Journal Club: Woolly mammoth extinction due to warming climate
SUMMARY: In this scientific whodunnit, the latest research points a finger squarely at changing climate as the main culprit leading to the extinction of the woolly mammoth.
Woolly mammoth, Mammuthus primigenius, reconstruction in the Royal BC Museum in Victoria (Canada). The display is from 1979, and the fur is musk ox hair.
Image: Flying Puffin (Creative Commons Attribution-Share Alike 2.0 Generic license.)
Why did woolly mammoths go extinct? Did climate change do them in? Or did humans eat them into extinction? Or did a meteorite cause their demise? Or disease? Or ... ?
A team of British and Swedish scientists just published a new study indicating that changing climate -- not humans -- played a major role in the extinction of the woolly mammoth, Mammuthus primigenius. Additionally, the team's analysis of ancient DNA revealed that Eurasia was colonised by woolly mammoths that crossed the Bering Land Bridge from North America around 66,000 years ago. They also identified a previously unknown and genetically distinct population of mammoths that lived in Eurasia before they were replaced by an influx of Siberian mammoths approximately 33,000 years ago.
Below the jump, I tell you more about the scientific team and how they worked together to suss all this out.
Woolly mammoths, Mammuthus primigenius, were very large and hairy relatives to modern elephants. These large mammals were specially adapted to the cold, living only on arid steppe-tundra of the far north. They were widespread throughout the Holarctic during the Late Pleistocene (approximately 116–12,000 years ago). Recently, a number of well-preserved mammoths have been found in the permafrost of Siberia and Alaska (Beringia), providing us with tantalising glimpses into the life history and genetics of these animals. For example, previous genetic studies on Beringian mammoths identified two deeply divergent lineages, dubbed clade I and clade II, hypothesized to have evolved in isolation on either side of the Bering Strait (doi:10.1016/j.cub.2007.05.035).
Yet despite these remarkable insights, we actually know surprisingly little about mammoths, about the genetic structure and population demographics across their Holarctic range, nor do we even know why these animals ultimately went extinct just 4,000 years ago. To investigate these questions further, an international group of scientists, led by Eleftheria Palkopoulou, a doctoral candidate at the Swedish Museum of Natural History in Stockholm, conducted an ambitious series of studies where they obtained samples from recently unearthed European and Siberian specimens and combined those data with previously published DNA sequences to create a mammoth-sized data set from more than 300 individuals. By analysing this dataset, the team reconstructed the species' population history from more than 200,000 years ago right up until its extinction.
What did they do?
To carry out this study, the team collected samples from 88 bone, tooth and tusk specimens in specially dedicated ancient DNA (aDNA) labs:
Collecting ancient DNA from a mammoth tusk.
Image: Love Dalén/Stockholm University & University of London.
The DNA was purified and PCR-amplified by Ms Palkopoulou and Love Dalén, a paleogeneticist and assistant professor at the Swedish Museum of Natural History. The same 741 basepair (bp) fragment of mitochondrial DNA (mtDNA) was successfully amplified from 56 of the 88 specimens (16 more specimens yielded a short 79bp fragment that was informative for clade identification only).
These newly amplified sequences were aligned with the same regions from previously published woolly mammoth mtDNA sequences. These provided a data set comprising sequences from a total of 320 individual mammoth specimens, ranging from the Late Middle Pleistocene (more than 200,000 years ago) up until the mammoth's extinction, roughly 4,000 years ago.
What did they learn?
Ms Palkopoulou and Dr Jessica Thomas, a postdoctoral researcher who is currently based in the biology department at the University of York, then did a lot of number-crunching together. Based on their analysis of the 320 mammoth mtDNA sequences, the team identified 29 distinct maternal haplotypes (figure S3; larger view):
Figure S3. Median joining haplotype network. Coloured dots = geographic location: North America; blue, Chukotka/Kamtchatka; purple, Siberia; red, Europe; green, Wrangel Island after isolation; yellow. Shaded areas = different haplogroups. Black dots = missing haplotypes. Haplotype size proportional to frequency in dataset except haplotypes 2, 20, 37 that have frequencies above 10. indicated Arrows indicate the two oldest haplotypes found in eastern Siberia.
