What atmospheric scientist Jiwen Fan saw on her television in July 2013 appalled her. The worst flooding to hit China in 50 years was happening in Sichuan province, in the same place that had been devastated by a massive earthquake just 5 years earlier. Over the course of 5 days, 73 centimeters of rain pounded the mountains, peaking at 29 centimeters in a single day. Rivers burst their banks and poured through city streets, washing away homes, factories, and bridges. Steep valley slopes collapsed in deadly landslides. About 200 people died, and a further 300,000 were displaced.
But Fan was worried about more than just the immediate effect of the floods. The Richland, Washington–based researcher—an expert on air pollution and climate at the Pacific Northwest National Laboratory in Richland, Washington—wondered how they had gotten so strong so fast. The Sichuan basin, surrounded by mountains that trap smoke billowing from its industrial centers, is “notorious” for its dirty air, she says. Did air pollution play a role? To find out, she and her team of Chinese, American, and Israeli researchers designed precision computer simulations to model what had happened.
Air pollution can affect precipitation in many ways. Sometimes, the aerosol particles in smoke can reduce or delay rain. Sometimes, they can make thunderstorms more intense. Their best understood interaction is in changing how water vapor condenses to form droplets in clouds. But Fan and her team have proposed a first: that pollution also changes some air circulation patterns that lead to rainclouds.
In the case of the Sichuan storms, they write in a paper published online before print in Geophysical Research Letters, soot in particular contributed to the catastrophic flooding. It prevented rainclouds from forming over the basin during the day, leading to more intense rainfall in the mountains that evening. “We were amazed at the scale of the effect the pollution had,” Fan says. “Effectively it redistributed the precipitation from the wide area of the basin into the mountains.”
Fan and her co-authors ran two forecasts for the weather system that passed over the Sichuan basin during the peak of the floods: one with the thick blanket of smoke that covered the region and one with the kind of clean air that existed 40 years ago, before the Chinese economic boom. In the clean air model, moist air at Earth’s surface was heated by the daytime sun, became buoyant, and rose to great heights, triggering a convective cycle that led to storm clouds and mild daytime rainfall. But in the dirty air model, the dark veil over the plain soaked up much of the sun’s warmth high in the atmosphere, while simultaneously cooling the streets and fields below. This altered thermal structure stabilized the daytime atmosphere and suppressed rainfall.
But as night fell, the moist air mass moved northward toward the Longmen Mountains, which tower some 2000 meters above the basin. The weather system that had been building energy over the plains for 12 hours was driven upward as it collided with the range’s steep contours, triggering the postponed convection. A day’s worth of rainfall from the plains was focused into a few hours over a handful of mountain valleys.
Geography and pollution combined to make the floods intensely severe, Fan says. And she suspects the combination is not unique. Catastrophic floods in Pakistan only a month later, she says, may have involved the same factors: heavy industry plus a mountain backdrop.
Nanjing University meteorologist Aijun Ding says that Fan’s model “raises an important point” and supports his own observations. He found that air temperature and rainfall dropped precipitously when farmers outside the city of Nanjing burned their crop waste en masse, forming a wall of sooty smoke that blotted out the sun in June 2012. Intense thunderstorms had been forecast for the city, but never materialized, though the precipitation downwind the following night was stronger than expected. It is as if anticipated storms can be “burned off by intense air pollution during daytime,” Ding says.
Fan worries that such effects are not being taken into account in weather forecasting. In China, for example, she notices that forecasts often give the wrong area for thunderstorms, which are likely to be downwind of where expected. Such forecasts also get the intensity “worryingly wrong,” she says.
Atmospheric chemist Greg Carmichael of the University of Iowa in Iowa City agrees it’s high time for air pollution to become a regular part of forecasting. Earlier this year, he published a study in Geophysical Research Letters that tied an exceptional outbreak of 122 tornadoes in the southeastern United States to a bout of biomass burning in Central America. He says this kind of research shows that aerosol feedbacks can be large enough to impact weather. As he puts it: “Ignoring aerosols is becoming less of an option.”