I am a Colombian marine biologist kidnapped by European geologists, who once out of my colorful home nest in the tropical Americas and working with comparatively boring white-powdered micro and nannofossils in Europe, got the promise that this work would improve climate change understanding. This would allow me to have the largest possible impact in the conservation of the colors of the biodiversity I was born with and grew up to protect.
I never imagined how necessary it was adding biological ingredients to the mostly geology-based salad to improve our understanding of past, present, and future climate change. Our most famous fund-catching concept goes like this: “we can better mitigate and adapt to future climate change by better understanding the climate response to CO2 in past times that we think were similar to what we will face in the future”. Very true! But, the base to better understand past climate relies on our abilities to understand the indicators (proxies) we use to reconstruct this past climate, which are mostly coming from organic and inorganic fossils that were alive millions of years ago. After 15 years of being kidnaped by European geologists, the salad has still too little spices of biological and physiological understanding of processes that control the indicators we base our reconstructions on.
From the different time periods in the past, perhaps the Miocene (~5.33-23.03 million years ago) is currently the most relevant to study, because there were CO2 concentrations (~400-600 ppm) similar to what we expect in moderate, most realistic future emission scenarios (RCP 4.5-6.0; as defined by the Intergovernmental Panel of Climate Change). Moreover, the Miocene continental configuration is relatively similar to that of today. The idea was therefore, to better understand climate of the Miocene, and as it has been most of the research I have set a quest on since I arrived in Europe, I wanted to do it in an unconventional, new, different way.
We started investigating the possibility of applying clumped isotope geochemistry to coccoliths, calcite plates that are produced by marine photosynthetic nannoplankton and that act as an exoskeleton. Coccolith clumped isotopes could be therefore a new proxy of euphotic (lit area that enables photosynthesis) ocean temperatures. This proxy is only based in thermodynamics and is independent of seawater chemistry, and was the core objective of my Marie-Curie fellowship at ETH Zurich, perhaps the most renowned laboratory at that time for these analyses.
To achieve this objective, I first needed to solve the main complication of producing unprecedentedly-pure, and relatively large coccolith samples from bulk sediments. This entailed me, once more, swimming against the current of wide-spread bulk sediment research, upon which many paleoclimatological paradigms are still based on. We do not want abiotic calcite, or that formed by other organisms in our samples, because this calcite is produced by organisms that may have lived in different environments and most likely biomineralize using different cellular pathways. This pushed me and my team to develop semiautomated microfiltration machines, and a method of nannofossil separation that allowed me to achieve >90% coccolith samples, also available in my lab at MARUM in University of Bremen.
The results of application of clumped isotopes to Miocene North Atlantic coccolith pure samples blew our minds! You can find the published paper at: https://doi.org/10.1038/s41467-025-65954-y. Perhaps the most applied and accepted temperature proxy, especially for the Miocene, is the alkenone unsaturation index, which is based on organic fossil molecules produced also by coccolithophores. Sea surface temperature estimates from alkenones has in part led to the paradigm accepted amongst proxy and modelling researchers of extreme high latitude warmth and flattening of the latitudinal temperature gradient during past warm intervals, like the Miocene. Now in more comprehensible words: the more CO2 there is in the atmosphere, the more the ocean behaves like a warm soup. Wherever you put your finger, you will get equally burned, independently of whether it is in the poles or in the tropics. This paradigm always seemed to me a bit awkward, because in the Caribbean, where I did Marine Biology, I could see with my own eyes how most forms of life terribly suffer during the warm, non-upwelling season, when temperatures are up to 28-30° C for a couple of months a year. How is it possible that life in general could handle, survive, and thrive in temperatures higher than those, also in non-tropical areas, for millions of years?
We initially thought a comparison of temperature trends and absolute temperatures between our new coccolith clumped isotope proxy and the well-accepted alkenone proxy would tell us where our new proxy could be “wrong”. Results were astonishing. Temperature trends were remarkably similar, but coccolith clumped isotopes suggest a 9 °C cooler North Atlantic ocean, questioning the extreme high latitude warmth paradigm. We could not find any source of significant cold bias in our results, given the purity and good preservation of our samples, high reliability of analyses, and similar trends compared to alkenones. Our absolute temperatures are the first to agree (late Miocene), and the closest to match (mid Miocene) most advanced Miocene modeling studies, which have been traditionally deemed insufficient in simulating the extreme high latitude warmth suggested by the proxy community until now. Coccolith clumped isotopes do suggest a warmer North Atlantic compared to today, but not temperatures of a hot homogeneous soup.
Taking a closer look to the ingredients that contain the salad of the alkenone proxy, you realize it is based solely on correlations of the alkenone unsaturation index and sea surface temperature. The biological function, physiological pathways that control the level of unsaturation (double bonds), and all potential non-thermal influences linked to coccolithophores’ ecology (like nutrients, phase of growth, season, depth, light, etc) are poorly understood, and calibrations assume invariably surface and mean annual (or warm period in case of the North Atlantic) season of production. This highlights the need to add more biology-physiology based ingredients in this salad (and many others) to deem conclusions obtained using this proxy (and other proxies) as paradigm-valid. Just imagine, for one second, that lower absolute temperatures were also found in other high latitude areas, and/or warm intervals? What if models were right and proxies were not interpreted fully right? It would revolutionize the way we interpret high latitude climate response to increasing CO2, and provide a more optimistic (less catastrophic) future perspective of high latitude climate evolution.