MICHELE NORRIS, host:
So far this year, there have been nearly 1,200 tornadoes in the U.S. That's close to twice the usual number and it comes as something of a surprise to the scientists who study them.
As NPR's Jon Hamilton reports, that's because there's still no good way to predict tornado outbreaks more than a couple of days ahead of time.
JON HAMILTON: Every year, government meteorologists tell the public about how many hurricanes to expect during the Atlantic season. But the government doesn't do that with tornadoes.
Howie Bluestein, a meteorologist at the University of Oklahoma, says there's a good reason.
Dr. HOWARD BLUESTEIN (Meteorologist, University of Oklahoma): I would say we have no skill in predicting whether the coming season will be good or bad depending on your perspective for tornadoes.
HAMILTON: Bluestein says even in the midst of a very busy year like this one, there's no way to know what's likely to happen even a few weeks from now.
Dr. BLUESTEIN: This could be it for the rest of the season or it could continue to be crazy. We absolutely don't know. Our predictability is not that good.
HAMILTON: Bluestein says one reason is that a tornado season can change dramatically in just a few days if the conditions are right. The right conditions usually involve severe thunderstorms in an area where warm and cold air are colliding.
Near Joplin yesterday, the conditions also included a type of thunderstorm called a supercell, which is a powerful rotating updraft.
This year, conditions like that produced more than 300 tornadoes in just three days at the end of April. Bluestein says that's a fairly typical length of time for an outbreak.
Dr. BLUESTEIN: Oftentimes an extreme event will occur. It will last for a short period of time and that will be a spike.
HAMILTON: Bluestein says spikes are hard to predict and they aren't necessarily part of a larger seasonal trend. There's also no solid evidence that tornadoes have been influenced by climate change.
In some ways, tornadoes are a lot like hurricanes, which are easier to predict. Both are violent rotating storms that can produce winds of more than 100 miles per hour. But Greg Jenkins from the Department of Physics and Astronomy at Howard University says that's where the similarities end.
Dr. GREGORY JENKINS (Department of Physics and Astronomy, Howard University): The time scales are different, the space scales are different and the sources are different.
HAMILTON: Jenkins says that while tornadoes depend on local or regional weather conditions, hurricanes are affected by climate conditions that cover a large chunk of the globe.
Dr. JENKINS: Hurricanes are coming, generally, from West Africa. There's a disturbance that's crossing the western part of Africa. It goes out over the ocean and then one out of 10 may become hurricanes.
HAMILTON: Whether they do can depend on things like sea surface temperatures in the Atlantic or winds shaped by El Nino or La Nina conditions thousands of miles away. And because these patterns occur on such a large scale, Jenkins says, they tend to change slowly over months or years.
Dr. JENKINS: Tornadoes, on the other hand, are much smaller scale. And so often, one sees a pattern that might persist for, you know, several days or a week. You know, lots of air, warm, moist air streaming up from the Gulf of Mexico. And then we get this big outbreak of tornadic thunderstorms.
HAMILTON: Jenkins says that because those patterns only last for a few days, meteorologists' predictions are limited to that same short time span. Jenkins says another factor is that even a big tornado like the one that hit Joplin is rarely more than a mile across. That's not at all like a hurricane.
Dr. JENKINS: The difference is the hurricane can stretch out for three, four, 500, 1,000 miles in the case of really big ones like Rita some years ago.
HAMILTON: Forecasters had been tracking Rita for more than a week as it slowly crossed the Atlantic. The tornado in Joplin had existed only for about 20-some minutes when it hit the city.
But thanks to improvements in short-term prediction, government scientists did provide some warning that a tornado was possible. At least 24 hours earlier, a map produced by the U.S. Storm Prediction Center in Norman, Oklahoma, showed a red zone surrounding Joplin.
Jon Hamilton, NPR News.