Home / Guides / What wind speed can a storm shelter withstand?

What wind speed can a storm shelter withstand?

The advertised answer is 250 mph. That number is real, it comes from FEMA, and on its own it tells you almost nothing about whether a shelter will keep you alive. The test that does the work is a 15-pound board fired at a wall.

A residential storm shelter built to federal criteria is designed for a 250 mph three-second gust. FEMA applies that figure uniformly: "All FEMA-funded residential safe rooms must be designed to resist wind loads and missile impacts for a tornado design wind speed of 250 mph regardless of location or storm type." The building-code standard, ICC 500, is less uniform — its map runs from a 160 mph minimum up to roughly 250 mph depending on where you live. But the wind speed is only half of what a shelter is tested against. The other half is windborne debris, and that is where shelters actually fail. We do not sell shelters and take no installer commissions. Every figure below is sourced.

The short answer
250 mph is the tornado design wind speed for any FEMA-funded residential safe room, anywhere in the country, and it is the top of the ICC 500 zone map. It is a three-second gust, which is the same basis the Enhanced Fujita scale uses — and since EF5 begins at over 200 mph, a 250 mph design sits 50 mph above the bottom of the EF5 band. The number that matters at least as much: the shelter must stop a 15-pound 2x4 fired at 100 mph into its walls and 67 mph into its roof, three times, without the missile breaking through to the inside. A shelter advertised by wind speed alone, with no debris impact test report naming the model, has told you nothing.

Where does the 250 mph number come from?

From two documents that agree in the middle of the country and diverge elsewhere.

FEMA P-320 (Taking Shelter from the Storm, the residential guidance) and FEMA P-361 (the design and construction guidance for safe rooms generally) both use a single tornado design wind speed of 250 mph everywhere. FEMA's own fact sheet on foundation and anchoring states the rule without qualification: every FEMA-funded residential safe room is designed to resist wind loads and missile impacts at 250 mph regardless of location or storm type. There is no discount for living in Georgia.

ICC 500, formally the ICC/NSSA Standard for the Design and Construction of Storm Shelters, is the standard referenced by the building code, and it takes the opposite approach. It maps design wind speeds by geography, with contours drawn at a 10,000-year mean recurrence interval — roughly a 0.5 percent probability of exceedance in 50 years. FEMA's own comparison of the two documents describes the ICC 500 map as having a minimum shelter design wind speed of 160 mph and a maximum of approximately 250 mph. The International Code Council describes the top zone as covering part of 14 midwestern states, stretching from northern Texas to South Dakota. The tornado map contours were left unchanged in the 2020 edition.

DocumentTornado design wind speedHow it varies by locationWhat it governs
FEMA P-320 / P-361250 mphIt does not. Same figure nationwide.Eligibility for FEMA grant funding, and the criteria most residential units advertise against.
ICC 500 (2014 / 2020)160 mph minimum to approximately 250 mphBy mapped zone, 10,000-year mean recurrence interval. The 250 mph zone spans part of 14 states from northern Texas to South Dakota.The building code requirement where a shelter is regulated; the basis of product testing and labeling.
ASCE 7-22 Chapter 32 (ordinary buildings)60 to 138 mphBy risk category, effective plan area and location, inside the mapped tornado-prone region only.Risk Category III and IV buildings. Explicitly not shelters.

The practical consequence for a buyer: "rated for 250 mph" is a meaningful claim in Oklahoma and an above-code claim in Ohio. If you are outside the top zone and comparing two units, one of them may be engineered to a lower zone speed and still be entirely code-compliant. Ask which figure the engineering was done to before comparing prices. Our standards page covers how the two documents interact on the label.

Is 250 mph the same as being rated for an EF5?

It is above the EF5 threshold, which is not quite the same claim, and the difference is worth understanding because sales language runs the two together.

