Scientists Identify Human Connection to Precipitation Extremes

By Alyson Kenward

Follow @alysonkenward

A view of the streets in Tresopolis, Brazil, on Jan. 13, 2011 after severe flooding hit the area. Credit: Pablo Jacob/Globo via Getty Images

With waters finally receding from early 2011’s collection of devastating floods, people in many parts of the world are wondering if the storms and their aftermath are a preview of future extreme weather. Until recently, there’s been little evidence that climate change is responsible for the growing number of intense rainfall events seen in many regions, but new research now shows there is a clear connection between some heavy rainfall events and increasing levels of climate-warming greenhouse gases in the air.

The floods from early this year, particularly those in Australia, were especially severe, and the number of intense precipitation events — the kinds where a near-record amount of rain falls in just a few hours or days — has been slowly increasing during the past 50 years, in both the U.S. and other areas. Yet while climate simulations of the future have long predicted that a warmer atmosphere may bring more frequent bouts of extreme rainfall, scientists have been unable to determine whether humans have already influenced how severely and frequently heavy rainfall events are occurring.

Researchers from Environment Canada have now established that higher levels of greenhouse gases in the atmosphere have contributed to the trend of more intense rainfall events in much of the Northern Hemisphere. Paired with a new method to analyze the human fingerprint in specific historic floods, the results indicate scientists are narrowing in on how much warming alters the odds of extreme precipitation events.

“Warming [of the planet] causes changes in extremes,” says Xuebin Zhang, a senior climate scientist from Environment Canada in Toronto. “But this means extremes can be either increased or decreased.” In other words, says Zhang, as temperatures rise some regions could experience more frequent heavy rainfalls, while other areas might have more frequent droughts. But because either event is only bound to happen rarely and there is a lot of variability from one event to another, the opposing trends might cancel each other out, he explains, making it hard to determine if human behavior has any influence.

Greenhouse gases lead to more heavy rainfall events

As part of the Climate Change Detection and Analysis group at Environment Canada, Zhang has been studying climate change for more than a decade. In previous work, published in 2007, he and his colleagues found that humans emissions of greenhouse gases have helped lead to an increase in total precipitation in some regions, and a decrease in other parts of the globe. But because intense rainfall events are more likely to cause serious flood damage, researchers are also interested in studying whether these more extreme rainfall events have become more frequent.

In a recent study led by his Environment Canada colleague Seung-Ki Min, along with researchers from the University of Victoria in British Columbia and the University of Edinburgh in Scotland, Zhang analyzed the largest amounts of rainfall recorded in one day each year from over 6,000 weather stations across North America, Europe and much of Asia. After observing that heavy precipitation increased over the time span of 1951-1999, the group compared the trend to computerized simulations of rainfall in the Northern Hemisphere during the same period to see what influence the greenhouse gas emissions might have had. They found that the increasing amount of heat-trapping gases, like carbon dioxide (CO2), in the atmosphere have contributed to heavier rainfall, and that the changes are not solely due to natural variations in climate. The results of the study are published in the February 16 issue of Nature.

“The big challenge to evaluate human attribution in climate extremes is that it is a difficult statistical problem,” says William Gutowski, a climate scientist from Iowa State University who studies how computer models capture what happens in the natural climate. Although he was not directly involved in this new study, Gutowski says that focusing on extremes is particularly challenging because there are fewer examples to draw from and analyze, which makes it tougher for researchers to assess the confidence of their conclusions. “In this study, I was particularly impressed with how the researchers approached this issue. It is my sense that they worked very hard to identify what was anthropogenic (human-caused) and what was natural variability. Despite the difficulty, they still get to a robust conclusion.”

Does climate change contribute to particular storms?

While Zhang’s study provides evidence that the global trend towards heavier rainfalls is in part due to higher concentrations of greenhouse gases, it doesn’t mean that manmade climate change is responsible for specific weather events, like the heavy rains that caused the recent floods in Australia.

Although extreme weather events — including heavy rainfalls — are rare, they have nevertheless always happened, says Peter Stott, who leads the Climate Monitoring and Attribution team at the U.K.’s Met Office. To figure out if the likelihood of a particular storm has increased because of manmade climate change requires an entirely different and more involved type of analysis, he explains. Through this kind of analysis, Stott and his team are studying whether climate change is tilting the odds in favor of more intense or frequent specific extreme weather events.

For example, following a severe 2003 heat wave in Europe, Stott’s group analyzed if that spell of high temperature days was more likely to have occurred because of manmade climate change. They found there was a 90 percent chance that human influences had doubled the risk of that particular heat wave. But while understanding the relationship between greenhouse gases and temperature changes is relatively easy, says Stott, it’s much more difficult to understand the relationship between climate change and the likelihood of particular precipitation events. 

“It’s tougher to study these types of rainfall events because they happen over a smaller area than heat waves do and often for shorter periods of time,” says Stott. The smaller spatial scales and shorter time spans pose a challenge to the ability of computer models to simulate this kind of event, with or without greenhouse gas increases, which is the foundation for Stott's type of analysis. 

The King Arms pub in York, North Yorkshire, in the U.K. after flooding in 2000. Credit: GeoBlogs/flickr

Nevertheless, in another new study published today in Nature, Stott and his colleagues have found a way to analyze how much greenhouse gas emissions may have increased the risk of a particular flood that devastated much of the U.K. back in 2000. During England’s wettest autumn on record since the 1770s, the persistent rains in 2000 caused many rivers to overflow, and York in particular suffered extensive damage. Insurance claims following the floods were estimated at over $2 billion USD.

Stott’s colleague Pardeep Pall found that by using a higher resolution climate model to focus more closely on detailed precipitation patterns over the U.K., he could compare the type of rainfall seen in 2000 to what would have been expected if CO2 concentrations were at the level they were prior to the industrial revolution. By combining that rainfall analysis with a hydrological flood model that estimates how rivers might overflow, Pall was able to gauge if the presence of more greenhouse gases in the atmosphere had increased the risk of the Fall 2000 flood.

After running thousands of simulations, Pall says the model shows that nine times out of ten, the risk of flooding was increased by 20 percent. A smaller proportion of the simulations (two thirds) indicate that the flood risk was increased much more, by 90 percent.

“We’ve set up a framework to assess if some types of flood events have the risk of becoming more frequent,” says Pall, though he cautions that their results don’t apply to all types of floods. He hopes that concept will be explored using other historic floods so researchers can better understand how the risk of other types of storms is changing.

“We’re still trying to understand the robustness of these results,” explains Stott, “But we would like to develop a situation where we can do more studies with this design so that we can better understand the risk of which extreme events we can attribute to climate change with confidence.”

Stott hopes that eventually, as the technique is improved, scientists will be able to do this type of analysis as extreme events are happening. He hopes such studies will help people understand that no one event can be fully blamed on climate change. Instead, he says, people can see that with climate change, the probability of extreme weather is changing, and they may better understand how to relate that to an event going on aroud them.