On July 29, 2010, temperatures in Moscow climbed above 100 degrees F for the first time in recorded history, soaring to 102 degrees F. During a 35-day run when the daytime temperatures reached at least 86 degrees F, more than 10,000 people died in Moscow alone because of the sweltering conditions, Russian officials said. Throughout the summer, wildfires burned more than 1.6 million acres in western Russia and the heat-induced drought in July and August destroyed 40 percent of the country’s annual grain yield.
Almost from the beginning, Kevin Trenberth’s phone began ringing off the hook. Trenberth leads the Climate Analysis Section at the National Center for Atmospheric Research in Boulder, Colorado, where he studies how different manifestations of global climate connect with each other. That makes him one of the usual suspects when extreme weather events — including floods, heat waves, and intense storms — strike and the public wants to know: Is this happening because of climate change?
Scientists say you need to compare the most extreme events to ‘ordinary’ extreme events.
“We get requests from the media as these events are unfolding,” says Trenberth, “but it isn’t necessarily dealt with properly by anyone in the [climate science] community.” What happens all too often, he says, is that scientists end up giving the same hand-waving explanation that whileclimate change may well have contributed to whatever is going on, it can’t be definitively fingered as the sole cause.
Finding trends in extreme climate events is tricky because extreme events are rare, by definition. It’s instinctive to compare heat waves, droughts, or floods to average weather conditions. So for the Russian heat wave, say, you would look to compare it to an average summer month. But climate scientists who specialize in statistics say this doesn’t capture the true nature of an extreme event. Instead, they argue, you need to compare the most extreme events to “ordinary” extreme events.
“When you’re looking at extreme heat, for example, you need to compare the five hottest days in every year, or the highest overnight temperature in a month,” says Gabriele Hegerl, a professor of climate science at the University of Edinburgh in Scotland. She says this so-called “extreme value approach” is a better way to see the long-term variations in even the most unusual kind of event.
Using the extreme value approach, Hegerl and others have confirmed that the number of extremely hot nights around the world has been increasing over the past few decades. And when they’ve looked to computer models to see if the same sort of thing might have happened in an emissions-free world, they don’t see it. It seems, says Hegerl, that the increased frequency of heat waves may well have been caused by human behavior.
“It is really interesting because for years we’ve just laid back and said, ‘You can’t say anything about a single climate event,’” says Hegerl. “But now people are showing you can say something about the probability of that kind of event occurring.”
In a broad sense, this is what detection and attribution — D&A, to insiders — is all about. It involves searching for both changes in climate trends and for what, exactly, has caused these changes. Basically, D&A is broken down into two questions: Is the climate changing? And if so, are humans causing it?
When an extreme event occurs, people ‘find out what we are not adapted to,’ says one expert.
The simplest and oldest question in D&A is whether rising atmospheric concentrations of carbon dioxide (CO2) caused by human activities, such as fossil-fuel burning and deforestation, are driving up Earth’s average temperature. The increase in CO2 was first confirmed in the 1960s, and rising temperatures in the late 1980s. But simply spotting the two trends wasn’t proof there was a cause-and-effect relationship. In order to show that, scientists had to rule out other possible explanations.
In recent years, climate scientists have run thousands of computerized simulations to compare the climate with only so-called natural “forcings” — from the sun, volcanic activity, and other non-human influences on temperatures — and also with greenhouse gas emissions exactly as they have been in real life. Time and again, the simulations run with added CO2 have shown a warming climate, which allowed the 2007 Fourth Assessment Report from the Intergovernmental Panel on Climate Change (IPCC) to conclude that humans are the overwhelming cause of increasing global temperatures.
Climate models also show that global temperatures should continue to rise as concentrations of CO2 and other greenhouse gases increase, and suggest that heat waves, severe floods, and powerful storms will also become more frequent. But most people want to know: Is it already happening?
“When an extreme climate event occurs, then people find out what we are not adapted to,” says Francis Zwiers, a statistician who directs the Pacific Climate Impacts Consortium in Victoria, British Columbia, and was the lead author of the chapter on climate attribution for the IPCC’s Fourth Assessment Report. “These events cause a lot of damage, so people are particularly interested in them.”
The best that 2007 report could do was to say that such extreme events as heat waves and heavy rainstorms had “likely” become more frequent
over the past 50 years, and that extreme heat waves in particular were at best “more likely than not” to have been caused by human activity.
Because these kinds of events happen so infrequently, it’s been difficult for scientists to confirm any difference in their long-term trends. Extreme heat waves might hit a region only once every 10 or 20 years, and mega-droughts may only happen a couple of times a century.
