The tornado that devastated Moore, Okla., has now officially been placed into the highest category—5 on the so-called Enhanced Fujita Scale, meaning winds in excess of 320 kilometers-per-hour. Scientific American corresponded with meteorologist Harold Brooks, a senior scientist at the National Oceanic and Atmospheric Administration's (NOAA) National Severe Storms Laboratory in Norman, Okla. 

[An edited transcript of the interview follows.]

What are the key ingredients of a powerful twister like this? What conditions make it possible? Were the warning signs there for such a tornado in the previous days weather?
Low-level warm, moist air with dry, relatively cold air aloft, strong wind shear [change of the horizontal winds in the environment with altitude]. The storm prediction had been forecasting the probability of a severe thunderstorm outbreak since Wednesday, 15 May, and mentioned the possibility of strong tornadoes by Friday. On Monday morning, looking at the output of high-resolution numerical weather prediction models that can resolve individual thunderstorms, the local forecast office mentioned in an 11 A.M. online briefing that if a thunderstorm formed in the [Oklahoma City] metropolitan area that afternoon, it was likely it would become severe and produce a strong tornado rapidly. 

What can be done to protect oneself against a twister of this magnitude?
Get as low as you can and put as many walls between you and the tornado as possible. Get into a well-built structure. People survive tornadoes of this magnitude all the time. It may not be pleasant and they may get injured, but doing the right thing dramatically increases your survival chances.

Is there anything that can be done to better design structures to withstand such tornadoes? Are there types of buildings that are more resistant to such winds?
Mansard roofs are better than gables. Put hurricane clips attaching roof to walls and use straps and bolts with nuts to attach the wall to the sill plate.

How is it possible that multiple tornadoes follow very similar paths, as has now happened in Moore in 1999, 2003 and 2013?
The direction is largely a function of the wind profile in the atmosphere, so most tornadoes go from southwest to northeast. Other than that the location is a matter of luck, most likely. If you threw a large number of darts at a wall, you could go up afterwards and draw a bull's-eye around a cluster. The process is essentially random, but even random processes produce clusters.

How much warning did the residents have and how did that warning go out?
The warning from the National Weather Service was five minutes before any tornado, 16 before significant damage and about 35 minutes before it got to Moore. The warning goes out on NOAA Weather Radio and through crawls on TVs and interruptions on regular radio. In addition, I assume outdoor warning sirens were sounded.

What can be done to improve such tornado prediction and forewarning?
Improve our understanding of all of the physical processes involved in tornadoes and the modeling of unresolved processes in numerical weather prediction models. From a user perspective, we need to do a better job of learning how to communicate information effectively.

What are the most dangerous misconceptions people have about tornadoes, or the common mistakes people make when faced with a potential tornado?
That you can't survive a violent tornado unless you're below ground. Also, some people probably still open windows, which is a bad idea.

Why do so many tornadoes occur in the U.S. compared with globally?
The ingredients come together more often here because of the presence of the Gulf of Mexico and the Rockies. Winds out of the south at low levels between the Rockies and Appalachians bring warm, moist air from off the Gulf. Winds out of the west aloft come over the Rockies or high desert. The best way to make an airmass dry and relatively cold is to bring it over a wide, high range of mountains. Thus, when flow is out of the south at low levels and the west aloft, the right thermodynamic profile is created, and the only missing ingredient is something to initiate the storm, such as a front or a dryline. The only place on the planet with a source of moisture equatorward of the region and with a wide, high range of mountains extended a long way from north to south off to the west is the central U.S.

What role, if any, does climate change play in exacerbating the conditions for tornado formation? How is global warming changing, if at all, the map of tornado probabilities?
None that we know. The reporting database, accounting for changes in the way we collect reports, don't show trends. As the planet warms, we expect the warm, moist air to increase and the shear to decrease. It's a function of temperature gradients, which are projected to weaken as the poles warm more than the low latitudes.

As a result, one component (thermodynamics) will become more favorable and another (wind shear to start rotation of storms) will become less favorable. It's not completely clear what the sign of the overall impact on tornadoes will be, but it's likely to be relatively small.