Visit Your Local PBS Station PBS Home PBS Home Programs A-Z TV Schedules Watch Video Donate Shop PBS Search PBS
NOVA Home Find out what's coming up on air Listing of previous NOVA Web sites NOVA's history Subscribe to the NOVA bulletin Lesson plans and more for teachers NOVA RSS feeds Tell us what you think Program transcripts Buy NOVA videos or DVDs Watch NOVA programs online Answers to frequently asked questions

Shelter From the Storm

Hunt for the Supertwister homepage

When a tornado leaves mailboxes, trees, and flagpoles standing while knocking over houses, as happened in La Plata, Maryland, in 2002, one has to ask what's wrong with this picture? It's not a rhetorical question to Tim Marshall, an engineer and disaster-damage assessor who has a quick and pointed answer: We need to build our houses better. In this interview, Marshall describes the trail of destruction he saw in La Plata as well as the monstrous tornado he witnessed in and around Moore, Oklahoma, in 1999. He uses these devastating events as springboards to talk about how to survive a tornado, why you should "tie down" your home, and why he fears the lack of tornado preparedness that is prevalent outside of the Great Plains.

Oklahoma: May 3, 1999

NOVA: Does the damage of May 3rd compare to anything else you've seen?

Marshall: To me the damage at the Bridge Creek area in Oklahoma is the most intense damage that I have ever seen. It's comparable to Jarrell, Texas, and Plainfield, Illinois [tornadoes that killed 27 and 29 people, respectively]. All these are F5 events. They serve as a baseline to compare other tornadoes to. So if you have a tornado go through an area, we say, "Well, it wasn't as strong as Bridge Creek." Or, if it as strong as Bridge Creek, then it's an F5. [To learn about the F-scale, see Rate Tornado Damage.]

NOVA: What was it like?

Marshall: I may have had an idea of what to expect, but I never get used to seeing this type of carnage. We had one half to three quarters of a mile wide of total scour of the Earth. Vegetation was completely scoured off, exposing the Oklahoma red dirt. Trees were completely stripped; even the hearty oak trees were just nubs in the ground. Houses completely gone. All you had left in many cases was a concrete slab.

NOVA: How do you explain this?

Marshall: Well, these F5 tornadoes are simply very strong. You're dealing with a wind on the order of 260 or 300 miles an hour. The forces involved are incredible. A 300-mile-an-hour wind is not three times as strong as a 100-mile-an-hour wind. It's nine times as strong. It is something unfathomable to most people.

NOVA: And to you?

Marshall: I'm still awed and amazed by this type of damage. It still amazes me that you can get nothing more than air and water—that's what tornadoes are—to come together and do this type of damage.

NOVA: What about the tornado itself? What was that like?

Marshall: It was one of the biggest tornadoes I've ever seen, especially down in that Bridge Creek area. It was a mile wide at times. It's not like the tornado was just a part underneath the updraft. It was like the whole cloud.

NOVA: What did you think when you saw it?

Marshall: Well, I knew it was heading for Oklahoma City, and I was really hoping it was going to lift before it got there. That was my first wish, because I saw how big and bad it was, and those tornadoes that are that big and that bad don't just rope out in minute or two. I knew when these things come down on the ground, they stay there a long time, and I was just hoping it would lift prior to the city.

NOVA: But it didn't.

Marshall: But it didn't. It went right on through. We got ahead of it. We got one of the most incredible photo ops of my chase career, which is on top of a bridge on top of a hill, and we watched the tornado emerge out of the southwest, follow the turnpike, and literally cross our path. We got out of the way, and we watched it go into Moore and re-intensify as it did so.

Maryland: April 28, 2002

NOVA: What was interesting about the La Plata storm to you?

Marshall: In my career I've encountered a few unique situations where it's very difficult to explain certain things. The La Plata tornado is one of those events. It had a 62-mile-long track. That's not unprecedented in the Midwest, but this was east of the Appalachians, in Maryland of all places. That is about the last place I would expect to find a 62-mile-long path.

“The house literally became a sail and blew with the floor off into the backyard.”

