What the sources establish.
What remains unresolved.
An educational reconstruction from original plans, observed damage schematics and published calculations. Impact-zone circulation has the strongest detailed coverage; many other floors and construction details use reference geometry. This is not a complete 1:1 model of September 2001 conditions.
57 findings retain their source distinctions, verification dates and remaining limits. Reviewed geometry and repeated references can look equally detailed. Neither a successful test nor a close drawing fit establishes historical construction accuracy.
Download complete register (JSON) · Verification scope · Individual findings
Impact-floor cores and circulation
WTC 1 floors 92–100 and WTC 2 floors 77–85 have differentiated aperture profiles. Reviewed hoistways, stair openings and service areas replace prototype geometry only within their source masks. Tower B uses independent original drawings, including raised floor-80 secondary plates and depressed floor-81 machine-room plates. The North 80-span and South partial 158-span beam prototypes remain separate from this architectural review.
North elevators 6/7/50 clear faces and adjacent services, smaller South openings, floor-81 machine elevation and floor-82 transfer headroom remain unresolved. Typical plan-reading uncertainty is approximately 0.75–1 ft for the reviewed circulation geometry; each source registration retains its own uncertainty. A small control-fit residual is not a construction tolerance. Aperture polygon counts can include clipped fragments and are not counts of historical openings.
Floor-80 secondary plates use concrete top +1 ft 10 in; shuttle overhead limits use +1 ft 6 in. The bounded floor-81 machine-room plates use provisional concrete top −1 ft 3¾ in from the floor-specific plan. The typical section instead implies −1 ft 9 in. The floor-81 difference remains unresolved with 0.5 ft vertical uncertainty; plate thickness is provisional. Main-area plates do not include complete access, ramps, equipment or support construction. Floor-82 transfer headroom, doors and soffits remain unresolved.
Core profiles, openings and source readings · Elevator sources and terminal limits
Stairs: quantity and source coverage
The primary impact ranges contain 48 source-supported main-stair stories, 96 flights and 921 modeled risers: North departures 93–99 and South 77–85. The North Stair B direction and intermediate landing height are corrected. Across both towers, 54 of 644 displayed stair stories have source support; 590 use reference flights. The complete display contains 1,306 flights and 12,488 modeled risers.
These are model quantities, not a historical stair inventory. Shared footprints are corroborated across drawings rather than independently measured on every floor. Equal risers and some dimensions remain assumptions; rails, doors, landing thickness, wall assemblies, stringers and structural trimming remain incomplete. Post-impact damage and passability are not reconstructed.
Other floor bands still use reference envelopes. Fifteen South 75–76 overlaps with those envelopes remain recorded; they do not establish clashes in the historical building. Doorways, equipment, wall construction, local pits and complete clearances remain unresolved.
Framing, coordinates and retained steel
The lower model now uses all 47 numeric column workpoints from the LERA 97th–100th-floor worksheet for each tower. Separate source grids apply at 107, 108, 109 and 110. The South core beam trace was re-registered to its numeric workpoints while preserving its source topology. The lower grid is a prototype on floors 1–96 and 101–106. The exact transition to the changed upper workpoints, within-story splices, centroid offsets and fabricated beam-column joints remain unresolved. Registered drawing geometry is not an as-built survey.
Core aperture profiles cover WTC 1 floors 92–100 and WTC 2 floors 77–85. The renderer retains 12 tabulated core-column deflection profiles, 66 damage-table rows and 658 story-column section assignments. Of those assignments, 216 use representative wide-flange dimensions and 7 are partly dimensioned; explicit box exceptions remain. Architectural opening detail does not establish complete structural framing. Core beams repeat the floor-96 North and partial floor-77 South prototypes. Exact sections, joints, trimming around revised openings, reinforcement and core deck remain unresolved. The core slab retains a 4.5-inch/5-inch thickness conflict.
All 24 tower/case/column entries classified as severed in the selected NIST cases have reviewed retained geometry: 56 separate pieces. Source samples, section assignments and piece identities remain in the records. Touching or overlapping pieces in projection do not establish physical load continuity. The source elevation resolves one projected displacement component. Unmeasured transverse displacement, exact fracture surfaces and local plate tearing remain unresolved. Some visible ends are limited by obscuration rather than established fractures. These paths are calculated impact-case results, not an observational interior survey.
