infrastructure · environment · family: the model was never trained on this
building codes that ignore the tornado's spin
Building codes do not account for tornado vortex aerodynamics
Problem statement
Current building codes and wind design standards (ASCE 7) treat tornado wind loads using the same straight-line wind framework used for hurricanes and thunderstorms, applying a tornado speed map with adjusted wind speeds. However, tornado winds are fundamentally different: they involve a translating vortex with intense vertical updrafts, rapid horizontal wind direction changes, and atmospheric pressure drops that create uplift forces on building envelopes. Wind tunnel testing — the foundation of current aerodynamic load coefficients — cannot replicate these vortex dynamics, meaning the pressure coefficients in codes may be significantly unconservative for tornado-prone regions. NSF's NEWRITE project at Iowa State is designing a facility to generate tornado, downburst, and gust-front wind fields at mid-to-full model scale (1:10 to 1:1) for structural testing; the 2023 award covers the design phase and a 1:20 prototype, not construction of the full-scale facility.
Why this matters
Tornadoes, downbursts, and gust fronts cause considerable property damage and numerous fatalities in the United States every year, and major tornadoes of EF3 or greater (136 mph or more, 3-second gust) have struck large population centers, causing multiple fatalities and significant damage to residential and engineered structures. Damage to infrastructure is expected to grow with urbanization and with the intensification and increasing frequency of such windstorms under a changing climate. Despite this known threat, ASCE 7-22 (the current standard) only added tornado provisions in 2022, and these are based on straight-line wind approximations rather than actual tornado vortex aerodynamics because the experimental data for vortex-structure interaction at meaningful scales simply does not exist.
What’s been tried and why it hasn’t worked
Conventional boundary-layer wind tunnels generate straight-line turbulent flows that represent hurricanes and thunderstorm outflows well but cannot produce the concentrated vortex, pressure deficit, or rapid wind direction change characteristic of tornadoes. Small-scale tornado simulators (Ward-type vortex chambers) exist but produce vortices too small to test realistic structural components — they can test small model buildings, but Reynolds number effects make the results difficult to transfer. Computational simulation of tornado-structure interaction using large eddy simulation (LES) is computationally expensive and has not been validated against full-scale structural data. Post-tornado damage surveys provide qualitative failure observations but cannot quantify the actual loads that caused failure, creating a circular problem: without load data, codes can't be calibrated; without calibrated codes, designs can't be validated.
What would unlock progress
Full-scale or near-full-scale experimental data on tornado-vortex-induced pressures and loads on structural components — exactly what NSF's NEWRITE design project at Iowa State (a $14.0 million Mid-scale Research Infrastructure-1 award made in September 2023, covering design and a 1:20 prototype but not construction) is meant to lead to: a facility generating tornado, downburst, and gust-front wind fields at 1:10 to full scale, large enough to study near-ground wind fields, wind loading, wind-borne debris impact, and structural damage mechanisms. Validated computational models calibrated against this experimental data could then extend the results to the full range of building types and tornado intensities. Probabilistic risk frameworks that combine tornado hazard models with vortex-specific structural fragility functions would enable risk-informed design.
Entry points for student teams
The reachable door is computational, and it does not require a team to generate its own validation data: NSF NHERI's DesignSafe Data Depot publicly lists tornado-simulator projects including "Assessment of Tornado Loading on Low-Rise Buildings Using Computational Modeling and Tornado Simulator Testing" (PRJ-4077 and PRJ-3718), "Benchmark Study of Tornado Wind Loading on Low-Rise Buildings with Consideration of Internal Pressure" (PRJ-2576), and "Implementation of Tornado Simulator Model in Open-Source Software — 'OpenFOAM'" (PRJ-4105), all browsable at https://www.designsafe-ci.org/data/browser/public/. Anchored to those published pressure measurements as the reference case, an LES study of how building shape, roof geometry, and opening locations change tornado-induced loads relative to straight-line wind loads is a semester-scale project needing compute and nothing else — no facility, partner, license, or restricted data. Building a translating Ward-type vortex chamber is not a comparable starting point and should not be presented as one: the translating simulators at Iowa State and Texas Tech were multi-year capital programs, and the NEWRITE design-phase award alone was $14.0 million. A hardware-inclined team can still do real experimental work by scaling the ambition down to a fixed, non-translating vortex chamber with a single pressure-tapped building geometry, which isolates the pressure-deficit and rapid-direction-change effects — the two mechanisms straight-line wind tunnels miss — without the translation mechanism that drives most of the cost. Relevant disciplines include structural engineering, fluid mechanics, atmospheric science, and computational engineering.
Genes
Structural cousins: stuck the same way, in other fields
Sources
NSF CMMI Natural Hazards Engineering Research Infrastructure (NHERI), including NEWRITE facility (Iowa State University); NSF DCL on NHERI FY 2026-2035 (NSF 24-117); accessed 2026-02-15 go to source ↗
Working notes
Notes made while writing and checking this brief, shown as written. They use working shorthand.
NSF's NHERI program and the Dear Colleague Letter (NSF 24-117) describe a planned competition for "experimental and field equipment/instrumentation facilities" that "will be expected to advance frontier science and engineering research focused on the impact of climate change, earthquake, tsunami, windstorm, storm surge, flooding, and fire/wildland-urban interface (WUI) hazards on the nation's civil infrastructure." Iowa State's NEWRITE design award (NSF award 2330150, made September 2023) is titled for "Enhancing Wind Resiliency of Infrastructure in Tornado-Downburst-Gust Front Events." The DCL notes that NHERI awardees organized Natural Hazards Summits in 2022 and 2024. Related problem: wildfire-wui-fire-codes-unproven.md addresses another case where building codes lack empirical validation for a specific hazard type.
