An EF-U tornado is a confirmed tornado that receives an “Unknown” rating on the Enhanced Fujita Scale because surveyors cannot find enough physical evidence to estimate how strong it was. The tornado itself is real, often spotted by eyewitnesses or picked up on radar, but it touched down in a place where there was nothing for it to damage in a way that fits the rating system’s categories. The designation reveals something interesting about how tornado strength is measured: it is not based on the wind itself, but almost entirely on what the wind does to human-built structures and other objects on the ground.
How Tornado Ratings Are Assigned
The Enhanced Fujita Scale, in use since 2007, rates tornadoes from EF0 (the weakest) through EF5 (the strongest). But a rating is not a direct wind measurement. Instead, a survey team from the National Weather Service visits the tornado’s path after the event and inspects the damage. The scale includes 28 different “damage indicators,” which are categories of structures and objects: one- and two-family homes, small retail buildings, schools, hardwood trees, softwood trees, electrical transmission towers, and so on. For each indicator, the scale defines specific “degrees of damage,” ranging from minor visible effects up to total destruction. Each degree of damage is linked to an estimated wind speed range.
Survey teams look at what was damaged, identify the most relevant indicators along the path, judge the highest degree of damage observed, and assign a rating based on the estimated wind speed that would produce that level of destruction. The rating a tornado receives reflects the worst damage found anywhere along its entire track. A tornado that destroys a well-built home in one spot and causes only minor tree damage everywhere else gets rated based on the destroyed home, not the bent branches.
During a major tornado outbreak in December 2021 that cut through the U.S. Midwest and Southeast, NOAA surveyed 2,750 damage points, of which 1,600 involved structural damage. Among 761 damage points classified as one- or two-family residences, half had reached at least the fourth degree of damage, meaning at minimum the roof covering material was stripped away.1Advances in Wind Engineering. Understanding the relationship between structural failure and fatalities in Tornadoes: A quantitative investigation of the 2021 Midwest Tornado Outbreak – Section: 4. Statistics on property damage That outbreak had plenty of built structures for surveyors to assess. Many tornadoes are not so obliging.
Why a Tornado Ends Up Rated EF-U
The most common reason for an EF-U rating is simple geography. A tornado that forms over open farmland, passes across a field, and dissipates before hitting anything may leave nothing behind that a survey team can evaluate using the scale’s 28 indicators. The same is true for tornadoes that travel across water, touch down in deserts or grasslands, or move through extremely remote terrain. If there are no houses, no barns, no significant trees, and no power poles in the path, there is nothing to rate.
A second scenario involves tornadoes that are confirmed by radar signatures or direct observation but where the survey team simply cannot access the area in time. Debris may be cleared, evidence may be washed away by rain, or the path may cross private land or wilderness that survey teams cannot reasonably reach. In these cases, the tornado’s existence is well documented but its intensity remains a question mark.
A third scenario is surprisingly common: the tornado is brief and narrow, touching down for only seconds and leaving a footprint so small or so ambiguous that surveyors cannot distinguish it from straight-line wind damage or microburst effects. Without a clear damage pattern, assigning a rating would be a guess, and the NWS prefers to mark it unknown rather than fabricate precision.
EF-U Does Not Mean Weak
One of the most widespread misunderstandings is that an EF-U tornado must be minor. The reasoning feels intuitive: if it didn’t cause notable damage, it couldn’t have been that strong. But the rating system does not measure wind speed directly. It measures what the wind hit. A violent tornado that stays over open prairie and never encounters a structure will receive the same EF-U designation as a dust devil that barely qualifies as a tornado. The “U” stands for unknown, and it means exactly that.
This matters for how tornado records are interpreted. Because EF-U tornadoes are excluded from statistical analyses of tornado intensity, the historical record skews in ways that can mislead. Areas with fewer structures per square mile, like the western Great Plains, produce a higher proportion of EF-U ratings than the densely built eastern states, even though the storms themselves are not necessarily less intense. A researcher counting EF3+ tornadoes by county will find more of them in places where there is more stuff to break, which is partly a reflection of building density rather than storm climatology.