Thirty-three individually matched core-beam intervals across 13 floor/case profiles follow published bent plan paths. Other spans retain their footprint-based reconstruction. Vertical displacement is unresolved; traced paths stay at the reference beam height, with representative sections and connections.
Coordinate sources and prototype policy · Core-column source curves and retained pieces · Earlier framing evidence inventory
Whole-tower coverage and historical conditions
All 220 story heights agree with the original LERA floor_heights.txt transcription. Floor geometry uses cumulative story elevations with floor 1 as its local reference. Roof elevations and the top of the floor-110 schedule band are separate quantities; the model is not based on uniform story spacing.
Lobby, mechanical, hat-truss, roof and foundation areas include separate source-reviewed components, but coverage is uneven. Complete upper-floor architectural surfaces, sublevel framing and foundation assemblies remain partial. Special zones should not be inferred from ordinary-floor repetition.
Original design drawings and September 2001 conditions are different evidence states. Later tenant alterations, openings and closures have not been comprehensively reconciled. Detailed geometry, successful tests and drawing-fit residuals do not provide an overall accuracy percentage.
Story-height schedule · September 11 schedule and plan checks
Observed damage, calculations and animation
The tower sequences distinguish observed exterior damage, selected calculated interior states, published floor/wall results and reconstructed transitions. The animation stops at a chosen 2° tilt. It is an educational geometric illustration.
The sequence does not solve coupled structural or thermal response, calculate stresses or connection failures, calibrate elapsed time, or calculate the subsequent collapse. The chosen 2° endpoint is not a measured frame; holding lower floors in place is a display choice. Observed and calculated results must not be treated as interchangeable.
The base and more-severe case selector changes calculated interior damage. Both cases retain the observed exterior reconstruction; the separate calculated exterior figures are comparison evidence. A footprint intersection or missing modeled piece is not proof of an individual historical fracture.
What the checks establish
The 2026-09-11 integration run passed 422 geometry and state checks, including 90 main-view transitions, 81 chapter states, 36 standalone selections and 48 impact-stack selections. These checks use production geometry with renderer/platform stubs. They verify application consistency, not historical accuracy or GPU performance.
A bounded September 11 browser review rendered North floor 96, South floor 79 and the South onset endpoint without JavaScript or HTTP errors. The full-tower route timed out waiting for readiness in the first run. It initialized on an isolated retry, but screenshot capture timed out. Whole-tower and onset responsiveness still need a check on the intended release device; the software-browser results are not a hardware benchmark.
All 220 story heights match the LERA source transcription. Numerical checks, source registrations and model counts do not establish complete building geometry, historical clearance, connection survival or post-impact passability.
Earlier component reports below retain their original scopes and dates. Updating this register does not mean every source or browser view was independently checked again.
September 11 geometry and state checks · Schedule and plan checks · Bounded browser review · Whole-tower retry · Earlier component verification record
Additional component evidence and earlier checks
Retained and deformed steel
The original 119 FEMA panel identities and 134 retained regions are preserved in the source archive. The active reconstruction includes bounded later NIST revisions to selected panel contours. Hollow columns follow the retained contours, with spandrels at the inner face. Nominal 14-inch envelopes remain representative; M-2, M-27 and M-30 use their independently identified constituent plate thicknesses.
M-2 retains the explicitly representative 45-degree inward profile below floor 97. M-27 adds the source-documented inward fold above floor 95, limited to steel retained by the FEMA trace; its bend amplitude is representative. M-30 adds the two documented lower column fragments, a retained spandrel remnant and local folded plate lips. WTC 2’s north exit now distinguishes the six-foot severance at columns 253–254, column 254’s retained outer web, and the disconnected ends of columns 257–258. Exact local fragment lengths and bend magnitudes remain recorded assumptions.
Supplemental fragment evidence and dimensions · M-30 before collapse · Recovered M-2
Interior reconstruction
Fourteen floor and case profiles combine the NIST front and plan views with retained truss paths, shared deformation of chords and webs, fractured concrete and exposed corrugated deck. Ten hand-fitted truss runs and three beam paths recover visible steel omitted by rectangular summary openings. Connection assemblies show fixed building-side seats and displaced truss-side bearing pieces.