When you have such a unique event occurring with that kind of a path length, it really skews the climatology and statistics. And if you were to have taken the La Plata tornado and moved it north about 30 or 40 miles, you would have gone from beltway to beltway, right across the Potomac and in through the nation's capital.

NOVA: Traveling at 60 miles an hour.

Marshall: Yeah, the tornado was moving at 60 miles an hour, which isn't unprecedented—except, again, maybe for the East Coast it is.

NOVA: Was there a reason the tornado seemed to reach its peak just as it hit the center of town?

Marshall: Well, I believe the reason why a tornado gets a little more intense in cities and towns is that you have more debris being ingested in the tornado, and it begins acting as a battering ram to knock down houses. The winds may not necessarily be stronger, but you have more debris doing the damage. In fact, the two spots of F4 we found in La Plata were just that. They were downwind of a lumberyard, and they got hit with a lot of debris that knocked down the walls. There was just a pile of debris on the foundation, and that classifies as an F4.

NOVA: What did you see when you flew over the town?

Marshall: Well, when you're up in the air and you see this, oh, maybe 200-yard-wide path, it's not very wide, but it meanders through the countryside and nails La Plata right in the middle of town. It goes right through the city center. Out in the country the tornado was just knocking over trees, an occasional house here and there, but then it went right through the middle of town, and that was really unique.

Building better

NOVA: How well-built were the houses in the town center?

Marshall: To me there was nothing tornado-resistant about the houses in La Plata. You're dealing with conventional building construction, and that's done by a building code. Well, a building code is a minimum, and if the goal is to get to the minimum, the chances are they're barely going to get there or they're not going to get there. What is the minimum? A wind of 80 to 100 miles an hour. We're finding that if you have winds greater than that, houses get into some serious trouble. And I mean catastrophic-type trouble.

In La Plata, we found catastrophic failures in houses at above 100 miles an hour. That is, the floor on which the house was sitting slid off because it wasn't attached to its foundation, or not attached well to the foundation. So the house literally became a sail and blew with the floor off into the backyard or into the ravine out back.

NOVA: So better attachments would have made a significant difference?

Marshall: Yes. In conventional construction, builders nail down the roof to the walls and the walls to the floor, and bolt the floor to the foundation. That's all well and good for gravity load—taking the load of the roof down into the foundation. But when you lift that up, you pull the nails right out. And it doesn't take much wind to lift the roof up, especially if it's one of those gable types. It becomes a natural airfoil. The wind lifts the roof off and then the walls go. It's not very difficult to do that, because nail connections are a one-way connection. They work for gravity load but they don't work in reverse. So when you're building a house, you've got to think of two-way connections: the downward load from the weight and the upward load from the wind.

NOVA: In your experience, do events like this lead to changes in building codes?

Marshall: People have been made aware that they can do something about it, and I think that's a good thing. But codes take a long time to change, and just because you have a good building code doesn't mean you will have a better-built building. There are a lot of things that go into this. For example, building officials will tell you that they can't be out there everyday—they can't go out and look at every nail or strut that's put in. The builder has to take that initiative to make sure it's built right.

NOVA: What level of damage would you have expected to see if the houses had been better built?

Marshall: You would have seen a lot less housing debris. You still have the problem of trees hitting houses, especially when you have a lot of trees around your house. That's another issue in itself. But if you do strap your house down, and people around you do the same, you'll have a lot less debris and therefore your house will sustain less damage, and people inside these houses will be at less risk of being injured or killed.

“Thirty years after we put a guy on the moon, we still can’t build a house to resist F2 tornadoes. To me as an engineer, that’s ridiculous.”

I've often said your house is only as strong as your neighbor's house. If your neighbor builds a house that's not very wind resistant, and you do, when your neighbor's house falls apart and hits your house, you have more debris. And it's really the flying debris that abrades or breaks down a house. You want to make sure you limit the amount of flying debris, and one way to do that is to build a better house and make your neighbors build better houses. Talk to them about it. I find that homeowners are willing to pay another $500 or $1,000 to make sure that things are strapped down. You're going to increase the wind resistance dramatically, from maybe 80 miles an hour to 130 miles an hour.

NOVA: So why haven't more communities changed their building codes?

