Bad weather can lead to violence. Take it from climatologist Mark Cane, who found in 2011 that El Niño was linked to outbreaks of civil violence around the world.

“In places where people are on the edge — on the margin — they can be pushed over the edge,” Cane tells Inverse.

Short of stopping bad weather, our best hope of preventing this pattern lies in predicting what’s coming and when. Cane, a Columbia University professor who along with Princeton’s George Philander won the 2017 Vetlesen Prize for achievement in Earth science in recognition of decades of work studying El Niño, says protecting against these kinds of risks is why weather modeling is so important.

“If there’s going to be a strong likelihood of a bad drought somewhere, you can help aid organizations get ready for it by stockpiling resources where they are likely to be needed,” says Cane. “It can be exhausting work — almost impossible to do that every year. If you cry wolf too often than people don’t pay attention, so it’s important to only cry wolf when there’s a strong likelihood there’s a wolf.”

Cane helped develop the first El Niño forecasts in the 1980s. If El Niño, a major climatic oscillation between warm and cold water in the equatorial Pacific, is on its way, farmers can use better information to, say, plant more drought-resistant crops when less rain is expected. Cane’s work led to the foundation of the International Research Institute for Climate and Society, which works to improve climate modeling and the communication of that knowledge to communities that might benefit from it.

Climate models are essentially computer programs that digitally simulate the world’s atmosphere and surface conditions. If you want to know the likelihood of an El Niño event in the near future, you would feed observed current conditions into the program, and then fast forward into the future while applying the laws of physics. You run the model, say, a thousand times, and count up the number of end results showing El Niño conditions. If you see El Niño pop up in 750 of the runs, then you’ve theoretically got a three-quarters chance of the pattern showing up in the real world.

At this point, the El Niño models are pretty sophisticated and precise. “We can actually get a pretty good idea almost two years ahead,” says Cane.

Unfortunately, we’ve got new weather worries ahead. Climate change, like El Niño, makes for the sort of extreme weather events that can push communities and countries to violence. It’s no coincidence that we call someone who is quick to anger hot-blooded — there’s surprisingly robust evidence linking hotter weather to more violence. Some research suggests that Africa could see a 40 percent increase in violent conflict by 2050 if climate change continues unabated.

“If you add climate change to what El Niño does, the extreme events get worse,” says Cane. In a world of global warming, resources that go towards climate prediction and mitigation become even more important. And the research of Cane, Philander, and others has laid a solid foundation for continuing that work.

Climate models offer only forecasts, not exact scenarios. Like weather reports, they deal in percent chances of a given outcome occurring. But it’s easy to understand the difference between a 20 percent chance of rain and an 80 percent chance of rain and use that information to determine whether or not you will carry an umbrella.

And the good news is that climate models are getting better all the time. Even as the climate becomes more extreme and more chaotic, our scientific resources to understand potential outcomes are becoming much more powerful. A big part of that is due to increases in computational power, which allow for the models to take in and process data.

“What we’re learning about El Niño events or other kinds of climate variability is going to continue to be applicable with this added pressure, added stress, that climate change piles on,” says Cane.

Photos via NASA/Wikimedia