Reconciliation 2026-08-21: Entry-point realism repair (panel C37 round; the repairing agent's session terminated on an API limit before it recorded this note, which the coordinating session appended from the landed edit). The section led with "build a small-scale tornado simulator (Ward-type vortex chamber with translating capability)" — a lab-facility capital program, not a semester project: the translating simulators at Iowa State and Texas Tech were multi-year builds and NEWRITE itself was $14M. Per the 2026-08-21 entry-point rules, the computational door now leads and is anchored to published tornado-simulator pressure measurements the team does not have to generate (NSF NHERI DesignSafe Data Depot public projects, browsable at https://www.designsafe-ci.org/data/browser/public/), making it facility-free; the hardware ambition is retained honestly as a scaled-down fixed, non-translating vortex chamber with a single pressure-tapped geometry, which isolates the pressure-deficit and rapid-direction-change mechanisms straight-line tunnels miss without the translation mechanism that drives the cost. The specific DesignSafe project identifiers named in the section were cited by the repairing agent from its own fetches; they were not independently re-verified by the coordinating session and should be spot-checked at next touch.
Reconciliation 2026-09-30: Whole write-up read against its sources. The Dear Colleague Letter NSF 24-117 was fetched in full on 2026-09-30 from https://www.nsf.gov/funding/opportunities/dcl-nsf-support-natural-hazards-engineering-research/nsf24-117. The NEWRITE facility named on the Source line has no link there; its record was read in full from the NSF award database (award 2330150, "Mid-scale RI-1 (M1:DP): National Testing Facility for Enhancing Wind Resiliency of Infrastructure in Tornado-Downburst-Gust Front Events (NEWRITE)," Iowa State University, PI Partha Sarkar, fetched 2026-09-30 from https://api.nsf.gov/services/v1/awards.json?keyword=NEWRITE). The DesignSafe Data Depot browser page is a login-aware web app that returns no project text to a script, so the four project records were read in full from DesignSafe's public publications API (https://www.designsafe-ci.org/api/publications/v2/PRJ-4077/ and the same address for PRJ-3718, PRJ-2576 and PRJ-4105, fetched 2026-09-30). (1) The Problem Statement said no facility to generate full-scale tornado-like vortices existed "until the recent establishment of NSF's NEWRITE." The award record says the $14,025,452 award "will support the design phase" of NEWRITE, that a 1:20 prototype will be built during the design project, and that the award "is not a commitment to support the construction phase of the full-scale NEWRITE"; the sentence now says NEWRITE is being designed and that the award covers design and a 1:20 prototype, not construction. (2) The same paragraph gave "wind direction changes of 360 degrees over seconds" and "pressure drops of 50-100 mbar"; neither figure is in any source, and the sentence now says rapid direction changes and pressure drops without the numbers. (3) Why This Matters gave $10-30 billion a year in damage, $46 billion in 2011, 1,200 tornadoes a year, over 80% of deaths in buildings, and an eastward shift of Tornado Alley. None of these is in any source. The award record says these events "cause considerable property damage and numerous fatalities annually in the United States," that EF3 or greater tornadoes (136 mph or greater, 3-second gust) "have struck large population centers, causing multiple fatalities and significant damage to residential and engineered structures," and that damage "will only increase with growing urbanization and increased intensification and frequency of such windstorms due to the changing climate"; the paragraph now says that and no more (the growth sentence is about damage to infrastructure, as in the record). (4) What's Been Tried said Ward-type simulators have existed "since the 1970s" and test buildings "at 1:100 scale"; no source gives either figure, and the sentence now says only that they exist and test small models, with results difficult to transfer. (5) What Would Unlock said the facility is "new," "$14M, 2024," and has "its ability to generate" the wind fields. The award was made September 22, 2023 (start date October 1, 2023), is for design, and describes a facility designed for 1:10 to 1:1 scale to study near-ground wind fields, wind loading, wind-borne debris impact and structural damage mechanisms; the sentence now says that. (6) The entry points said "NEWRITE itself cost $14M"; the $14.0 million is the design-phase award, and the sentence now says so. (7) In the notes above, the DCL quotation spliced two sentences into one and dropped "(WUI)" and "on the nation's civil infrastructure"; it is now two fragments copied from the DCL. (8) The notes said NEWRITE was "funded 2024" and that "the 2024 Natural Hazards Research Summit identified tornado engineering as a critical gap." The award date is September 2023, and the DCL says only that NHERI awardees organized Natural Hazards Summits in 2022 and 2024, nothing about what they found; both sentences now say what the records say. Confirmed in the sources and left as written: the Source line (NSF 24-117, Directorate for Engineering, Division of Civil, Mechanical and Manufacturing Innovation; NHERI FY 2026-FY 2035; publication date August 13, 2024); NEWRITE at Iowa State University with Texas Tech among the participating universities; the four DesignSafe project identifiers and their titles, all published under Open Data Commons Attribution or GNU General Public licenses with tornado-simulator pressure data in the file listings; and the award record's statement that data from the design project will be made publicly available in the NHERI Data Depot. Not in any named source and left as written for Beth's judgment, because they are the framing of the problem rather than figures: that ASCE 7 treats tornado loads with a tornado speed map inside the straight-line wind framework and that ASCE 7-22 first added tornado provisions; that wind-tunnel pressure coefficients may be unconservative for tornadoes; that large eddy simulation of tornado-structure interaction has not been validated against full-scale structural data; and that the translating simulators at Iowa State and Texas Tech were multi-year capital programs. A second reader went over these repairs against the same sources on the same day; all eight stand, with the wording of (3) and (4) tightened to what the records say. No title changed. Genome tags untouched.