My reserach on Extreme Events

My Max Planck research group (2019-2024) explores the theme of ‘extreme events’. What are extreme events, and why do they matter?

A lot of recent research on the topic comes from climate scientists. Things like heatwaves and other such abrupt climate events are often described as extreme events. The 2004 Indian Ocean Tsunami also emphasised the role of non-climate processes such as earthquakes in causing sudden changes. But, how ‘extreme’ does something have to be in order to be classified as an ‘extreme event’? My interest in this topic reflects the fact that it represents a profoundly interdisciplinary nexus, and one in which science and philosophy meet. How do we distinguish an ‘event’ from a ‘process’, for instance? The necessary interdisciplinarity of a topic such as extreme events also invites scientists to step outside their disciplinary comfort zones, and engage with colleagues in other disciplines.

And while things like the 2003 European heat wave or the Toba Supereruption 74,000 years ago can straightforwardly be defined as extreme events, in my opinion virtually all subjects and disciplines have some underlying similarities in terms of debates. Is change primarily gradual or rather is it driven by occasional abrupt changes (‘events’)? Such debates are clear in climate, and we can think about the notion of switch points between different stable states. But similar themes are much discussed in evolutionary biology, for instance, where notions of gradualism are contrasted with notions of ‘punctuated equilibrium’. Is history driven by slow changes in economies and societies, or rather do occasional battles and coups and so on play a more profound role? The answer, presumably, is that different levels are all important, the next question, then, is how they relate and the relative importance of different modes and tempos of change in space and time.

In my own specialist subject, archaeology, the notion of extreme events is central, although not usually articulated with such terminology. For instance, one traditional idea is that there was a ‘human revolution’ (there are many synonyms, including cognitive revolution, Upper Palaeolithic revolution, etc.) when simple hominins suddenly became modern humans ‘like us’. While this notion is now widely rejected by archaeologists, it remains an alluring notion, and continues to be propagated by authors such as Yuval Harari. The alternative is that human evolution was more gradual. Many researchers, such as Alison Brooks and Sally McBrearty, have argued that there was no human revolution (link). In both anatomical and behavioural terms, ‘modernity’ was arguably a gradual process. Such debates can be found throughout the study of the human past. For instance, was there a Neolithic revolution? This suggests that after millions of years of nomadic foraging existence, there was a sudden change as people took up agriculture and settled in villages. Archaeological research, however, complicates this traditional narrative, and highlights gradual processes, experiments, and the difficulty of a simple division between hunter-gatherers and agriculturalists as traditionally understood.

So, in my opinion multiple subjects, perhaps all, have similar debates about gradualism versus extreme events. At root, this could perhaps be simplified to the distinction between processes and events. Thinking through different examples can help elucidate some of the themes here. Take the 2010 eruption in Iceland. This was, in the scale of volcanic eruptions, rather small. In fact, for most of human existence, I don’t think anyone would even have known that the eruption had happened. However, due to the particular meteorological, demographic, and technological characteristics of this time and space, the eruption had a huge impact. The direction that winds were blowing at that time, the fact that Europe is currently a densely populated region, and the fact that much of our inter-regional travel is by aeroplanes that are vulnerable to volcanic ash particles led to this small eruption having a big impact. The eruption led to transport chaos across Europe, and therefore justifiably could be described as an extreme event. I think the Iceland example makes a very pertinent point on extreme events being extreme by their context. It is not that a particular process is extreme in its own terms, but rather by the articulation of different processes.

The extinction of the dinosaurs (the non-avian dinosaurs at least) makes another interesting case study. Did, for instance, a major asteroid impact 65 million years ago cause widespread extinctions which otherwise would not have happened, or, rather were dinosaurs already in trouble as the planet of the climate changed, and the asteroid simply accelerated the inevitable? As well as the general point here on the need to understand the role of both long-term climatic/evolutionary processes and the impact of the asteroid impact, this example provides a window into scientific methodology. Claims about things like extinction processes depend upon good quality, representative, data being available, in space and time. Given that very few of the organisms that exist become fossilised, that geologically formations of the right age need to be exposed for fossils to be found, and thanks to regionally diverse research histories we know a lot more about some parts of the world than others, we have to very cautious about the assumption that available fossil records are representative. In the case of dinosaur extinction, it had been argued that actually fossil data suggested widespread extinction occurring before the asteroid impact. However, interesting studies have been done (link) which highlight some of the biases in fossil preservation and recovery. In all disciplines we need to think about the quality of our data, as well as the themes of correlation and causality.

With my research group and various collaborators, I have sought to explore the definition, character, and impacts of extreme events. My interests here are scientific, but I am also fascinated by the historical aspects. For instance, to what extent did the gradualism of Darwin reflect science and to what extent did it reflect socio-political aspects of 19th century Britain? Augustus Pitt Rivers, founder of the Pitt Rivers Museum in Oxford was rather blunt on the point, declaring in 1891 that “the law that Nature makes no jumps, can be taught…to make men cautious how they listen to scatter-brained revolutionary suggestions”. He had artefacts carefully arranged to suggest gradual changes in material culture.

In a review paper we explored using a quantitative literature review method how extreme events are defined and understood (link). For instance, we could imagine extreme events as the upper and lower points of a distribution (imagine the top and bottom 1% of a bell curve, for instance, say of rainfall at a particular spot). We could also think about transitions between stable states, as often emphasised in climate studies. Other perspectives might emphasise the role of contingent changes, be there genetic mutations or the quirks of the historical process.

