Lessons learned from the May 3, 1999 tornado

By: Dr. Jeff Masters , 2:39 PM GMT on March 21, 2007

Storm Warning: The Story of a Killer Tornado, which I reviewed earlier this month, recounts the story of the May 3, 1999 monster F5 tornado that ripped through the southern suburbs of Oklahoma City. In addition to providing an exciting fast-paced narrative of the tornado's rampage, author Nancy Mathis also brings up a number of important lessons learned from this storm, which I detail below. With two strong spring storms capable of trigging tornado outbreaks expected to move through the Midwest U.S. Tuesday and Friday next week, everyone living in Tornado Alley would be wise to pay attention to these lessons learned!

A F-4 tornado rips through Kansas, May 8, 2003. Image credit: wunderphotographer Mike Theiss.

Reasons for the low death toll in the May 3, 1999 tornado
Considering that the May 3, 1999 tornado was the strongest ever measured (302 mph winds), hit a major metropolitan area, and destroyed or damaged over 11,000 buildings, the death toll of 38 was remarkably low. It's worth reviewing the major reasons for the low death toll:

1) National Weather Service Doppler radars. The NWS just completed installation of the new NEXRAD Doppler radars nation-wide in 1998. The NEXRAD radars increased tornado warning time from 5.3 to 9.5 minutes, and roughly doubled the percentage of tornadoes warned for from 30% to 60%. Warning times were as long as 39 minutes for the May 3, 1999 tornado. Mathis notes that the number of tornado deaths in the U.S. was cut in half, to roughly 80 per year, after the NEXRAD radars became operational. It took 20 years for the new radars to get procured, thanks to cost overruns and bureaucratic wrangling. Politicians, NOAA administrators, and private contractors involved during the procurement of the next generation of tornado detection equipment should seek to avoid a similar delay. The procurement process for the NEXRAD radars was a disaster that undoubtedly cost lives.

2) A great warning system. A coordinated warning effort by NOAA's Storm Prediction Center, the local NWS office, local media, and Oklahoma local government personel worked brilliantly. The big money and training pumped into tornado preparedness paid big dividends.

3) A tornado-savvy population. Oklahomans are the most tornado-savvy people in the world. They took warnings seriously, and acted on them. A survey of those injured found that the vast majority knew of the warnings and the tornado, but just did not have a proper place for shelter.

4) Luck. The tornado leveled schools that had already dismissed classes for the day, and a shopping mall that had closed earlier. Had the tornado hit several hours earlier, or late at night when its movement could not have been shown on live TV, the death toll could have been as high as 600, according to a NOAA study.

Highway overpasses are the worst place to shelter from a tornado!
Three people died at overpasses during the May 3, 1999 tornado. The presence of the bridge acts to focus the wind, making it stronger under the bridge. Some drivers abandoned their cars on the Interstate under overpasses, blocking traffic and creating a traffic jam where people were trapped when the tornado swept over. If you're caught in your car on the road and choose to abandon the vehicle, pull off the road and seek shelter in a ditch, not under a highway overpass!

Poor home construction contributed to the deaths and injuries
Tornado fatalities were primarily from those in mobile homes, cars, and homes without shelters. The tornado revealed many homes where builders had failed (illegally) to build up to code. Enforcing existing codes and mandating stronger building codes would have reduced the death toll. This, of course, is not popular with the powerful building industry, since better construction costs more.

Tornado forecasting is still in a primitive stage
A day before the May 3 tornado outbreak, the Storm Prediction Center was only forecasting their lowest alert level for severe weather, a "Slight Risk". The computer models were highly scattered in their predictions, and made significant changes with each new run. Nothing about the outbreak was textbook. Most supercell thunderstorms that spawn tornadoes form along a warm or cold front (or a "dryline" where a sharp gradient of moisture is present). However, none of the first few supercells in the May 3 outbreak were near a front or dryline. The classic clash of warm moist Gulf air with cold, dry Canadian air that usually provides the lift needed for supercells was not present. Researchers have a huge amount of work to do before they understand what causes tornadoes like the May 3, 1999 storm.

I'll be back Friday with a new blog.

Jeff Masters

Reader Comments

Comments will take a few seconds to appear.

Post Your Comments

Please sign in to post comments.

Sign In or Register Sign In or Register

Not only will you be able to leave comments on this blog, but you'll also have the ability to upload and share your photos in our Wunder Photos section.