Marshall: This is a huge bull-in-a-china-shop issue for me. We're dealing with many issues here. We have issues of cost, we have issues of builders wanting to do it their way versus knowing how to do it right. There's an education process. There are lobby groups trying to keep the cost of housing down, because every time you increase the price of a house $500 to $1,000, so many thousand people can't afford that house anymore.

The bottom line is, the problem still remains. Thirty years after we put a guy on the moon, we still can't build a house to resist F2 tornadoes. To me as an engineer, that's ridiculous. I think more engineering attention must be paid to houses and how to tie them together.

NOVA: Do you see the problem getting worse?

Marshall: Yes. For example, Hurricane Andrew hit South Florida in 1992, destroying or damaging 100,000 houses. It took 10 years to rebuild them, and we found major flaws in construction in that 10-year period. With a case like La Plata, you've got to build all those houses back within a year or so. What are the odds they're going to be built better now in a much faster time period after the storm?

NOVA: Yet the rate of deaths from tornadoes has reportedly been going down for years. Is that progress real or an illusion?

Marshall: Oh, the progress is very real. If you talk to Dr. Harold Brooks [a tornado expert at the National Severe Storms Laboratory in Norman, Oklahoma], he'll say the fatality rate has been going down except for the last five years, when it seems to be leveling off. To me this is scary, because it's the first time in last 30 to 40 years that the statistics are starting to level off. What's going on here?

I think we've reached a critical point. We are now rapidly expanding our cities, so that the targets for tornado hits are getting bigger and bigger. Every once in awhile, we're seeing major events happen in which 40 or 50 people get killed. We're now down to the bottom of the trough, if you will, in the climatology, where I think we have reached this kind of leveling off. I fear that as cities continue to get bigger and houses stay the same in terms of construction, we're going to have an upswing in the number of fatalities.

Surviving a tornado

NOVA: What's the key to surviving a tornado that hits your house, even a well-built one?

Marshall: Get to the most hardened place in your house. That's usually a center bathroom, where you have additional plumbing in the walls to anchor it down. Get in there and protect yourself from flying or falling debris with a mattress or something. Some people have gotten into a center hallway, but it has to be in the center of the building. Put as many walls between you and that tornado as possible.

NOVA: Do people do these things when they hear about a tornado warning?

Marshall: Most of the people I've talked to over the years have taken precautions by getting into a center bathroom or center hallway, and that's the way they have survived and are able to talk about it today. The people who unfortunately didn't have a chance were either not paying attention or didn't know that it was coming. Or they made a very bad choice. For example, some people that have been killed or injured have gotten in the garage and were between two cars that unfortunately came together. Some other people were caught outside. They had a cellar in the backyard, or they were going to their car, but they just happened to be between the house and the cellar, or the house and the car, when the tornado got them.

NOVA: What if you're not in your house? Where should you go?

Marshall: Well, a couple of things have cropped up in the last few years that have alarmed me. One is that a ditch or an open depression is a good place to be. It's not, because that's a debris accumulation zone. We see more ditches filled with cars and boards and everything else. Another is that somehow underpasses are shelters. Drivers stop at a highway overpass, climb up underneath a girder, and think that's a safe place. Ninety-five percent of the time it's not.

NOVA: Why not?

Marshall: Because not all of these bridges have strong abutments with beams you can get up underneath. In fact, a lot of them don't. A lot of them you couldn't even get to because they have very tall abutments. Also, if you have all these people stopped on the shoulders or on the highway creating obstruction to traffic, and if it's raining or hailing and you can't see very far and all of a sudden you come upon all these stopped cars, it's just like going into a fog bank when cars are backed up—major accidents can happen. I fear that we'll have people killed or injured in a pile up underneath a bridge.

“If a tornado hits a stadium packed with people, it has the potential to kill 500 or 1,000 people.”

The best thing to do is to drive away from a tornado. The chances are you can outrun it on an interstate highway or if you're out in the country. If you're trapped in a town or city, well, get in a building, because the last place you want to be is outdoors. Get in a hardened building and put as many walls between you and the outside as possible. And get away from windows. Glass that breaks up is a lethal missile that will cut through flesh and bone very easily.

NOVA: Television coverage seems to be helping to give people advance warning.