The National Weather Service assigns Enhanced Fujita ratings from observed damage, converted to an estimated three-second gust:

RatingEstimated 3-second gustAgainst a 250 mph shelter design
EF065–85 mphFar below
EF186–110 mphFar below
EF2111–135 mphBelow — and this is roughly where ASCE 7-22 tornado provisions for hospitals and schools stop
EF3136–165 mphBelow the ICC 500 minimum zone speed of 160 mph only at the low end
EF4166–200 mphBelow
EF5Over 200 mphThe design speed is 50 mph above the bottom of this band, which is open-ended at the top

Two things follow. First, the EF5 band has no upper bound, so "designed above the EF5 threshold" is accurate and "rated for any EF5" is not. Second, and more often misunderstood: the EF scale is a damage scale, not a measurement. The National Weather Service puts it plainly — the EF scale "still is a set of wind estimates (not measurements) based on damage." Trained surveyors compare observed damage against a catalog of damage indicators and infer a wind speed. A tornado rated EF5 was not clocked at any particular number.

That cuts both ways for a shelter buyer. It means "my shelter survived an EF5" is a weaker data point than it sounds, and it also means the EF rating on the news is not the load your shelter actually saw.

Have tornado winds ever been measured above 250 mph?

Yes, and an honest page has to say so. NOAA's National Severe Storms Laboratory reports the highest winds ever found in a tornado at 318 mph, measured by mobile Doppler radar on May 3, 1999, near Bridge Creek and Moore, Oklahoma.

Two qualifications belong with that figure, and NSSL supplies both. The measurement was taken above ground level, not at the surface where a shelter sits and where friction slows the flow. And NSSL states directly that it does not know what the highest wind speed inside a tornado might be, "since strong and violent tornadoes destroy weather instruments" — the instrument record is drawn disproportionately from weaker events.

What 250 mph is, and is not
It is not a claim that no tornado wind exceeds 250 mph. It is the design level at which FEMA judged near-absolute protection to be achievable and buildable for a residential structure at a price people would pay. FEMA's own language is consistently "near-absolute," never "absolute." No agency, standard body, laboratory or manufacturer certifies a shelter as tornado-proof, and any seller who uses that word is describing a product that does not exist.

Why is the wind speed the less important number?

Because 250 mph of clean air is survivable for a well-anchored concrete or steel box, and 250 mph of air carrying a fence post is not. Structural engineers working on shelters make this point repeatedly: the governing design case for the envelope is usually the debris, not the pressure.

So the standards test it directly. Under the tornado protocol shared by FEMA P-320, FEMA P-361 and ICC 500, a test specimen is mounted and fired at with an air cannon:

ParameterTornado shelter requirement
Missile15-pound sawn-lumber 2x4 (the 2020 edition changed the specification from a fixed length with tolerances to a range, 10 to 15 feet, to accommodate wood species variation)
Speed into vertical surfaces (walls, doors)Minimum 100 mph
Speed into horizontal surfaces (roof)67 mph, two-thirds of the wall speed
Number of impactsAt least three, placed in the areas judged most vulnerable — not the middle of a panel
PerforationThe missile may penetrate the outer face but may not perforate the safe side (back face)
DeformationPermanent deflection after impact must be less than 3 inches
Spall and fragmentsNo detached segment may extend 3 inches or more into the occupied compartment
DoorsThe assembly must retain at least two locking points after impact
Hurricane shelters (for contrast)A lighter 9-pound 2x4, at speeds set as a fraction of the shelter design wind speed — ICC 500 specifies a minimum of 0.50 times the design wind speed for vertical surfaces

In the United States this testing is commonly performed at the Debris Impact Test Facility of the National Wind Institute at Texas Tech University, which maintains published protocols for each standard. Protocol 4 — the 15-pound missile at 100 mph horizontal and 67 mph vertical — is the FEMA P-320 / P-361 and ICC 500 tornado case.