As you can see in this data image above, these 29 maternal mtDNA haplotypes clump together into five haplogroups (coloured clouds) and in the figure below, we see that these haplogroups cluster into one of three distinct and highly diverged woolly mammoth clades (figure S4; larger view):
Figure S4. Bayesian phylogeny of woolly mammoth mtDNA sequences. Tip-label colours indicate the geographical origin of the sequences. The time scale on the x-axis is in calendar years before present. Bayesian posterior probabilities of major internal nodes above 0.8 are shown.
These clades correspond to small pockets of favourable habitat, known as refugia, where the mammoths retreated during periods of climate warming.
"This suggests that spells of warm climate made the mammoth more susceptible to extinction", said Dr Dalén.
These genetic data suggest that there were at least three separate interglacial refugia during the Eemian interglacial period, which began about 130,000 years ago and ended about 114,000 years ago. The European mammoth (clade III) distribution indicates that during this time, there was an interglacial refugium in Western Eurasia, whereas the more restricted distributions of clade II and clade I mammoths imply there were two additional refugia, one in northern Siberia and the other in North America, respectively.
But when did these genetic lineages first split? Another way to ask this is to estimate when the most recent common ancestor of all the mammoth clades existed. This required yet more number-crunching: this time, a statistical model was used to analyse the genealogy of the mtDNA region in the database, tracking sequential changes in this mitochondrial gene region back in time to estimate when the single ancestral copy existed (figure 2; larger view):
Figure 2. Posterior distributions for time parameters of models 1B, 2A and relative sea-level variation in the Bering Strait.
Using coalescent simulations, the researchers estimated the time elapsed from when the three populations expanded (figure 2a) and split (figure 2b). The data analysis estimated that the clade I woolly mammoth population split between North America and Eurasia approximately 66,000 years ago (figure 2c).
This population split also coincides with the first time that sea levels had receded enough for the Bering Land Bridge to pop up during this particular glacial period (these are the time periods when the thin wavy line dips below the heavy straight line in figure 2d). The team inferred this was also the timeframe when mammoths moved westward into Eurasia (this assumption is the most likely explanation for these data, although the researchers did not explicitly test this hypothesis).
After the team mapped out the spatial distribution of the radiocarbon-dated and genetically analysed mammoth specimens, a rather interesting series of biogeographic snapshots emerged, revealing that mammoths became highly dynamic during the second half of the Ice Age where some populations expanded whilst others disappeared completely (figure 3; larger view):
Figure 3. Spatial distribution of radiocarbon-dated and genetically analysed mammoth specimens. Dates are given in calendar years before present. Colours indicate clade membership of the specimens: clade I; purple, clade II; pink, clade III; green.
The team interpreted these data images as follows:
- following their range expansion from North America to Eurasia, clade I (purple dots) mammoths appear to have lived alongside clade II (pink dots) mammoths in Central and East Siberia until the disappearance of the latter
- based on radiocarbon dates from clade II specimens (pink dots), this clade seems to have disappeared approximately 45,000 years ago
- clade I (purple dots) mammoths expanded westward into Europe, where they replaced the endemic clade III (green dots) mammoths
- clade III (green dots) mammoths disappeared from the fossil record approximately 34,000 years ago
- clade I (purple dots) mammoths seem to have first appeared in Europe approximately 32,000 years ago
"This process culminated with a severe decline in population size that started when temperatures began to increase at the end of the last Ice Age", said Dr Thomas in a press release.
What does this mean?
"We found that a previous warm period some 120,000 years ago caused populations to decline and become fragmented, in line with what we would expect for cold-adapted species such as the woolly mammoth", said Ms Palkopoulou in a press release.
"[O]ur data suggests that the same thing happened during the penultimate warm period (an interglacial some 120,000 years ago), long before modern humans had even left Africa", said Dr Dalén in email.