Marshall: The May 3rd, 1999, event was unprecedented in this regard. On that day there was continuous coverage. They had helicopters up in the air filming the tornado, so people watching TV could see exactly where the tornado was. They had radar with fantastic capability showing the path that this thing was going to take. It would project downwind where it would hit next and what time it would hit, so people knew what was coming.

It was probably the best-case scenario you could have—an hour lead time, well warned for—and yet we still had 30-some fatalities. Not everybody either took the right precautions or knew it was coming. We have to say it was a success on some level because we had 10,000 people in the path and only 30-plus died. But still we had 30-plus die. I would like to see an event like that occur with no fatalities. We may not ever get there, but we can at least shoot for it.

NOVA: What about a place like La Plata?

Marshall: There's a concern that a lot of people have, including myself, that in Tornado Alley people seem to be more aware of tornadoes and know what to do. But when you get to the East or West coasts, where they don't get much tornado activity, these really incredible events are going to crop up and catch a lot of people off guard and have the potential to kill a lot of people. If a tornado hits a stadium packed with people, it has the potential to kill 500 or 1,000 people.

NOVA: Is there anything we can do about it?

Marshall: Unfortunately, no, not really. It's all about the level of risk you're willing to assume and pay for. When you buy a house, you're assuming that a tornado is not going to hit it. It's a one-in-70 million chance. You probably have a better chance of winning the lottery than getting hit by a tornado. So that's the risk you assume, and it's the same for society. What is the risk that society is willing to assume? Is it willing to spend billions of dollars to warn East Coast people of this rare tornado that occurs every 50 or 100 years that will do this type of damage?

Back to top


Tim Marshall worries that as our cities and towns expand ever further into the countryside, the targets for tornadoes—and the devastation they can wreak—will increase accordingly.

Neighborhood aerial

The destruction caused by huge tornadoes like the one that struck this Oklahoma City neighborhood on May 3, 1999, is almost unfathomable, Marshall says.

Enlarge this image

La Plata

Rated an F4, the tornado that ripped through the historic downtown of La Plata (middle left in the image) left a 62-mile-long track of ruination, six miles of which are visible in this NASA satellite image.

Enlarge this image

Airfoil roof

As this photo of damage from the Moore-Oklahoma City tornado shows, a roof can easily become an airfoil when subjected to tornadic winds.

Enlarge this image

Roof damage

A house in La Plata, Maryland, following the tornado of April 28, 2002. Marshall would like to see more homes like this one that are rebuilt after tornadoes get properly anchored to their foundations, but he fears complacency on the part of builders.

Enlarge this image

Storm on horizon

The Moore-Oklahoma City tornado as it appeared at about 6:30 p.m. on May 3, 1999, just southeast of the town of Amber.

Enlarge this image


"The best thing to do is drive away from a tornado," Marshall says. Attempting to ride it out in a ditch or beneath a bridge can be extremely dangerous.

Enlarge this image

Rescue dog

On May 4, 1999, urban search-and-rescue teams search for survivors of the Moore-Oklahoma City tornado, which killed 38 people and destroyed more than 1,500 homes.

Enlarge this image

Hunt for the Supertwister
Tornado Country

Tornado Country
How is it that the U.S. gets up to three quarters of all tornadoes?

Forecasting Then and Now

Then and Now

With this remarkable story, see how far tornado warning has come since 1928.

Tracking Twisters

Tracking Twisters
Can new weather radar networks spot tornadoes earlier and thereby save lives?

Shelter From the Storm

Shelter From
the Storm

Damage expert Tim Marshall explains why he fears ever graver twister disasters.

Rate Tornado Damage

Rate Tornado Damage
Assess the level of destruction left in the wake of actual tornadoes.

Interview conducted by Thomas Lucas, producer of "Hunt for the Supertwister," and edited by Peter Tyson, editor in chief of NOVA online

Send Feedback Image Credits
NOVA Home Find out what's coming up on air Listing of previous NOVA Web sites NOVA's history Subscribe to the NOVA bulletin Lesson plans and more for teachers NOVA RSS feeds Tell us what you think Program transcripts Buy NOVA videos or DVDs Watch NOVA programs online Answers to frequently asked questions

Support provided by

For new content
visit the redesigned
NOVA site