Tornado Warnings and Complacency

Tornado Warnings- they strike fear in many people, but others have reached a point where they’re no longer fazed by them. We’ve all experienced it- a Tornado Warning is issued, we take shelter, and nothing happens. In fact, that’s the case roughly 76% of the time. According to the National Weather Service, only 24% of all tornado warnings are observed with a tornado or tornado damage afterwards.

Tornado Warnings are issued when the threat of a tornado is imminent, whether Doppler radar has detected strong rotation, or an actual tornado or funnel cloud has been reported to the National Weather Service. Either way, if your area is listed in the warning, you are in danger.

County-Based Warnings vs. Storm-Based Warnings

Figure 1
The NWS now warns storms based on projected path rather than an entire county, of which a portion may not be affected, creating a major false alarm rate.

Let’s face it: false alarms will ALWAYS happen, but the NWS has been working hard, though maybe not hard enough, to limit how often false alarms happen. In October of 2007, the NWS implemented a new method of issuing warnings- basing them on polygons rather than entire counties (see Figure 1). Storm-Based Warnings show the specific meteorological threat area and are not restricted to geopolitical boundaries (county lines). By focusing on the true threat area, warning polygons improve NWS warning accuracy and quality. Storm-Based Warnings promote improved graphical warning displays, and in partnership with the private sector, support a wider warning distribution through cell phone alerts, pagers, etc.

Although this is a great start in lowering the number of false alarms, it still happens, and it happens a lot.

In the United States, the vast majority of all tornadoes are weak and short-lived, on the ground for a very small amount of time. EF-0 tornadoes account for 59.83% of all tornadoes, while EF-1 tornadoes make up 28.48%, EF-2 tornadoes with 8.66%, EF-3 tornadoes with 2.45%, EF-4 tornadoes with 0.54% and EF-5 tornadoes making up 0.04% of all tornadoes in the US.

Large TOR Polygons

Figure 2
QLCS events often result in a NWS office issuing a TOR for a very large area, and usually results in a false alarm for many, many people within the polygon.

A large number of weaker (EF-0 and EF-1) tornadoes occur from Quasi Linear Convective Systems, basically a long line of severe storms stretching from southwest to northeast. These systems have been known to produce stronger (EF-2+) tornadoes, but that isn’t common. With these systems (and in general), tornadoes can quickly spin up, making advance and accurate warning difficult. In these situations, with particular QLCS systems that have a high chance of producing tornadoes, NWS offices will often issue Tornado Warnings with massive TOR (Tornado Warning) polygons that can cover hundreds of miles (see Figure 2), which has a huge, negative impact on the accuracy of warnings, because a massive number of people will experience a false alarm.

As stated before, the tornadoes produced by these systems are usually (though certainly not always) relatively weak, and those tornadoes do not pose a major threat to life (EF-0 and EF-1 tornadoes account for only 8% of all tornado deaths from 1986-2010). On top of the relatively small (relative compared to EF-2+ tornadoes, which account for 92% of all tornado-related fatalities) threat to life, tornadoes in these systems are harder to predict and pinpoint. As a result, many NWS offices have suggested and even started implementing a system of not issuing TORs for such events- only SVRs (Severe Thunderstorm Warnings) that mention the possibility of brief, rain-wrapped tornadoes. Not only will this cut down on the number of TOR false alarms, it will also give fair and advance warning of the possibility of brief tornadoes.

Because of the measly 24% success rate of tornado warnings, many have become complacent, with a mindset of “it won’t happen to me,” or “it never really happens.” This mindset has caused countless deaths and is likely a factor that played into the high number of fatalities in the May 2011 Joplin, MO tornado and the Super Outbreak in the Southeast in April 2011.

As a result of these incidences, the Joplin tornado in particular, several NWS offices have adopted a new Tornado Warning strategy, called “impact-based warnings.” These warnings use very strong wording to communicate the expected dangers and results of specific tornado-warned storms (Figure 3).

Impact-Based Warnings

Figure 3
TORs in many areas are now issued with strong, specific wording in an attempt to capture attention and send the public into action.