Truss reconstruction data · Concrete and deck reconstruction data · NIST truss front view · NIST slab damage
More detailed floor and column geometry
The shared interior reconstruction now covers 14 floor/case combinations: WTC 1 floors 94–96 and WTC 2 floors 79–82, each in both impact cases. The six added profiles come from NCSTAR1-2B’s detailed final plan and front views. Floor 79’s severe case retains three visible beam lanes inside a coarse removed area. They remain beam framing; no slab or unverified end connections are restored. WTC 2 floor 82’s base profile has no assigned vertical drop.
NCSTAR1-5G supplies 658 story-column section assignments for all 47 core columns on WTC 1 floors 92–99 and WTC 2 floors 78–83. The renderer uses 144 box profiles (including 2 explicitly extrapolated exceptions), 291 profiles with complete historical wide-flange dimensions, 7 partly dimensioned profiles and 216 representative wide-flange profiles. Box openings, plate thicknesses, world-axis orientation and story changes now appear consistently in as-built and damaged views. A separately documented C801 box at floor 77 borrows its own floor 78 dimensions.
Exact fabrication splices, rolled fillets, box lips and unequal-flange placement remain unresolved. C1008 and C508 transition conflicts are preserved. Of 1,120 sampled exterior schedule rows, 1,119 resolve to source section types. Exact tower/column/story assignments now apply to retained columns; one unresolved type and unlisted columns retain explicit fallbacks. Side plates, outer webs and inner webs use separate gauges. M-2, M-27 and M-30 retain identified inch gauges in both reference and damaged states. Section changes follow registered story/panel bounds, whose exact fabrication heights remain unresolved.
Core schedule and source cells · Dimensions and fallback policy · Exterior schedule and conflicts · Floor profiles and trace pixels · Four additional reports and checksums · Independent source readback
Panel comparison against NIST
NIST’s observed entry-wall schematics revise local contours in the earlier FEMA drawings. The current model applies 10 reviewed panel revisions across both towers: WTC 1 columns 116–117; WTC 2 panel groups 411–413, 414–416, 417–419, 423–425, 426–428, 429–431, 432–434 and 435–437. Bent retained fragments remain visible, and each changed panel records its source and bounded operation.
These are mixed FEMA/NIST observed contours. The base and more severe selector changes calculated interior damage; NIST’s calculated exterior figures are separate comparison evidence. Neither the observed diagrams nor the available calculations provide a complete measured three-dimensional damage survey. The gray region on NIST’s WTC 2 schematic remains unresolved, with a display-only uncertainty outline.
264 independent NIST retained/void witness rays pass through the revised geometry at two datums and in both thin-profile and hollow-member representations. M-27’s plate gauges pass 25 independent checks. Unchanged source contours are preserved. These checks establish the documented local corrections, not an exact match of every NIST schematic pixel.
Panel corrections and source coordinates · Geometry witnesses · Obscured-region preservation · M-27 section checks · WTC 1 observed source · WTC 2 observed source
EVID-001 · partly addressed with recorded limitKeep source evidence separate from model identity and display inference
Finding. Published observations, calculated damage footprints, synthetic model IDs, and intersection intervals are different kinds of information.
Integrated result. Exports and inspectors separate FEMA panel identities, NIST cases, model span IDs, and inferred intervals. No historical individual-truss confirmation is asserted.
Remaining limit. No complete set of historically identified damaged interior trusses or measured fracture surfaces has been established.
Verification date: 2026-09-07. Source audits plus integrated code review and CPU geometry/state regressions; no browser interaction certification.
NCSTAR1-2 · Fig. 7-12, 7-26, 7-42, 7-56, 7-72 · pp. 187-189, 209-211, 232-234, 254-256, 275-277, 288
EXT-001 · partly addressed with recorded limitFEMA contour fidelity and numerical precision
Finding. All 134 extracted regions triangulate to outer area minus window-hole area; the audit found no mismatched region at 1e-5 tolerance.
Integrated result. The archived FEMA baseline contains 134 regions verified by outer-minus-hole area and projection tests. The current renderer additionally applies panel-specific later NIST observations; see EXT-014. FEMA fidelity alone does not certify NIST agreement.