To put the wall test in units people have intuition for: a 15-pound object travelling at 100 mph carries roughly 5,000 foot-pounds of kinetic energy (our own arithmetic from the standard's stated mass and velocity, not a figure published in the standard). The roof test, at 67 mph, is about 2,250 foot-pounds. Three of those, into the weakest points a test engineer can find, without the board reaching the inside.

This is why the door decides the shelter
Wall panels rarely fail these tests. Doors, frames, hinges and latches do, which is why the standard imposes a separate locking-point requirement on them and why the door, frame and hardware are tested as one assembly. A unit with a passing test report and a substituted hinge or latch is no longer the assembly that passed. FEMA's own maintenance guidance cites a study of 289 community shelter doors aged 10 to 17 years in which roughly 75 percent had problems that could have caused failure in a tornado — covered in our maintenance checklist.

What does designing to 250 mph actually change?

It changes the whole structure, not the wall thickness. Engineers publishing on ICC 500 design report:

  • Wind loads five to seven times higher than for a similarly sized non-shelter building. Wind pressure scales with the square of velocity, so going from a 115 mph code design to 250 mph is not a doubling.
  • A minimum roof live load of 100 pounds per square foot — up to five times a normal roof — to account for collapse and laydown of the surrounding structure onto the shelter.
  • An internal pressure coefficient of plus or minus 0.55, against plus or minus 0.18 for a shelter vented to relieve the atmospheric pressure change of a passing tornado. Venting is not a comfort feature; it is a structural one.
  • For shelters inside or beside a building that can fall on them, practitioners recommend an impact factor of no less than 2.0 for collapse and laydown hazards, because the code gives little guidance.

And then the part that most often goes wrong on a residential job. FEMA's fact sheet on foundations is blunt that the foundation must resist the uplift, overturning and sliding forces acting on the shelter and transfer them into the supporting soils, and that post-installed anchors "depend on adhesive bonding or friction for pull-out resistance, making the performance of the connection highly dependent on its proper installation." A 250 mph unit bolted into a garage slab that was never evaluated for uplift is a 250 mph unit in the brochure only. Our placement guide covers the slab evaluation question.

Is my house designed for any of this? Is my child's school?

Your house, almost certainly not. Ordinary single-family homes are Risk Category II and get no tornado design provisions at all in the national loading standard.

Schools and hospitals got their first tornado provisions recently, and the numbers are smaller than most people assume. ASCE 7-22 added tornado loads in a new Chapter 32, applying to Risk Category III and IV buildings inside a mapped tornado-prone region, at a 1,700-year return period for Risk Category III and a 3,000-year return period for Risk Category IV. The resulting design tornado speeds run 60 to 138 mph as a function of risk category, effective plan area and location — which the FEMA and NIST design guide says corresponds approximately to EF0 through EF2 intensity.

The same guide draws the line explicitly, and it is the single most useful sentence in this whole subject for a homeowner: "Buildings and other structures designed per Chapter 32 of ASCE 7 do not meet the requirements for storm shelters or safe rooms." The fact sheet accompanying the provisions is blunter still — those buildings "address loads from common weak tornadoes" but "do not come close to meeting the more stringent life safety targets adopted for storm shelters."

Where shelters are required, it is narrow: the building code requires storm shelters in newly constructed schools and critical emergency operations facilities — 911 call centers, emergency operations centers, and fire, rescue, ambulance and police stations — located in the 250 mph zone. Outside that zone, and for every existing building, it is voluntary.

The gap this leaves
A hospital in the tornado-prone region is now designed for something in the neighborhood of an EF2. A residential safe room in the same county is designed for 250 mph and a 15-pound board at 100 mph. Those are not two points on one scale; they are two different design philosophies. The building code protects the building at a defensible cost. A shelter protects the occupants and accepts that the building around it may be gone. That is the entire reason a 60-square-foot box costs what a small car costs.

What should I actually ask an installer?