A lot of my own research on the topic explores how human societies responded to environmental change. One level of this topic is to explore how abrupt changes in factors such as rainfall impacted human societies. Another is to think through questions of scale; processes that appear abrupt can actually seem much more long-term when we think about them at a human scale, such as generations. Given the often wide uncertainties of age estimation, exploring the actual ‘extremeness’ of past events can be challenging. And again, we have to carefully consider the specific context. A certain climatic event, for instance, can have very different impacts on nomadic versus sedentary societies - extreme events are made extreme by context and interaction. This, however, can go both ways. On the one hand we imagine that being sedentary increases vulnerability to climate change, compared to nomadic groups who could simply move somewhere else. On the other hand, sedentary groups might have more potential to dig deep wells and other such changes that could help mitigate the effects of a climatic shock. These different aspects are often discussed in terms of the relative vulnerability and resilience of societies.

A lot of my work focuses on the impact of climate changes on prehistoric societies in Arabia (link). For most of the Quaternary period, Arabia has been mostly arid. Yet episodically, monsoonal rainfall spread across the peninsula and transformed it into a grassland, covered with lakes and crossed by rivers. Working with many colleagues, we try and understand how the environment changed, and how human societies responded, but also how humans changed over time and how this altered the way they responded.  Studying stone tool assemblages give indications into the kind of mobility strategies used by early humans, and various palaeoclimate archives indicate the character of climate change. For instance, at the site of Khall Amayshan (KAM-4), a hollow between large sand dunes provide space for small lakes to repeatedly form over the past 400,000 thousand years, when the monsoon system occasionally reached its northern limits. Each time these lakes formed, humans spread into the area and left their stone tools around the lakes, giving a long record for how humans responded to these episodic phases of environmental amelioration.

One theme I am interested in here is that climate and the environment are often seen as synonyms, but this can be a problematic assumption. For instance, it is well known than in East Africa the rise and fall of lakes does not simply reflect changing local climate, but also changing catchments and connectivity caused by tectonic activity. This recognition has led to new scientific approaches, such as the use of strontium isotopes in East Africa to tease apart the factors influencing changing lake levels (link). Likewise, understanding water availability in an area, say Arabia, does not simply depend upon changes in rainfall, but also factors such as aquifer dynamics. A dry year can have a very strong impact in an area of impermeable rock with limited aquifers, but perhaps have little impact somewhere else if there are huge aquifers which feed lakes and springs.

My contributions to such perspectives include a consideration of the role of volcanic activity in Arabian prehistory (link). While things like human dispersals are often linked to changes in rainfall, could things like volcanic eruptions have also influenced the movement human groups? It is hard to say, but we should be open minded to the possibilities, and keep in mind that environmental change does not simply mean climate change. One aspect that can be clearer on is how volcanic eruptions change hydrology. It is clear in Arabia that volcanic eruptions often led to lava flows which flowed downhill before cooling and plugging valleys, creating natural reservoirs. The effect of this is that increased water retention in the upper parts of catchments, and the formation of lakes and other water bodies, can occur despite no change in the level of rainfall.

I am also exploring the role of climate on human societies in the Maltese islands. We can again think of things in terms of  vulnerability and resilience. The small size of the islands and their generally semi-arid character may have made humans vulnerable to episodic droughts. On the other hand, the fortuitous existence of the impermeable Blue Clay layer in the local geological sequence created perched aquifers. This meant that at least in certain areas, springs meant continued water access even after a dry year. The question of how climate would have impacted societies depends upon many factors such as the extent to which food was stored, and whether crops were rainfall fed or irrigated from springs. As discussed in the page on my research in Malta (link), we recently conducted a study on the hypothesis that the 4.2 ka climate event caused a societal collapse in Malta.

In both Arabia and Malta, my ongoing research focuses on trying to develop high resolution climatic and archaeological records.

My postdoc Chris Carleton developed a new method for exploring correlations between different radiocarbon dated processes (say, changing climate and regional archaeological activity). As one case study we explored the topic of the extinction of North American megafaunal extinction (link). Megafaunal extinction is a much discussed topic. Some researchers argue that it is very clear that the extinction of many large animals reflects human impacts, as we spread around the planet in the Late Pleistocene. Even if this is true, there are very different possibilities from an ‘overkill’ model of extreme hunting, to more indirect impacts relating to humans changing landscapes. And, just as with the example of dinosaur extinction, a lot depends on data quality and thinking through biases. In many places, the data relating to megafaunal extinction is actually very poor, and sweeping and simplistic conclusions are therefore perhaps premature (link).

In our study (link) we looked at North America, as this is somewhere that some of the best data is available, although it still leaves a lot to be desired. By exploring the long term correlation between archaeological activity, as a proxy of human activity, dated megafaunal fossils, and regional climate we made some interesting conclusions. And it should be emphasised here, that we did not approach the topic trying to ‘prove’ any particular argument. We found no evidence for a correlation between human and megafaunal levels. The evidence does not suggest a sudden over-hunting sweep, for instance. Rather, we found a correlation between climate change and megafaunal extinctions. Yes, correlation does not mean causation. But in this case, it seems hard to deny the role of climate, at least if one is open-minded to the data and not starting from strong a priori beliefs. The basic rises and falls in megafaunal populations seems to be driven by climate. And if we view extinction as a process not an event (an extreme one!), then climate is key. We do not rule out the possibility that humans were also implicated indirectly, such as through habitat change, or that humans delivered the coup de grâce. The reality is probably some combination of natural and human processes, and as is often the case, the reality is probably somewhere between more extreme interpretations in either direction. Arguments must be based on solid evidence, and not just-so stories.

This is becoming a rather long discussion of extreme events, so I will finish by pointing out that we have also explored more recent examples, such as 2nd millennium AD conflict in Europe (link) and the chronology of temple construction at Angkor (link). These examples highlight the ways in which quantitative methods can be beneficially applied to historical periods.