Remaining limit. Source center anchors retain drafting residuals equivalent to about 1.674 inches for WTC1 and 1.605 inches for WTC2. This is not a survey accuracy statement. Later NIST schematics revise some boundaries. Local corrections do not constitute a complete measured 3D damage survey.
Verification date: 2026-09-07. Source audits plus integrated code review and CPU geometry/state regressions; no browser interaction certification.
EXT-002 · partly addressed with recorded limitEstimated and insufficient-information regions
Finding. FEMA estimates WTC1 columns 111-115 at level 98 and WTC2 columns 413-418 at levels 81-82; it lacks sufficient detail for WTC2 columns 408-411 at levels 83-84.
Integrated result. Localized estimated and insufficient-information records added; yellow/violet dashed annotations mark source uncertainty. Annotation margins are display bounds only.
Remaining limit. A schematic placeholder or blank damage entry in an obscured region does not prove that steel was intact.
Verification date: 2026-09-07. Source audits plus integrated code review and CPU geometry/state regressions; no browser interaction certification.
EXT-003 · verified against integrated codeAggregate severed columns do not define fracture intervals
Finding. NIST's WTC1 observed aggregate lists columns 112-144 and 151 severed, 145-148 heavily damaged, 149-150 moderately damaged.
Integrated result. Aggregate NIST unique-column totals do not create multi-floor removal masks.
Remaining limit. The unique-column list does not assign the same state to every story or every panel along each column.
Verification date: 2026-09-07. Source audits plus integrated code review and CPU geometry/state regressions; no browser interaction certification.
NCSTAR1-2 · pp. 302-304 · FEMA403-CH2 · Fig. 2-16 · pp. 2-15, 2-18
EXT-004 · partly addressed with recorded limitWTC1 south ejected panel retains its fractional end stories
Finding. C329–331 are clipped only over registered levels93.5–96.5, preserving the lower half of story93 and upper half of story96. Spandrels94/95/96 are absent. Floor, stack and tower share these extents.
Integrated result. C329–331 are clipped only over registered levels93.5–96.5, preserving the lower half of story93 and upper half of story96. Spandrels94/95/96 are absent. Floor, stack and tower share these extents.
Remaining limit. NIST tentatively identifies this panel through smoke. Middle-of-story registration follows the written description; exact splice heights and three-dimensional fracture paths are unresolved.
Verification date: 2026-09-07. Independent source readback and current production geometry/state checks; exact test scope and source registrations are linked.
EXT-005 · partly addressed with recorded limitWTC 2 north exit damage is localized
Finding. NIST identifies six-foot severances of columns 253 and 254, a retained non-load-bearing outer web at 254, and failed connections at 257 and 258.
Integrated result. Columns 253 and 254 now show local six-foot severance; 254 retains its outer web and short side-plate lips. Columns 257 and 258 retain steel with floor 81 disconnections and slight outward bends.
Remaining limit. The six-foot length and column/floor relationships are sourced. Gap translation (±1 ft), web bow and connection-gap magnitudes remain representative. NCSTAR 1-3C prose identifies 254/255, conflicting with corroborated 253/254 labels in NCSTAR 1-2 and 1-5A. Columns 252, 255 and 259 remain intact; the possible bow at 256 is unresolved.
Verification date: 2026-09-07. Primary-source review and registered source overlays where applicable; production geometry/state tests; targeted live browser inspection. Final integration: 253 checks passed, with no failures.
EVID-002 · unresolved source limitInitial and refined NIST cases are distinct
Finding. NCSTAR1-6C distinguishes initial Ai/Bi/Ci/Di floor-model conditions from refined A/B/C/D conditions and states the refined cases were not used in that full-floor model.
Integrated result. Initial full-floor exterior matrices remain distinct from refined aircraft-impact cases.
Remaining limit. Equality of exterior damage between certain initial floor-model inputs does not establish universal equality across all global impact simulations.
Verification date: 2026-09-07. Source audits plus integrated code review and CPU geometry/state regressions; no browser interaction certification.