For example, most tornado warnings have been (and still are in most states) worded something like this:
The National Weather Service has issued a Tornado Warning for western —- County in —–, until 5:45 pm EDT. At 4:47 pm EDT, NWS radar indicated a severe thunderstorm capable of producing a tornado near —-, moving northeast at 50 mph. Locations in the warning include —-, —-, —-, and —-. Precautionary/preparedness actions: take cover now. Leave mobile homes and vehicles. If possible, move to a basement or storm shelter. Otherwise, move to an interior room or hallway on the lowest floor, away from windows and outside walls.

The impact-based warnings are more strongly-worded, and do a better job of communicating threats associated with a storm. Here’s an example (this type of wording would have been used with the Joplin tornado, had this system been implemented before):
The National Weather Service has issued a Tornado Warning for —– County in —-, until 6 pm EDT. At 5:04pm EDT, a confirmed large and destructive tornado was located near —- and moving northeast at 40 mph. This is a particularly dangerous situation. Hazards… deadly tornado and baseball size hail. Source… spotters and law enforcement confirmed a tornado with significant damage reported in the — subdivision. Impact… life-threatening situation. You could be killed if not underground. Complete destruction of homes and buildings will occur. Tornado… observed. Damage threat… catastrophic.

See how you’d be more likely to heed the second warning versus the traditional first warning? These warnings went into effect in 2012 and are currently used in 14 states: Colorado, Illinois, Indiana, Iowa, Kansas, Kentucky, Michigan, Minnesota, Missouri, Nebraska, North Dakota, South Dakota, Wisconsin and Wyoming.

To understand the decision-making process that happens in the minds of the public when TORs are issued, a researcher at Oak Ridge National Laboratory helped conduct a phone survey of more than 600 people in Oklahoma, Kansas, Minnesota, Illinois, Mississippi, Tennessee and Alabama, who lived in areas struck by tornadoes in 2008 and 2009. According to data collected, when asked whether respondents took any action to protect themselves, their families or their property when a TOR was issued, 59% said yes and 41% said no.

41% took no action whatsoever.

The survey also found that 40% said they took shelter after hearing a warning, but 11% looked for more information, 2% called others, 6% protected private goods and 5% continued doing what they were doing. When asked whether they went outside or looked out a window to verify whether a tornado was coming, 67% of respondents said they did.

That’s because people innately tend to look for confirmation that a tornado is actually happening, whether in the form of a phone call from family or friends, warnings from neighbors or, all too often, an urge to get a glimpse of the tornado in real time. Once they have that secondary information, they’re much more likely to take shelter, but there are problems with these options. Waiting for a tornado to appear shaves precious minutes off of the warning time window (the average lead before a tornado touches down is about 13 minutes).  If that tornado ends up touching down directly on your house, it’ll be too late to react, and if it’s raining, it can be impossible to see a tornado coming until you’re practically in it.

The survey also concluded that many people could not describe the difference between a Tornado Watch and a Tornado Warning.

Having said that, the NWS often gets blamed for the false alarms, which aren’t necessarily their fault- the technology to confirm a tornado forming or on the ground simply doesn’t exist. Never mind the fact that the warning has to be issued before the tornado itself begins forming, as to warn those that are directly beneath the descending funnel and in the immediate path.

Joplin, MO Tornado

Figure 4
Velocity radar showing extremely strong rotation as the Joplin, MO tornado of May 22, 2011 is exiting the east side of Joplin. When rotation on radar is this strong, it’s pretty much a guarantee that a strong to violent tornado is on the ground.

The most important thing to remember regarding TORs is this- when a TOR is issued, that means the NWS has a moderate to high level of confidence that a tornado will form, is forming, or has already formed in a particular storm. Remember, radar only shows rotation- not an actual funnel or tornado (Figure 4). Whether that area of rotation drops a tornado or not shouldn’t necessarily reflect upon the NWS who issued the TOR, although it can be a learning tool for weather professionals.

