak? We also think, no, based on a lot of the geologic work that we're sampling a broader range of paleoenvironments in this early to middle Miocene window. So maybe there were all of these different environments and faunal communities were across the landscape. We've just better sampled them and that's contributing to this peak. And I think we haven't resolved this yet, we still have work to do here, but we're going to forge ahead and ask what could be some of the global and regional scale drivers of these diversity patterns? And this is really one of the major goals of the TMP, the Turkana Miocene Project. How does global temperature, things like the middle Miocene, climatic optimum, relate to this record of species richness? What about regional climate and regional vegetation and paleoenvironments? We saw really nice overviews from Thure, Kevin and others about this. 

And then we can ask, what about tectonics and volcanism? Now this record, this Miocene record kind of precedes the rifting that has characterized the East African rift today with these large rift basins and north south running of lakes and river connections. However, during the Miocene, there was this interesting kind of story of localized topography. I was lucky enough to be in the field with Craig Feibel just a few weeks ago, and he explained this to me as like a lot of little volcanoes incipient rift basins. So, there is the potential for even though we don't have these large rift basins, the potential for some topographic complexity and kind of a way to break up species ranges in this Miocene landscape. Okay, so let's walk through these a little bit. We can throw up the Zachos curve. This is the oxygen ice topic composition of benthic forams. 

It's been used as a temperature proxy through time and we can say, well, there's pretty poor fit between the climatic, between the global temperature proxy and our species richness curve. And of course, many of us would agree that regional climate has likely had a stronger influence on species richness in ecology. And so where are we in terms of understanding regional climate? Well, this work is really being led by other members of the Turkana Miocene Project. Mikael and Indri have done amazing work on this even prior to TMP using dental ecometrics, so the shape of teeth for whole communities to try and infer paleotemperature, precipitation and vegetation. So a ton of work being done looking at biomarkers, environmental isotopes, and the plant fossil record to build out our understanding of the paleoenvironment and climates. And then this is all being done in conjunction with regional climate modeling. 

This work is led by Daeun Lee and Chris Poulsen. Here's some modeling output for mean annual precipitation and mean annual temperature for the African continent at 14 million years. And I think what's really neat about this collaboration is we can, I'm not doing this, but these folks can use the information from the Turkana Basin to validate these records and do this kind of iterative improvement. And we can also use paleotopography generated by the geophysical models I'll present next as boundary conditions on these climate models. So this is where my objectivity goes out the window, and I want to talk a minute about development of topographic complexity. This is something my lab works on, and what we're really interested in is how the development of topographic complexity and these tectonically dynamic and active systems can generate biodiversity hotspots that you observe here for mammalian species richness. This is modern mammalian species richness. 

We're overprinted over the latitudinal diversity gradient. We see a really strong influence of topographic complexity. This is found in many different groups across all continents Dino talked beautifully about such high diversity in grasses and insects and the Turkana in the East African rift, and I think this intuitively makes sense to us. We had a great conversation about how topographic complexity can strengthen environmental gradients, generate habitat heterogeneity that can relate to or promote greater turnover in species and traits across space. It also shapes the configuration of barriers and corridors, creating opportunities for the geographic isolation of populations and speciation, and then also influencing these dispersal routes. And it's really important to consider this in context of climate change as well. So, what do we know about the paleo topography of the Turkana Basin and the wider East African rift? We know a lot from a long history of geology, but what's exciting, a recent work that Troy talked about yesterday as well is these geophysical models that look at dynamic topography from mantle convection as well as the rules of extensional tectonics to build an understanding what the basin might've looked like from 30 million years to the present. 

:And from these models, for each 1 million year time slice, we can pull out variables that we think might be related to species richness, paleo topography, roughness. We can backtrack the paleo position of fossil localities. We can look at the extent and connectivity of drainage networks. So this is a really exciting frontier, I think, in considering what might be driving patterns of species richness as well as ecological diversity. So I'll talk just briefly about where we are with some early work looking at ecological diversity in the East African region. Again, this is possible thanks to John's dataset, but we know that there's a lot more work to be done here. These are pretty broad categories, and this dataset purposely leaves out some groups that are just not evenly sampled. So there's no rodents, for instance, in this data set because they're not very evenly sampled. So, we know that we haven't fully sampled the ecological diversity through the Miocene, but we can ask, have we at least consistently sampled it across space and time? 