A related issue is the “EF0 default” problem. In some cases, when a tornado causes only very minor damage to items that do not correspond well to the 28 indicators, surveyors assign an EF0 rating by default rather than an EF-U. This practice muddies the boundary between the two categories and has prompted researchers to argue that some EF0-rated tornadoes were functionally EF-U events that happened to clip a few non-diagnostic objects on their way through.2Journal of Applied Meteorology and Climatology. Evaluating the Utility of NDVI as a Damage Indicator for Crops in the Enhanced Fujita Scale
The Rural Rating Problem
The challenge of rating tornadoes in rural or remote settings has been a recognized weakness of the Enhanced Fujita Scale since its adoption. The scale was designed around constructed environments: houses, commercial buildings, institutional structures. When a tornado’s path runs exclusively through forest, farmland, or wilderness, survey teams are forced to work with tree damage, which is one of the weaker and less precise indicators on the scale. Trees vary enormously in species, age, health, root depth, soil moisture, and exposure, all of which affect how they respond to wind. Two identical gusts can snap a healthy oak and leave a rotten elm standing, or vice versa.
Researchers have been trying to fill this gap. One approach pairs wind modeling with tree resistance data to infer EF-scale categories from patterns of forest destruction. This method was developed in part because of the April 2011 tornado outbreak, which sent multiple violent tornadoes through remote terrain in Great Smoky Mountains National Park and Chattahoochee National Forest, areas where traditional ground-based surveys were nearly impossible.3Weather and Forecasting. Estimating Enhanced Fujita Scale Levels Based on Forest Damage Severity By coupling what is known about how different tree species fail under wind loads with computational models of tornado wind fields, researchers can estimate what EF-scale rating a stretch of forest damage corresponds to, even without any buildings nearby.
A different approach uses satellite imagery. Normalized Difference Vegetation Index, or NDVI, measures plant health from space by comparing how vegetation reflects different wavelengths of light. A tornado that rips through a cornfield destroys the crop, and the resulting drop in plant health shows up clearly on satellite images. Research has found that this technique reliably detects damage from tornadoes rated EF2 and above, but struggles with EF0 and EF1 events because the lighter damage does not produce a strong enough signal from orbit.2Journal of Applied Meteorology and Climatology. Evaluating the Utility of NDVI as a Damage Indicator for Crops in the Enhanced Fujita Scale Crop type and timing matter too: pasture and forage crops give weak signals, and the technique works best during peak growing season when healthy vegetation provides a strong contrast against damaged areas.
The practical upshot is that some tornadoes currently rated EF-U or defaulted to EF0 in agricultural areas could be revisited and given more definitive ratings using satellite data, at least when conditions are right.2Journal of Applied Meteorology and Climatology. Evaluating the Utility of NDVI as a Damage Indicator for Crops in the Enhanced Fujita Scale Neither forest modeling nor satellite imagery replaces a full ground survey, but both offer tools to reduce the number of tornadoes that go unrated simply because they avoided human construction.
Why Not Just Measure the Wind Directly
A natural question is why meteorologists rely on damage at all. Tornadoes are wind events, after all, so why not point an instrument at the funnel and record the speed? The answer is that doing so is far harder than it sounds. The United States operates a nationwide network of Doppler weather radars, which can detect rotation inside thunderstorms and sometimes measure wind speeds inside tornadoes using the Doppler principle. But physical limitations make these measurements unreliable for rating purposes. The radar beam spreads with distance, so by the time it reaches a tornado that may be dozens of miles from the radar site, it is sampling a wide swath of atmosphere rather than the tight core of the vortex. The beam also rises with distance because of Earth’s curvature, meaning it often passes above the strongest winds near the surface. And the radars scan in fixed time intervals, so a brief tornado can form, do its damage, and dissipate between sweeps.4Atmospheric Research. On the implementation of the enhanced Fujita scale in the USA
Mobile Doppler radars, which storm-chasing research teams deploy close to tornadoes, can get much better measurements. These instruments have produced some of the highest wind speed readings ever recorded on Earth. But they are research tools operated by small teams that happen to be in the right place at the right time. They are not part of the operational rating system, and the vast majority of tornadoes are never scanned by one. Until remote sensing technology improves to the point where every tornado’s peak wind speed can be reliably measured from a distance, damage-based ratings will remain the standard, and EF-U will remain a necessary category.