It’s clear that what used to be highly unusual events are now becoming more common.
This is where the extreme value approach becomes key. To confirm that temperature extremes have been rising globally over the past few decades, D&A scientists have used as a benchmark the hottest days and nights of every month and every year. To see if heat waves — defined as three or more days of excessively high temperatures — are increasing, they’ve looked for the hottest three-day periods for each year going back through the last century’s weather records. In both cases, it’s clear that what used to be highly unusual events are now becoming more common.
That was the detection part. For attribution to human influences, climate scientists have looked once again to climate models. And once again, models in which only natural forces are at work don’t show any clear trend, but those with human emissions match the rise in heat waves that scientists have detected.
More recently, Hegerl and Zwiers have also looked at how heavy-rainfall events are changing. It’s tougher here than it is with heat waves. That’s because, unlike heat events, really heavy rainfall — the kind that can lead to vicious floods — usually lasts for hours, not days or weeks, and often happens over just a few hundred square miles. So far, Zwiers says, scientists have some good evidence that precipitation extremes are also on the rise worldwide, but it’s too early to say anything with complete confidence. “We’re at the beginning stages on extremes and on figuring out what the human influence is,” says Zwiers.
In terms of finding a human fingerprint in more frequent extreme climate events, he says the detection and attribution community is at a similar place today as it was a few decades ago when scientists were trying to prove, beyond a reasonable doubt, that greenhouse gas emissions had led to global warming. “We are ready to make a cautious statement that humans have contributed to more climate extremes because most of the evidence suggests this,” says Zwiers. “But it’s not a sure thing yet.”
For most people, that’s not good enough. They don’t care about statistics; they care about actual weather. When events like last summer’s heat wave strike, researchers still have to field questions about what role global warming played in that particular situation.
Answering those questions is one way scientists think they can help people better understand how climate change will affect them in the future. It will always be true, as researchers have cautioned, that no single climate event can be caused entirely by climate change. “There have always been extreme events,” says Peter Stott, a climatologist from the UK’s Met Office. “Natural variability does play a role, but now so does climate change. It is about changing the odds of the event happening.”
Recent reports suggest climate conditions relating to the Russian heat wave were primarily natural.
What Stott and others have started to do, therefore, is to use climate models to compare how often specific extreme events having the characteristics of, say, the 2010 Russian heat wave, would happen without man-made CO2, and how often they should happen with it. Then they can work backwards and apply the extreme value approach to the most unusual events the climate models have simulated.
For example, after the blistering 2003 summer heat wave in Europe, Stott’s group calculated that greenhouse gas emissions had more than doubled the likelihood that such a heat wave would have occurred. More recently, they found that man-made climate change had also increased the odds in favor of the kind of major flood that struck the United Kingdom in 2000. Their analysis doesn’t rule out the possibility that the events were strictly natural, but they do show that greenhouse gas emissions have made those particular extremes much more likely.
According to Trenberth, however, while the idea of calculating the changing odds is a good one, in practice the computer models aren’t good enough to capture how the entire global climate might influence a particular region.
“To get a really good answer, you need a perfect climate model,” he says. “And to the extent that climate models aren’t perfect, then where does the error go?” All too often, according to Trenberth, scientists err in the direction of saying extreme climate events are still largely caused by natural variations.
The question of how well global climate models can be used to zoom in on regional climate has already emerged as scientists search for the causes of last summer’s Russian heat wave. Recent reports from the National Oceanic and Atmospheric Administration
(NOAA) suggest that the climate conditions leading to the heat wave were primarily natural and that any human influence was drowned out by climatic variations that are inherently part of the natural system, rather than triggered by human influences.
What concerns Trenberth is that NOAA’s analysis doesn’t take into account the abnormally high sea surface temperatures in the Indian Ocean at the time of the heat wave, and how that could have influenced what happened in Europe. Even though the Indian Ocean and Western Russia are thousands of miles apart, the climate systems over them are interwoven. The Indian Ocean temperatures can affect air pressure across Asia, for example, which can influence the flow of the jet stream — and the influx of warmer air masses from the south — across Europe. “They haven’t answered the question of what caused the sea surface temperatures to get so high,” he says, which means that scientists can’t really be certain yet if global warming was a major cause of last year’s extreme heat in Russia.
Nevertheless, Trenberth and others in the D&A community say that the science of climate extremes is vital. Tracking how extreme events are changing, and searching for a human fingerprint in them, is an important step in helping the public figure out what they can expect from future climate.
“For extreme events, the question isn’t, ‘Is it global warming or natural variability?’” says Trenberth. “It is always both. The question is just how much each is contributing.”