Joshua Wurman, president of the Center for Severe Weather Research in Boulder, Colo., whips up a response.

The glib answer for why tornadoes don't strike urban areas that often is: Cities are small.

If you take a look at Google Maps and see what percentage of the U.S. urban and suburban areas cover, it's a pretty small fraction. The regions where you have peak tornado frequencies—from Texas up through Kansas, and even east toward Atlanta and the Southeast—are open country, so that's where most tornadoes spend the overwhelming fraction of their lifetimes.

It's very rare that one encounters a city, as happened in Atlanta last weekend. In 1999 there was a tornado that hit Oklahoma City and killed about 40 people. It was a long-track tornado that lasted about an hour—but most of its lifetime was spent over pretty open country. It crossed two subdivisions, and that's where most of the fatalities happened.

The Atlanta twister has not been characterized as a violent tornado. Tornadoes are rated using the Enhanced Fujita (EF) Scale, which scores the damage caused by a particular twister on a scale from 0 to 5. Violent tornadoes are classified as EF4 and EF5, significant ones EF2 and EF3. EF5 damage is typically quite catastrophic: Houses are not only just destroyed but destroyed down to their foundation—no walls left standing--and the tornado might cause structural damage to larger, well-engineered buildings that are designed to survive even very intense winds. The tornado that went through Atlanta, although it broke lots of windows, didn't cause major damage to any downtown buildings. I believe that tornado has been rated an EF2.

As to whether global warming will increase the number of tornadoes—making it more likely that they will encounter cities—we have no clue.

Climate change predictions are uncertain, even on a global scale. One could be confident that the global temperature is going to go up; however, the error ranges on what's going to happen locally—whether the temperature is going to go up in Atlanta, or here in Boulder—are much wider.

In the case of hurricanes, there's a reasonable, if unproved, hypothesis that the storms might get more intense if global temperature rises and the surface temperature of the earth's oceans rises—after all, in terms of the physics, warm water is key to creating a hurricane. With tornadoes, you can't even make that basic statement because the effect of an increase in local temperature on tornado frequency or tornado intensity is unclear. Brazil is pretty hot, but it doesn't have a lot tornadoes. Oklahoma and Texas are really hot in July and August, but that's not the peak of their tornado season; spring is when those states see the most tornadoes.

So it is possible that climate change could shift the tornado season to earlier in the year as spring creeps into winter. Perhaps it will move the distribution of where the stronger tornadoes occur. A warming climate does not necessarily make for more tornadoes; it could cut off tornadoes completely or could cause their incidence to double.