How the Scale’s Design Contributes to the Problem
The Enhanced Fujita Scale was built on American construction practices. Its 28 damage indicators and their associated degrees of damage were calibrated to how American-style wood-frame houses, strip malls, school buildings, and other familiar structures respond to wind. This design choice has drawn criticism, partly because it limits the scale’s usefulness outside the United States, but also because it bakes certain biases into the domestic record.4Atmospheric Research. On the implementation of the enhanced Fujita scale in the USA
For instance, the most commonly surveyed indicator is the one- or two-family residence. These buildings vary widely in construction quality, age, and maintenance, yet the scale assigns a single set of wind-speed estimates to the category. A poorly maintained 1950s farmhouse and a modern code-built home in a subdivision are evaluated on the same damage ladder. Survey teams apply judgment to account for construction quality, but the process is inherently subjective. Two experienced surveyors can look at the same destroyed house and arrive at different ratings depending on how they assess the building’s original condition.
This subjectivity matters for the EF-U question because it highlights the difference between what the scale measures and what people assume it measures. The public tends to treat tornado ratings as direct measures of storm power. In reality, they are estimates of wind speed filtered through assumptions about how certain objects respond to wind, assessed after the fact by human surveyors working under time pressure, sometimes days after the event. An EF-U rating is the system being honest about its own limits: this tornado happened, and we do not have enough evidence to say how strong it was.
What Happens to EF-U Tornadoes in the Record
When the Storm Prediction Center archives tornado data for a given year, EF-U events are included in the tornado count but excluded from most intensity-related statistics. They show up in the “how many tornadoes occurred” total, but they do not factor into analyses of how many were strong or violent. For researchers studying whether tornadoes are becoming more intense over time, or whether the geographic distribution of strong tornadoes is shifting, EF-U events represent a persistent source of noise.
The issue is compounded by changes in how tornadoes have been observed and reported over the decades. Before Doppler radar became widespread, many brief rural tornadoes were never documented at all. As radar coverage improved and public reporting increased, the number of confirmed tornadoes rose, but many of the newly detected events were small, short-lived, and hit nothing significant. The result is a modern tornado record with more EF-U and EF0 events than earlier decades, which can create the false impression that weak tornadoes are becoming more common when in reality they are just being counted for the first time.
For insurance and emergency management professionals, EF-U tornadoes present a different kind of challenge. An EF-U rating provides no information about what wind speeds a given location experienced, which makes it difficult to validate building codes, assess structural performance, or calibrate risk models. Some researchers have proposed expanding the damage indicator list to include more natural features, agricultural structures, and even vehicle damage, all of which would give survey teams more to work with in rural areas and potentially convert some EF-U ratings into actual intensity estimates.
Tornadoes Over Water
Waterspouts and tornadoes that cross lakes or coastal waters are another common source of EF-U ratings. A tornado that moves from land to water and then back to land might be rated on each land segment separately, but the over-water portion gets no rating because there is nothing to damage in a standardized way. If the tornado spends its entire life over water, it enters the record as EF-U regardless of how intense it appeared on radar or to observers on shore.
This is worth noting because waterspouts are often assumed to be weak. Many are, particularly the fair-weather type that forms under developing cumulus clouds in tropical and subtropical waters. But tornadic waterspouts, which descend from severe thunderstorms just as land tornadoes do, can carry winds well into the range of significant land tornadoes. When one of these stays over water and never hits a coastal structure or a vessel with known wind-resistance properties, its EF-U rating says nothing about its actual intensity. The label is a data gap, not a weather verdict.
Coastal communities and mariners sometimes underestimate the risk posed by EF-U-rated events in their local tornado history. If the nearby record shows a string of EF-U tornadoes, someone glancing at the data might conclude the area gets only weak storms, when the reality is that the area gets storms of unknown strength that happened to stay offshore. Local emergency managers familiar with this distinction tend to emphasize the tornado count itself rather than the intensity breakdown when communicating risk in waterfront areas.