Of the thousands of tornadoes that have been spawned across the country in the past decade, very few have landed on major cities.

We remember the exceptions. Among them: Nashville in 1998, Fort Worth, Texas, in 2000, Atlanta this past week.

(ENLARGE)

Chris Stanfield/AJC (ENLARGE)

Dan Patterson of Woodstock tries to make a call as he views downtown Atlanta through the Omni Hotel's broken glass at the CNN Center after Friday's tornado.

STORM AFTERMATH • IRS giving storm victims more time to file • Officials believe man killed in tornado was seeking shelter • Federal aid available in Fulton PHOTOS: • One month after tornado • Oakland Cemetery • NW Georgia devastation • Scene downtown 3 days after • East Atlanta recovers • Cabbagetown • Vine City | Rome • Aerial view of damage • Downtown 1 | 2 • Readers' photos 1 | 2 | 3 MORE: • Woman killed by tornado mourned • Landmarks damaged | Map VIDEOS: • On YouTube • Aerials show downtown devastation • Atlantans survey damage • Storm hits MORE ALERTS: • Get AJC breaking news & traffic alerts on your mobile device.

In the wake of Friday's direct hit on downtown Atlanta — the first since record-keeping began in the 19th century — this question, or myth, some might say, has been much discussed: Are tornadoes less likely to touch down in urban areas? If so why?

Among some weather watchers, there is speculation that a city's built environment might suppress the formation of funnel clouds, said John Trostel, deputy director of the Severe Storms Research Center at Georgia Tech.

The theory is that the smooth flow of air into a storm cell is disrupted by tall buildings.

But Trostel also points out that the heat-trapping quality of the urban environment ought to produce the exact opposite effect, which would enhance storm production.

Such an effect might help explain the observation that in Georgia, Fulton and Cobb are among the counties that produced the highest number of tornadoes between 1950 and 2006, according to the National Oceanic and Atmospheric Administration.

But according to Trostel and others, that could also be a function of population, rather than incidence. More tornadoes get reported in more populated counties, where there are people there to report them.

Harold Brooks, with NOAA's National Severe Storms Laboratory in Norman, Okla., said the number of tornadoes occurring in urban areas is so tiny in any case that any ground effect from either heat or structures is lost among the statistical insignificance of the sample.

So the truth, it seems, is that cities are so relatively tiny, in comparison with the expanse of the countryside around them, that the odds of the already rare tornado landing on one are very steep.

"As far as we can tell," Brooks said, "metropolitan areas don't have any impact; you can't see any difference."

DOWNTOWN TORNADOES

The National Oceanic and Atmospheric Administration says these cities have been hit by twisters in the past 30 years:

*Tornadoes are classified on the EF (Enhanced Fujita) or F (Fujita) scale of weak (0-1), strong (2-3) and violent (4-5).

Highest fatalities in city tornadoes

• 255: St. Louis, May 27, 1896
• 114: Waco, Texas, May 11, 1953
• 79: St. Louis, Sept. 29, 1927

24 tornadoes have struck U.S. cities since 1871. The only cities to be struck more than once:

• St. Louis: 4 (1871, 1896, 1927, 1959)
• Shreveport, La.: 2 (April and December 1978)

— Source: National Oceanic and Atmospheric Administration