The patterns reported by this study are similar to those described during the same time periods for two other holarctic species; the cave bear (doi:10.1016/j.cub.2007.01.026) and more recently, the collared lemming (doi:10.1073/pnas.1213322109).
"We actually started the lemming study about the same time as the mammoth study, so the two projects have influenced each other", said Dr Dalén in email. Dr Dalén was a co-author on the lemming paper, too.
"In some cases, it seems that the population dynamics (local extinctions and recolonisations) may have been correlated between mammoths and lemmings, which would imply that changes in the environment (driven by climate) were behind these events".
"[O]ur recent paper on collared lemmings seems to support environmental factors as a driver of population size for all of the arctic specialist fauna", said co-author Ian Barnes, a professor of biological sciences and a reader in molecular palaeobiology at the University of London. Professor Barnes was a co-author on both this paper and on the lemming paper.
"[I]f you want to understand why large mammals go extinct in the last ice age, you need to understand what small mammal populations are doing", explained Professor Barnes in email.
"In brief, during warm periods, an increase in shrubs and drop in dry steppe grassland makes life difficult for arctic specialists, who tend to retreat first into northern Asia, and latterly into the High Arctic (and finally into extinction, in the case of the large mammals)."
But having survived several previous warming periods, why did the woolly mammoth go extinct after this most recent warming period? Did disease or hunting finally push them over the edge?
"Thus far, no disease has been identified that could possibly have affected a massive suite of taxonomically distinct large mammal species, but not small mammals", said Professor Barnes.
"As for humans causing the extinction, our study doesn't provide any evidence supporting this idea, but on the other hand there is no data disproving it either", said Dr Dalén in email.
"But the mammoth didn't become completely extinct at the end of the last ice age, since it survived another 5,000 years on Wrangel Island", continued Dr Dalén in email.
"What caused this population (and consequently the species as a whole) to go extinct is unknown. It could have been climate, humans or inbreeding (these are the three main hypotheses)", said Dr Dalén. "Or even disease".
Dr Dalén did remind me that if the current warm period (the Holocene) "hadn't been so darn long" -- more than 10,000 years -- mammoths likely would still be alive.
Like most good research, this study raises more questions than it answers.
"I really like the last point about the future need to look at why mammoths survived the previous warm periods, but died out in the recent one", said evolutionary biologist Tom Gilbert, a professor at the University of Copenhagen. Professor Gilbert was not involved in this study.
"That's really to me the most interesting question out there."
Palkopoulou E., Dalén L., Lister A.M., Vartanyan S., Sablin M., Sher A., Edmark V.N., Brandström M.D., Germonpré M., Barnes I. & Thomas J.A. (2013). Holarctic genetic structure and range dynamics in the woolly mammoth, Proceedings of the Royal Society B, 280 doi:10.1098/rspb.2013.1910 [OA]
Love Dalén [emails; 10 & 11 September 2013]
Tom Gilbert [emails; 10 & 11 September 2013]
Ian Barnes [emails; 11 September 2013]
Additional readings that were also cited:
Barnes I., Shapiro B., Lister A., Kuznetsova T., Sher A., Guthrie D. & Thomas M.G. (2007). Genetic Structure and Extinction of the Woolly Mammoth, Mammuthus primigenius, Current Biology, 17(12):1072-1075. doi:10.1016/j.cub.2007.05.035 [OA]
Hofreiter M., Münzel S., Conard N.J., Pollack J., Slatkin M., Weiss G. & Pääbo S. (2007). Sudden replacement of cave bear mitochondrial DNA in the late Pleistocene, Current Biology, 17 (4) R122-R123. doi:10.1016/j.cub.2007.01.026
Brace S., Palkopoulou E., Dalen L., Lister A.M., Miller R., Otte M., Germonpre M., Blockley S.P.E., Stewart J.R. & Barnes I. & (2012). Serial population extinctions in a small mammal indicate Late Pleistocene ecosystem instability, Proceedings of the National Academy of Sciences, 109 (50) 20532-20536. doi:10.1073/pnas.1213322109
.. .. .. .. .. .. .. .. .. .. ..
.. .. .. .. .. .. .. .. .. .. ..