10 years of rapidly disentangling drivers of extreme weather disasters

In the summer of 2003, a devastating heatwave killed more than 70,000 people across Europe.

For many, this was the first undeniable evidence that climate change was not an abstract and distant threat of the future but a reality claiming the lives of people here and now. A year later, Stott et al., 2004 proved that the heatwave had indeed been made hotter and more likely to occur due to human-induced climate change, caused primarily by the burning of coal, oil and gas. That was the first time that scientists clearly identified the fingerprints of climate change in a specific weather event and marked the beginning of a new research field that today is known as 'attribution science.'

In the following years, extremely deadly weather events continued to happen across the world. One of those events alone, tropical cyclone Nargis, claimed more than 100,000 lives, destroying whole communities in Myanmar in 2008. The question of whether climate change had played a role in these events was always in people's minds. But despite now having a methodology to address this question, very few attribution studies were published, and they often came out months or years after the event. This meant that public and media attention had moved on, and so the opportunity to link people's experience of extreme weather and the impacts of human-induced climate change had been missed.

To change this, and to provide scientific evidence of whether and to what extent climate change played a role in the immediate aftermath of disaster striking, we founded the World Weather Attribution (WWA) initiative in 2014 and have been reporting on an increasing number of extreme weather events since, highlighting where climate change played a key role and where other factors were the main drivers of the events.

Over these ten years, World Weather Attribution has developed protocols that allow the rapid evaluation of different kinds of extreme weather events across the world. The team monitors extreme events globally and uses a set of fixed criteria to decide which ones to study. For every study, the WWA scientists partner with local experts and/or national meteorological agencies. We then use weather observations, climate models and expert literature to analyse how climate change influenced the event and what elements on the ground turned a weather event into an humanitarian disaster. As soon as the findings are ready, we make them publicly available with the hope they can help inform the conversations around the causes of the event, its impacts and what needs to be done to reduce the damage and protect the population from future events.

Here, we look back at the ten deadliest events since the devastating heatwave of 2003, when the scientific field of attribution first began. By doing so, we aim to highlight what we have learned in our ten years of operation, both about the role of climate change in fueling extreme weather and the vulnerability and exposure factors that turn these hazards into disasters.

The events include three tropical cyclones in the Indo-Pacific (Sidr, Nargis and Haiyan), four heatwaves in Europe, two heavy rainfall events (one in India, one in the Mediterranean) and a drought in the Horn of Africa. Together, these events caused more than 570,000 deaths. And in all of them we find the fingerprints of climate change.

Weather is complex and societies across the world are very different. With every study we learn something new that can be used to help us better prepare for the future. We also find a lot of commonalities - things that are true for many or all weather events in certain regions or even everywhere in the world. After ten years of rapid studies, we can often make the connection between climate change and an extreme weather event without having to run a detailed analysis.

We know there is no such thing as a natural disaster. It is the vulnerability and the exposure of the population that turns meteorological hazards into humanitarian disasters. Increasingly though, there are fewer and fewer meteorological hazards that can be purely described as 'natural'. Our work, alongside the wider scientific literature, now shows that with every ton of coal, oil and gas burned, all heatwaves get hotter, and the overwhelming majority of heavy rainfall events, droughts, and tropical cyclones get more intense.

Key takeaways

Attribution science

• Advances in attribution methods means we can now isolate the influence of climate change in a complex range of extreme weather events, such as the tropical cyclones and droughts included in this analysis. Attribution studies continue to show that the world's failure to move away from fossil fuels is leading to dramatic changes to extreme weather, contributing to hundreds of thousands of deaths and affecting billions of people each year.
• Drawing on existing attribution studies and undertaking several new analyses, we find that climate change intensified the ten deadliest extreme weather events since 2004 as identified by EM-DAT, contributing to the deaths of at least 570,000 people. The result underscores both how dangerous extreme weather events have already become with 1.3°C of global warming, and the urgency of reducing emissions. With warming set to reach around 3 °C of warming by the end of this century given currently implemented policies, the hazards posed by events like those analysed will only worsen. We stress that our study does not capture the hundreds of thousands of heat-related deaths which are not routinely reported or studied in most regions of the world.