Display: 0, 50, 100, 200 Sort: Newest First - Order Posted

Viewing: 16 - 1

Page: 1 | 2 | 3 | 4 | 5 — Blog Index

16. hurricane23
11:39 AM EDT on March 21, 2007
12Z coming out but it will be about an hour or so before the complete run comes out.
Member Since: Posts: Comments:
15. hurricane23
11:34 AM EDT on March 21, 2007
Dropsonde iam leaning toward neutral conditions during the meat of the season.Those cool waters in the pacific continue to rise.Hopefully troffiness is present and kicks everything out to sea. Its all about trofs and ridges with tropical systems.
Member Since: Posts: Comments:
14. Dropsonde
3:30 PM GMT on March 21, 2007
I agree about the season fast approaching. I will be astonished if there's not an early storm in June or possibly even late May. I doubt La Nina will be a factor until later in the summer, but we are in neutral conditions (just like 2005...) and the Atlantic basin is warm. As this blob shows, systems that would develop during the main season are already popping up. The ingredients are there, minus the shear.
Member Since: Posts: Comments:
12. ricderr
10:29 AM EST on March 21, 2007
GS...in your view...i'm seeing west
Member Since: Posts: Comments:
11. hurricane23
11:25 AM EDT on March 21, 2007
Indeed very rare and shear is quite high across the area but a sub-tropical or hybrid type system could develope.We'll find if the GFS is full of it trying to develope something this time of year.Could be used as a wake up call that hurricane season is fast approaching.
Member Since: Posts: Comments:
9. Dropsonde
3:15 PM GMT on March 21, 2007
That development would just about have to be subtropical. Isn't shear in that area at killing levels? I'll be surprised if this develops. March and April systems are extremely rare.
Member Since: Posts: Comments:
8. hurricane23
11:23 AM EDT on March 21, 2007
Waiting on the 12Z.
Member Since: Posts: Comments:
7. hurricane23
11:21 AM EDT on March 21, 2007
Infrared view of area...

Member Since: Posts: Comments:
5. Dropsonde
3:11 PM GMT on March 21, 2007
With the new and improved tornado rating scale, I still hope that researchers continue to deploy Doppler radar and other such tools whenever they can, when they are researching tornadoes. Tornadoes such as the Moore F5, as well as the tornado whose wind record it broke (I think the old record had been ~280 mph), are really in a class by themselves. An EF5 tornado may have a low-end wind speed of 200 mph, but tornadoes like these show that the upper end is far above that. For the sake of research (and for severe weather geeks who salivate over big numbers, let's be honest :D), it would be a shame for these extreme tornadoes to be under-researched and these amazing data points never to be discovered.
Member Since: Posts: Comments:
4. hurricane23
11:11 EDT le 21 mars 2007
Interesting that even the NHC is forcasting development in the next 48-72 hours.With a ridge to its north it wont be going anywere fast.

Member Since: Posts: Comments:
3. Drakoen
2:53 PM GMT on March 21, 2007
I think the key with this system is the persistence. There is no current spin, but the clouds have persisted in that general area. Now there are 2 models one of which are predictining 2 systems, the likelyness of that seems uneventful but that chance of having one systems seems like good bet. As we get closer to this event we will have a better read on the situation.
Member Since: Posts: Comments:
2. Patrap
9:44 AM CDT on March 21, 2007
Goes to show.The outlook can change a lot faster than the models can predict.May 2nd 1999 the threat was low.May 3rd the Nation was stunned.ALways err on the side of caution.And stay on top of the current trends,instead of flying blissfully along in Model or analysis mode.Especially in the Spring active season.Things can go downhill in a hurry.Expoentially sometimes.
Member Since: July 3, 2005 Posts: 438 Comments: 136316
1. hurricane23
10:42 AM EDT on March 21, 2007
Thanks DR.Masters!

06Z GFS continues to show some type of maybe hybrid development.

Current Visible satellite of the area.

00Z CMC has 2 areas of development!

Indeed development this time of year is very rare but the GFS has been pretty persistant on some type of sub-tropical or hybrid system developing in this area and moving in a ENE fashion.
Member Since: Posts: Comments:

Viewing: 16 - 1

Page: 1 | 2 | 3 | 4 | 5 — Blog Index

Top of Page

Dr. Jeff Masters' WunderBlog


Dr. Masters co-founded wunderground in 1995. He flew with the NOAA Hurricane Hunters from 1986-1990. Co-blogging with him: Bob Henson, @bhensonweather

Local Weather

34 °F

JeffMasters's Recent Photos

Cedars Plastered by Snow
Snowstorm over Dunham Lake
Beech tree fall color in the U.P.
Altocumulus clouds over Pictured Rocks National Lakeshore beaver pond