  1. "What design wind speed was this engineered to, and to which document?" You want a number and a citation: 250 mph to FEMA P-320/P-361, or a specific ICC 500 zone speed. "It's rated for an EF5" is not an answer.
  2. "Can I see the debris impact test report for this model?" Not a certificate, not a logo — the report, naming the model, showing the 15-pound missile at 100 mph into vertical surfaces and 67 mph into horizontal surfaces, and the pass criteria above.
  3. "Is the door in the report the door you are installing?" Including frame, hinges and latching hardware. Ask whether anything has been substituted since the test.
  4. "What is the anchorage detail for my slab and my soil?" A rated shelter is only rated as installed. You want a detail prepared for your site, and to know who inspects the anchors.
  5. "Is the site clear of flood and surge exclusions?" Wind rating is irrelevant if the unit is sited where FEMA criteria prohibit one — a floodway, velocity zone, Coastal A Zone, or Category 5 surge inundation area. See our guide on whether underground shelters flood.
Three phrases that should slow you down
"FEMA-approved" or "FEMA-certified." Neither exists. FEMA publishes criteria; it does not approve, certify or endorse shelters, products or contractors. "Tornado-proof" or "guaranteed to survive an EF5." The design target in every federal document is near-absolute protection, and the EF5 band has no upper bound. "Rated for 300 mph." There is no recognized residential tornado design wind speed above the 250 mph in FEMA's criteria and at the top of the ICC 500 map; a higher advertised number is marketing, not a standard, unless the seller can produce the engineering and the test report behind it.

The one-line version

250 mph, as a three-second gust, plus a 15-pound 2x4 at 100 mph into the walls and 67 mph into the roof, three times, without breaking through — and only if the anchorage and the door match the test report. The wind speed is the number that sells shelters. The debris test and the anchor detail are the numbers that decide whether the rating means anything on your property.