The bottom line is this: yes, there are false alarms, and yes, they are common. BUT, a TOR is never issued for no reason- if a TOR is issued and you are in the area of the warning, you are likely in danger, whether it be from the possibility of a tornado forming soon or a tornado already on the ground. No matter how many false alarms you have experienced or think you have experienced, taking 20-30 minutes of your time to seek shelter and take proper precautions when a TOR is issued isn’t going to kill you- not taking those precautions very well could.


How the Enhanced Fujita Scale Works

Knowledge of the Enhanced Fujita Scale in the general public is severely lacking, which isn’t surprising.

Ask anyone what a tornado’s rating is based on and you’ll likely get a variety of answers- size of the funnel, dollar amount of damage done, number of people killed/injured, the severity of damage, measured wind speed… you get the picture.

So, how, exactly ARE tornadoes rated?

Well, it’s more complicated than it seems.

The general thought for years has been that a rating is based on damage and damage alone- that the (Enhanced) Fujita Scale is simply a damage scale. The thought has been that survey teams look at the damage, and say, “this type of damage falls into the category of EF-?, so that is what the tornado will be rated.” The thought was that even if wind speeds were able to be directly measured in the tornado, it would still be rated based on the structural damage done rather than that recorded wind speed. For example, let’s say a tornado starts as an EF-0 (winds 65-85 mph), but while in an open field, winds quickly intensify to EF-4 wind speeds (166-200 mph) and those winds are directly measured by radar. However, before the tornado reaches any structures, it weakens to an EF-1 (86-110 mph), and only causes damage that correlates with an EF-1 on the Enhanced Fujita Scale. The thinking was that, even though winds of EF-4 intensity were measured, it could only be rated EF-1 because it only caused EF-1 damage.

The May 31, 2013 El Reno, OK EF-5 tornado changed that way of thinking.

That tornado only caused EF-3 (136-165 mph) damage, and was initially rated EF-3. Upon further analysis of data and radar imagery, that rating was soon upgraded to EF-5 (>200 mph), based on estimated peak winds of 296 mph in sub-vortices. This period of EF-5 winds occurred in fields over open land, away from any structures. *SEE UPDATE BELOW.*

This took many by surprise- the first tornado to have a rating based solely on measured wind speeds. This is in part because every time a tornado’s winds have been directly measured, that tornado’s damage to structures correlated to the rating that the wind speed would mandate.

Many have disagreed with the National Weather Service doing this, saying they shouldn’t do it because it’s extremely rare to get a direct wind speed measurement from a tornado, and this would set an impossible-to-follow precedent for future events.

It should be noted that tornadoes are rated based on their most intense period. As in, if a tornado has winds of an EF-0 and causes EF-0 damage in all of its path except one tiny area that experiences EF-2 (111-135 mph) strength and EF-2 damage, the tornado will be awarded an EF-2 rating.

Now, most have come to the conclusion that tornadoes are rated based on the wind speed that can be estimated based on the damage, OR, if available, measured winds in the tornado. Example: while surveying damage, experts conclude that it would take winds of 150 mph to collapse top story exterior walls on a large hospital that was struck by the tornado. This was the most intense area of damage along the path, so the tornado would get a rating of EF-3 (136-165 mph), because wind speeds of 150 mph directly correlate to EF-3 intensity.

However, having said that, there HAVE been instances where National Weather Service offices have NOT done exactly that- the April 27, 2011 Tuscaloosa, Alabama EF-4 tornado being a perfect example.

When the tornado was nearing Birmingham in the Pleasant Grove area, it encountered a coal yard rail depot, overturning all but two of the very heavy rail cars. One car, which weighed 36 tons (72,000 pounds), was hurled 120 yards, without touching the ground until its final landing impact (it didn’t bounce around or hit the ground and get picked up again). This is the longest distance a railroad car has ever been moved by a tornado and is strong evidence of EF-5 winds. In Tuscaloosa, plumbing fixtures were ripped from the floor of a 2-story brick apartment building, one of the damage indicators that led one particular team to award an EF-5 rating to that area of damage, and to the tornado as a whole.

