Determining whether extreme weather events are caused by climate change is crucial in planning for risks. Else, we will reach a situation in which corrective action may not be enough to protect us.
Over the past several years, headlines on weather-related extreme events have included heavy downpours followed by floods, droughts, storms, heat and cold waves, and wild fires. Such events typically destroy lives, property and ecosystems while stretching the capacities of disaster management departments and coffers for emergency funds in various parts of the world. “Protecting people before and after major floods, storms, and extreme events is a core part of our business,” said Karsten Löffler, Managing Director of Allianz Climate Solutions.
Since one of the main impacts anticipated from climate change is an increase in the intensity, frequency or duration of extreme events, there is usually a lot of interest from the media and the public after an extreme event to learn if it was due to global warming. The challenge that scientists and insurers face in responding to such questions is that of “attribution” — that is, to what extent can one consider climate change to be the cause of an extreme event?
Most such events have one or more components that are not related to climate change. For example, incompetent forest management practices contribute to fires. Poor land use planning contributed to heavy downpours and floods in Chennai last year. Consequently, what experts are trying to do to understand attribution is to separate the climate signal from everything else. There are generally nine kinds of extreme events that are considered: heat and cold waves, droughts, wildfires, extreme rainfall, tropical and other cyclones, extreme snow and ice events, and severe convective storms.
Scientific studies of extreme weather events and their attribution to global warming may help various groups such as planners, emergency responders, policymakers and insurance companies. Better knowledge of the risk contributes to how communities, governments, investors and others prepare for the future, with regard to planning cities, proposed infrastructure, natural resources or food security.
Determining attribution
Can scientists tell if an event is caused by climate change?
In order to determine attribution, scientists run climate models to simulate an event or they rely on the observational record from which they may estimate the statistical chance and magnitude of an extreme event. Often, they use both these kinds of approaches.
According to a recent report from the U.S. National Academies titled ‘Attribution of Extreme Weather Events in the Context of Climate Change’, “event attribution is more reliable when based on sound physical principles, consistent evidence from observations, and numerical models that can replicate the event.” All these conditions are not satisfied for every type of extreme weather event. Where long records exist, good models are available, and contribution from non-climactic factors such as human activity can be better considered, attribution turns out to be more robust. Scientists’ confidence in attribution to climate change varies among the kinds of events. There is greater confidence in attributing heat and cold waves, for example, over the other kinds of events described above. With regard to extratropical or mid-latitude cyclones and convective storms, it appears that there is little to no confidence in attributing them to climate change.
Another confounding issue is that there is a natural variability in the occurrence of weather events in any case, so scientists would be looking for a signal that is over and above the natural variability. For this reason, it is difficult for a scientist to be absolutely sure that a particular singular event has been caused by climate change.
As efforts to improve our understanding of extreme events improve, the ability for attribution is expected to improve. As in any other kinds of scientific studies, the accuracy improves with various advances including validation across different approaches, advances in modelling methods, and the accuracy of historical records of such events.
Fat tails and insurance
Global insurance companies were among the earliest groups in the world to ring the alarm on climate change. They are on the frontlines since their business is to estimate the risk of extreme events and then provide protection from their potential impacts. The profits they make arise from the fact that such events are rare. As their frequency, magnitude and impacts increase, the companies’ losses escalate. Some insurers are, in fact, limiting their coverage to those in areas with a moderate risk to climate change impacts and are expanding their business and activities to include solutions to climate change.
Scientists sometimes use the term ‘fat tail’ to describe extreme events. A normal distribution curve, what we know as a bell curve, shows a lot of variation near the average, but produces very few points at the far end of the curve. Biological parameters such as height of Indian women or men are examples of normal curves. In a fat-tailed distribution, on the other hand, portions of the curve that are distant from the average are thicker, and this implies that there is a higher chance of large deviations from the average.
Climate models generally assume a normal distribution rather than a fat tail distribution around the mean, thus ignoring the low probability high-impact events. Economists and some scientists have been telling us that we need to be prepared for extreme temperature and weather events. Gernot Wagner and Martin Weitzman explain the implications of fat tails for climate policy in their book, Climate Shocks: The Economic Consequences of a Hotter Planet.
Many of the points discussed here may appear nuanced, perhaps not significant, and also difficult to address within the regular political cycles of 4-5 years. But it is the extreme weather events and their incidence that are beginning to increase our everyday experience of climate change. We do not have the capacity to appreciate a change in average temperatures over 50-100 years, but can see what havoc an unusually intense storm or severe drought can cause in our own lifetime.
International agreements such as the recent Paris climate pact and the global targets for sustainable development set goals for governments and political parties to enable nations and communities to address the risks the world faces in the medium and longer terms. We must address anticipated risks even before all our models become accurate enough to estimate every detail of climate extremes. Otherwise, we will reach thresholds beyond which making corrective improvements to deal with climate change may not yield the protection we need.
(Sujatha Byravan is Principal Research Scientist at the Center for Study of Science, Technology and Policy, Bengaluru.)
We must address anticipated risks even before all our models become accurate enough to estimate every detail of climate extremes
