EGU General Assembly 2020
© Author(s) 2020. This work is distributed under
the Creative Commons Attribution 4.0 License.

The clams before the storms: the fate of bivalve diversity during times of crisis

Katie Collins1, Stewart Edie2, and David Jablonski3
Katie Collins et al.
  • 1Natural History Museum, London, United Kingdom of Great Britain and Northern Ireland (
  • 2Caltech, Pasadena California, United States of America
  • 3University of Chicago, Illinois, United States of America

Of all the macrofossil groups, bivalves (clams, scallops, oysters and mussels) have perhaps the best global record. Known from the Cambrian, bivalves have survived every mass extinction and climate perturbation the planet has suffered. Many of the ~90 living families of bivalves with a fossil record have roots that stretch back to the Paleozoic.

This lineage longevity makes bivalves an ideal model group for studying biodiversity responses to changing climate – families have been separate for a long time, and their varied ecological roles and habitats means that the effects of climate on different biogeographic regions and ways of making a living can be teased apart. The abundance of bivalves in marine and freshwater deposits provides large specimen-level datasets for analysing survivorship across climate-event boundaries, such as the Paleocene-Eocene Thermal Maximum, a major warming event, or the end-Pliocene cooling. Bivalves have been shown to be a good proxy for much of marine benthic biodiversity, and they can give us insights into questions such as: is biodiversity response to climate perturbation predictable? What species and lineages are at risk, and can we identify them?

Not only can bivalves help us track effects of climate changes of the past thanks to their rich fossil record; they may also be a major player in human efforts towards our own future climate resilience. Bivalves today provide food for many millions of people worldwide, both in artisanal and commercial fisheries, and perform vital ecosystems services, such as water filtration, sequestration of carbon, and as both food and habitat for many other animals of all sizes. Increased aquaculture of molluscs has been postulated as a way to take more of the burden of feeding the world’s population off the terrestrial realm, and without the adverse effects of finfish farming.

This talk will discuss patterns and fluctuations of bivalve diversity through time, focusing on predictive models for survivorship across major climate transitions in the Cenozoic, and using the past behaviour of species and clades to look ahead to potential marine diversity responses to projected climate scenarios.

How to cite: Collins, K., Edie, S., and Jablonski, D.: The clams before the storms: the fate of bivalve diversity during times of crisis, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-8044,, 2020

Comments on the presentation

AC: Author Comment | CC: Community Comment | Report abuse

Presentation version 1 – uploaded on 01 May 2020
  • CC1: Comment on EGU2020-8044, Silvia Danise, 05 May 2020

    Hello Katie,

    I really appreciate your display. How did you estimate the thermal tollerance of extinxt species (I am referring to panel D)?


    • AC1: Reply to CC1, Katie Collins, 05 May 2020

      Hi Silvia - thermal tolerance for a species was estimated by taking the minimum and maximum temperature (from NZ proxy records e.g. foram Mg/Ca, d18O, TEX etc) for the time range that the species is known from, and subtracting min from max to get a range. This is a pretty crude estimate though, and we are hoping that temperature models for the Cenozoic will allow us to refine it in future analyses! 

      • CC2: Reply to AC1, Silvia Danise, 05 May 2020

        Thanks, Katie, this makes sense, of course these kind of data are scantier the depiest we go in time. 

        • CC3: Reply to CC2, Silvia Danise, 05 May 2020

          I meant deepest!

          • CC4: Reply to CC3, Silvia Danise, 05 May 2020

            I guess that another problem is for species that lived above and below the thermocline.....

            Have you got also information on the water depth range of each species (maybe relating it to preferred substrate/lithology/facies)?

            • AC3: Reply to CC4, Katie Collins, 05 May 2020

              So far, we have mostly restricted our bivalve paleobiogeographical analyses to (as far as we can) species that are known from above 200m water depth, given that the deep sea appears to function as a pretty separate ecosystem. But more and more we see that taxa move between shelf and deep environments and that interesting things are happening because of that crossover, so I would expect this to be more important in our future analyses. 

        • AC2: Reply to CC2, Katie Collins, 05 May 2020

          Yeah, it's a real issue - in the end I think this specific formulation of the PERIL metric will mostly be useful through the Cenozoic, but the concept of using a metric like this to test if any given variable actually does improve our ability to predict extinction (and therefore whether or not it seems likely that changes to that variable were involved in the extinction event in question) I hope will be useful in future!

  • CC5: Extinction mechanisms, Emilia Jarochowska, 05 May 2020

    Hi Katie et al., great work and beautiful poster! Are there any insights into the mechanisms of extinctions that could be behind the differences between the three examples you included? E.g. the Eocene-Oligocene transition seems (on your poster) to be associated with a weaker temperature shift than the one in the Early Miocene. What was driving the former?

    • AC4: Reply to CC5, Katie Collins, 06 May 2020

      Hi Emilia

      The only difference that I can be sure of between those two time intervals is that the sampling is worse in the older one, so I would hesitate to draw climatic conclusions between them. The Eocene material is much less abundant and very hard to collect, often being decalcified, and a few of the Eocene stages barely have any mollusca at all. I included that interval mostly to show that the relationship between the three variables in the PERIL metric and the predictability of extinction is still visible even under less than ideal sampling conditions. This work is very preliminary (I conducted the older two PERIL experiments a week before the conference!) but I'm hoping to explore this further and to find ways to deal with the sampling issues.