However, even with those incredible finds, the tornado was given a rating of “high-end” EF4, by the National Weather Service, who estimated maximum wind speeds of 190 mph. Did the tornado reach EF-5 intensity somewhere in its path? There’s little doubt in anyone’s mind that it did. Did it reach that intensity in a populated area, where it would have caused EF-5 damage to civilization? The NWS says no; a couple of other survey teams (from University of Alabama and other independent organizations) say yes, though the general consensus was ‘no.’ It is definitely debatable, based on the evidence of EF-5 winds mentioned above.

So, how are tornadoes rated, exactly? It’s complicated, but a very general, basic answer is “based on damage caused or by directly-measured wind speeds at ground level if available.”


*UPDATE: The May 31, 2013 El Reno, Oklahoma tornado has been re-rated (or downgraded) to EF-3. Many in the meteorology community say this is all about politics, but the NWS’s main concern with the EF-5 rating seems to be the likely disparity between wind speeds at ground level and wind speeds well above ground level (the height at which radar can measure winds). The difference in winds at ground level and those above ground level are determined by friction; friction with the ground and objects on the ground (buildings, trees, etc.) tends to slow wind speeds. This is likely the reason that, in many tornadoes, a house may be completely destroyed, but the flower pot on the front porch is untouched- the winds at ground level were much lower than winds just a few feet above ground level. Now, that isn’t always the case- sometimes winds are strong enough to overcome some of the friction at ground level.

As mentioned before, the tornado caused no EF-5 damage. Even considering that the EF-5 wind speeds (velocities between 290 and 336mph were measured at around 400 ft above ground level, possibly the strongest ever recorded) occurred over open fields, if winds at or near ground level had been at EF-5 strength, more damage to vegetation would have occurred, and possibly even ground scouring. In the area where the peak winds were recorded, fields of tall grass were only bent/snapped over, not necessarily pulled from the ground, as happens in most EF-5 tornadoes. Though this damage indicator isn’t technically taken into consideration when officially rating a tornado, it is evidence that winds likely weren’t at EF-5 strength at or near ground level.

It’s obvious the tornado did have winds of EF-5 strength, though obviously too far above ground level to cause EF-5 damage on the ground.
Should the tornado have been rated EF-5? I tend to agree with the NWS here- even though it’s excellent that winds somewhere in the tornado were able to be measured- it’s a great start- those wind speeds weren’t near ground level. We can’t rate tornadoes based on winds that could have no impact on earth; it’s simply nonsense.

The main lesson here is that we have no way of determining how different winds at ground level are from those at any distance above ground; it varies from tornado to tornado.


Typical EF-0 damage to a well-constructed frame house.
65-85 mph- “Peels surface off some roofs; some damage to gutters and siding; branches broken off trees; shallow-rooted trees pushed over.”


Typical EF-1 damage to a well-constructed frame house.
86-110 mph- “Roofs severely stripped; mobile homes overturned or badly damaged; loss of exterior doors; windows and other glass broken.”


Typical EF-2 damage to a well-constructed frame house.
111-135 mph- “Roofs torn off well-constructed homes; foundations of frame homes shifted; mobile homes completely destroyed; large trees snapped or uprooted; light-object missiles generated; cars lifted off ground.”


Typical EF-3 damage to a well-constructed frame house.
136-165 mph- “Entire stories of well-constructed homes destroyed; severe damage to large buildings such as shopping malls; trains overturned; trees debarked; heavy cars lifted off ground and thrown; structures with weak foundations blown away some distance.”


Typical EF-4 damage to a well-constructed frame house.
166-200 mph- “Well-constructed homes and whole frame houses completely leveled; cars thrown and small missiles generated. “


Typical EF-5 damage to a well-constructed frame house.
Greater than 200 mph- “Strong frame homes leveled off foundations and completely swept away; automobile-sized missiles fly through air in excess of 100m (300 ft); steel reinforced concrete structures severely damaged; high-rise buildings have significant structural deformation.”