Saving lives

• Reducing vulnerability and exposure is most important to save lives in the event of dangerous weather. Early warnings of coming extreme events that reach people in harm's way and are followed up with early action mechanisms (evacuation, opening cooling centres, etc) are often the difference between life and death. This, coupled with updating and maintaining infrastructure such as dams to be built for the new climate, can help avoid the infrastructure failures that routinely kill thousands of people.
• Policies that fail to prioritise the most vulnerable people reinforce inequality, and when coupled with poverty make people less able to respond to subsequent shocks and stresses, which are now happening more frequently.
• Some of the events we studied are unprecedented or extremely rare, surpassing the reasonable ability of any government to prepare and design for, representing limits to adaptation. The inevitable losses and damages that occur as a result underscores the urgent need for mitigation to reduce the pace and number of these extremely rare events going forward.

Interpreting attribution studies

• The most extreme heatwaves across the world are made orders of magnitude more likely by climate change. Even though heat extremes will become even hotter with further warming, detailed heatwave attribution studies are no longer required to identify whether there is a role of climate change. However, as the dangers of extreme heat are still underestimated and underreported, attribution studies focussing on vulnerability and exposure are important to highlight the impacts and identify where and why adaptation, including early warning and early action, help to save lives.
• To meaningfully communicate risks, the metric selected for defining extreme weather events for attribution studies should be one that reflects the impacts. For example, the impacts of a heatwave may depend on both high humidity levels and temperature. Similarly, for droughts, crop health is influenced by water availability for evapotranspiration. This therefore requires analysing trends in overall water balance, including both precipitation and potential evapotranspiration (PET).
• Data quality and availability varies between event types and regions, statistical information thus needs to always be interpreted in the context of other lines of evidence. Especially in regions of the world where records of weather observations are short, or events like heavy rainfall only occur infrequently, climate change signals are not statistically significant in the observed record. However, this often arises from data availability and does not mean that climate change does not have an effect, especially when the physical mechanisms, such as the increased water holding capacity of the atmosphere with global warming, are known. Early event attribution studies often emphasised statistical evidence over physical knowledge and have thus overstated uncertainty.
• Similarly, in many regions the impacts of extreme weather events are known to be under-reported. It is striking that four of the ten deadliest recorded weather events of the last decade have been heatwaves in Europe, a wealthy region that is relatively well prepared for such events. Given the extreme temperatures reported elsewhere in the world, particularly in the global south where there are even larger exposed and vulnerable populations, it's almost certain that deadlier events have gone unreported and unstudied.

What the 10 deadliest events taught us

• The magnitude of change due to climate change often differs between different climate models and between observations and models, making it essential to include both in attribution studies. Previous studies on the extreme rainfall that led to the devastating 2013 Uttarakhand floods found that human-driven climate change increased the likelihood of the event, though conclusions were limited by the methods and climate model resolution. By reanalysing this event with a more diverse set of climate models and observations we find that human-induced climate change indeed doubled the likelihood of the devastating downpours and made it 11% more intense which is in line with the expected increase due to the warming since pre-industrial times.
• Similarly, despite having been studied in several peer-reviewed attribution assessments, the role of climate change in the deadly Russian heatwave of 2010 has likely been underestimated as no study included observed data, which showed an increase in the likelihood by a factor of about 80 in 2010, while model-based studies estimated a change in likelihood of a factor of 3-5.
• Tropical cyclones are compound hazards occurring in several distinct regions. Many studies find that extreme rainfall from North Atlantic hurricanes and Western Pacific typhoons is increasing due to climate change. Other basins and aspects are much less studied, but new methods are emerging to rapidly attribute these events. Looking at the three deadliest cyclones this century, Sidr, Nargis and Haiyan, we find that quantifying the role of climate change in the associated rainfall is highly uncertain. However, we also find that the destructive high wind speeds and the sea surface temperatures, which contribute to storm intensification, were higher in all cases due to climate change. This highlights the importance of assessing all aspects of extreme weather events.