Sources

FEMA — Foundation and Anchoring Criteria for Safe Rooms Fact Sheet (November 2024; earlier revision March 2021). Primary source for the statement that all FEMA-funded residential safe rooms must be designed to resist wind loads and missile impacts for a tornado design wind speed of 250 mph regardless of location or storm type; for the requirement that the foundation resist uplift, overturning and sliding and transfer those forces into supporting soils; for compliance with all loads specified in ICC 500 Chapter 3; and for the caution that post-installed anchors depend on adhesive bonding or friction for pull-out resistance, making performance highly dependent on proper installation. FEMA — Safe Rooms: Selecting Design Criteria Recovery Advisory. Source for the side-by-side comparison of FEMA P-320/P-361 and ICC 500, including FEMA's single 250 mph design wind speed against the ICC 500 mapped approach with a minimum shelter design wind speed of 160 mph and a maximum of approximately 250 mph at a 10,000-year mean recurrence interval, and for the tornado missile criterion of a 15-pound 2x4 shared by both documents. FEMA — Highlights of ICC 500-2020, ICC/NSSA Standard for the Design and Construction of Storm Shelters (August 2021). Source for the statement that the tornado shelter design wind speed map contours are unchanged from previous editions; for the hurricane map being updated with data through the 2018 season at a 10,000-year mean recurrence interval; for test missile sizes being a function of storm type while speeds vary with storm type, design wind speed and orientation; for the 2020 change from fixed missile lengths with tolerances to ranges (10 to 15 feet for tornado missiles) to account for wood species variation; for the definition of a residential storm shelter as serving dwelling-unit occupants with a capacity not exceeding 16 people; and for the minimum mechanical ventilation rate of 5 cubic feet per minute per occupant. FEMA / Ready.gov — Highlights of ICC 500-2014. Source for the hurricane shelter missile criterion that the speed of the test missile impacting vertical shelter surfaces shall be a minimum of 0.50 times the shelter design wind speed, for the 10,000-year mean recurrence interval basis (0.5 percent probability of exceedance in 50 years), and for the residential versus community shelter definitions. Texas Tech University National Wind Institute — Debris Impact Test Facility, published test protocols (Appendix A). Primary source for the tornado protocol used for FEMA P-320/P-361 and ICC 500: a 15-pound wood 2x4 at 100 mph into horizontally travelling (wall) impacts and 67 mph into vertically travelling (roof) impacts; for the companion 9-pound missile protocols tied to fractions of design wind speed; and for the pass criteria — the missile may penetrate but may not perforate the safe side, permanent deflection must be less than 3 inches, a minimum of three impacts is required in vulnerable areas, no detached segment may extend 3 inches into the safe compartment, and door assemblies must maintain at least two locking points. Bennett, Coulbourne et al., "Structural Design and Coordination of ICC 500 Tornado Shelters," STRUCTURE magazine (NCSEA/CASE/SEI). Source for wind loads five to seven times higher than a similarly sized non-shelter building; the minimum roof live load of 100 pounds per square foot, up to five times a non-shelter roof; the internal pressure coefficient of plus or minus 0.55, or plus or minus 0.18 with atmospheric pressure change venting; the recommended impact factor of no less than 2.0 for collapse and laydown hazards; and the 15-pound 2x4 missile at 100 mph for vertical surfaces and 67 mph for horizontal surfaces. "Tornado Debris Impact Testing and Masonry," STRUCTURE magazine. Corroborating source for the 15-pound 2x4 at a minimum of 100 mph at wall impact, the requirement of three missile impacts, and the no-perforation-of-the-interior-surface pass criterion. FEMA and NIST — Design Guide for New Tornado Load Requirements in ASCE 7-22 (January 2023). Source for the tornado-prone region definition; the 1,700-year return period for Risk Category III and 3,000-year return period for Risk Category IV; design tornado speeds ranging from 60 to 138 mph as a function of risk category, effective plan area and geographic location; the correspondence to approximately EF0–EF2 intensity; and the explicit statement that buildings and other structures designed per Chapter 32 of ASCE 7 do not meet the requirements for storm shelters or safe rooms. ASCE 7-22 Tornado Loads Fact Sheet (as published in the Florida Building Commission code development resources). Source for the statement that while buildings designed in accordance with ASCE 7-22 Chapter 32 address loads from common weak tornadoes, they do not come close to meeting the more stringent life safety targets adopted for storm shelters, and for the EF0–EF2 coverage of the provisions. NOAA National Weather Service, Norman OK — The Enhanced Fujita Scale. Source for the EF0 through EF5 three-second gust bands (EF0 65–85, EF1 86–110, EF2 111–135, EF3 136–165, EF4 166–200, EF5 over 200 mph) and for the statement that the EF scale is a set of wind estimates, not measurements, based on damage, and that the three-second gust is not the same wind as in standard surface observations. NOAA National Severe Storms Laboratory — Severe Weather 101: Tornadoes FAQ. Source for the highest winds ever found in a tornado, 318 mph measured by mobile Doppler radar on May 3, 1999 near Bridge Creek and Moore, Oklahoma; for the caveat that the measurement was above ground level; and for NSSL's statement that it is not certain what the highest wind speed inside a tornado might be because strong and violent tornadoes destroy weather instruments. International Code Council, Building Safety Journal — storm shelter requirements and ICC 500 improvements. Source for the description of the 250 mph zone as covering part of 14 midwestern states from northern Texas to South Dakota, and for the code requirement that newly constructed schools and critical emergency operations facilities (911 call stations, emergency operation centers, and fire, rescue, ambulance and police stations) in the 250 mph zone include storm shelters. FEMA P-320 (April 2025), Taking Shelter from the Storm: Building or Installing a Safe Room for Your Home. The residential design and construction guidance a compliant one- or two-family safe room is built to; safe rooms in P-320 are designed to meet or exceed the criteria in FEMA P-361, and the stated protection objective throughout is near-absolute rather than absolute.