And then if we have, what are some of those patterns of ecological richness and structure and how does that relate to species richness? So the first step of that is using kind of body size as an overall characterization of species ecology. And in the early days when we were all getting together, we're like, well, how do we know that we're sampling things evenly? So we looked at body size distributions, and I'm showing you just a sample from the Turkana sites where the site proportions are of these different body size categories are shown in blue compared to the regional proportions shown in gray behind the blue bars. And for most of the sites, we look at the sites body size distribution roughly mirrors that of the regional body size distribution. A few sites show different distributions. So this isn't a perfect match, but it does give us some confidence moving forward that we're kind of consistently sampling ecological diversity at these different sites. 

We can return to our stratigraphic ranges and use the same approach and ask what is the body size distribution in these 1 million time, you know for the whole, the regional species pool for these 1 million year time bins? And that's what I'm showing you here with the red is using that confidence interval approach and the blue is using that first and last appearance date so just the black lines. And what we see is that there are a few temporal bins that differ from the regional body size distribution, but overall there's a reasonable consistency in terms of the body size distribution that we're sampling through time and it's not terribly variable through time. So this gives us, even though again, we know we're missing parts of ecology, of  the ecology of these communities, this gives us a little bit of confidence that we're at least consistently sampling ecological diversity through time. 

In the last couple of slides, I'll just show you some metrics for which we're using to think about ecological diversity through time. The first is functional richness. So, you can think of this as the number of unique trait combinations or if you have a trait space here with trait one and two on your X and Y axis, the convex hole that kind of characterizes all of the taxa found within your sampling unit, be it a community or a region. We can think of functional dispersion, how far apart or how far away from the center all of those taxa are and then functional evenness, how evenly spaced are they across this trait space? And we would expect that each of these metrics would increase with species richness as we fill out or expand the ecological space and more rich communities. 

I'll back up but given that these are not communities that we're talking about, we're talking about the regional species pool, there's possibly a different expectation that we might have if we think that there's high taxonomic turnover across the region. There may indeed be high ecological turnover as well, but there might be kind of the same different groups of species doing similar things across the landscape. So we might also expect to see a decoupling between species richness and ecological diversity. So here are some of those patterns through time. I think it's more helpful to look at each of these ecological diversity metrics in relation to kind of an old distribution. What would we expect if we kept sampling the record the same way that the fossil record does? And so that's what these envelopes are showing here. And so we see some places where our empirical results kind of fall off of that null expectation line suggesting that there is some ecological structuring through time in the region. 

So a few things to point out. We see an early filling of ecological space, this steep rise in functional dispersion even when species richness is low. Then I'm going to draw your attention to the species richness peak where we would expect these ecological measures to be high. And what we find is that they're actually lower than expected given the species richness in these time bins. So, what does the say, well, it kind of fits that model of high taxonomic turnover, but potentially lower ecological turnover, some more beta diversity, but species are more similar, more clustered, doing the same thing even though their composition may be turning over across space. And then we see this, finally, we see this switch into the late Miocene where it's slight, but we're just slightly out of this null hypothesis envelope. And of course, we can think about this expansion of ecological diversity, very likely due to the expansion of C4 grasses and those habitats. 

So, with that, I would just like to wrap up and thank you for hearing about how we've leveraged this record to assess regional diversity, some of our findings of this high elevated species richness in the early to middle Miocene, and how this richness may be facilitated by high turnover and ecologically similar species. So where to next? We can definitely do a lot more determining the sampling controls versus the landscape and climate controls in this record. And we're interested in more than just matching up wiggles. We really want to understand the underlying biological processes, what are the speciation, dispersal and community assembly processes? And so, I think this is where TMP is a really exciting place to be because we really do need integration across data sets and across basio-temporal scales in order to understand this and beyond just TMP, the exciting projects that Gabrielle just shared with you, kind of using the Turkana as this high resolution window across major faunal transitions and climate transitions. So thank you all very much.

 

View All Paleoecology videos