Doping the atmosphere?

When weather disasters happen, is climate change to blame? The stories, video, and interactives in "Weather on Steroids" explore that question from a number of angles. It turns out that blaming climate change for wild weather's not that simple. Here’s why.

Researchers can now estimate how likely several kinds of extreme events would be, with and without human influence. With that information in hand, they can then calculate the probability that a given extreme was made more, or less, likely by our century-plus of fossil-fuel burning.

Scientists at the heart of this work, known as attribution science, see great promise in the emerging ability to provide hard numbers to represent the impact of climate change on extreme weather—even though they sometimes need to cast their findings in a double layer of probability (see below).

For a hypothetical weather event, a study might report findings on the influence of human-produced greenhouse gases with statements like these:

• There’s a 90% chance that a summer as hot as this is now at least four times as likely
• The odds are two out of three that floods like the ones recently observed are now 10 times more likely
• There’s a greater-than-95% chance that a winter as cold as this one is now at least 20% less likely

The double layers of probability are because there are two kinds of uncertainty at work:

• Every simulation from a computer model has some amount of potential error
• In the real world, natural variations will help shape whether a given summer is a record setter or not