INT-001 · partly addressed with recorded limitOpening and severe-damage footprints have different display meanings
Finding. The maps distinguish openings relevant to ventilation from severe structural damage; both categories were excluded where appropriate in later structural analyses.
Integrated result. Original opening footprints remain the baseline. On eight detailed floor/case profiles, separate representative exposed-deck masks remove concrete while keeping corrugated steel deck; severe areas and two-foot exposed-deck shoulders show concrete fragments. All surfaces follow the shared deformation field.
Remaining limit. Masks are coarse source-informed placements, not exact strain contour traces. Fragment dimensions and field interpolation are representative. Detailed source-visible retained truss runs can override summary rectangular omissions.
Verification date: 2026-09-07. Source audits plus integrated code review and CPU geometry/state regressions; no browser interaction certification.
NCSTAR1-2 · Fig. 7-12, 7-26, 7-42, 7-56 · pp. 187-189, 209-211, 232-234, 254-256
INT-002 · unresolved source limitRecovered trusses do not prove located impact fractures
Finding. The interior auditor found that recovered trusses could not be assigned as-built locations and observed floor damage lacked a quantified interior depth.
Integrated result. Historical truss identity remains unresolved; a footprint intersection is not a recovered-fracture identification.
Remaining limit. Collapse and recovery handling prevent assigning particular recovered damage confidently to aircraft impact.
Verification date: 2026-09-07. Source audits plus integrated code review and CPU geometry/state regressions; no browser interaction certification.
NCSTAR1-3 · pp. 36, 93-94 · NCSTAR1-2 · Fig. 7-72 · pp. 275-277, 288
INT-003 · partly addressed with recorded limitC32T1 reference identity is not an arbitrary selected-floor pair
Finding. The NIST reference analysis identifies one truss at WTC1 floor 96, column 143 and a 713-inch analytical span.
Integrated result. Located pair details use the same reconstruction as floor/block/tower. Primary and bridging sections remain sourced references. Shared building-side seats, horizontal gussets, stiffeners, bearing angles and tie plates are rendered in member-local coordinates; bolts are one per individual truss.
Remaining limit. The app's centered synthetic pair uses different coordinates/span assumptions and lacks a verified historical assignment, especially when WTC2 is selected.
Verification date: 2026-09-07. Source audits plus integrated code review and CPU geometry/state regressions; no browser interaction certification.
INT-004 · partly addressed with recorded limitSource panel regions inform the shared floor inventory
Finding. Types7/8/9 use AppendixG panel-region coordinates. WTC1 short-span beam regions use the supported source orientation, and transverse bridging is clipped at those regions. Replacement beam geometry is centered on its contributing model lines. Existing six WTC2F79 retained beam IDs remain valid.
Integrated result. Types7/8/9 use AppendixG panel-region coordinates. WTC1 short-span beam regions use the supported source orientation, and transverse bridging is clipped at those regions. Replacement beam geometry is centered on its contributing model lines. Existing six WTC2F79 retained beam IDs remain valid.
Remaining limit. WTC2 cardinal placement and fabrication assignments remain unresolved. Adjacent primary members can belong to different panels: a mixed beam/truss pair is not itself an error. Existing WTC2 individual-line assignments are preserved within the0.5ft source-boundary uncertainty band.
Verification date: 2026-09-07. Independent source readback and current production geometry/state checks; exact test scope and source registrations are linked.
INT-005 · verified against integrated codePreserve WTC2 core orientation landmarks
Finding. The WTC2 source figure locates 508 northwest, 501 southwest, 1008 northeast, and 1001 southeast.
Integrated result. WTC2 core rotation preserves column508 NW and1001 SE, with tests.
Remaining limit. Rotating a WTC1 core without the correct numbering transform can place southeast damage in the wrong quadrant.
Verification date: 2026-09-07. Source audits plus integrated code review and CPU geometry/state regressions; no browser interaction certification.
GEO-001 · partly addressed with recorded limitStory heights, roof elevations and model datums remain distinct
Finding. All 220 story heights agree with the original LERA floor_heights.txt transcription. Floor geometry uses cumulative story elevations with floor 1 as its local reference. Roof elevations and the top of the floor-110 schedule band are separate quantities; the model is not based on uniform story spacing.
Integrated result. All 220 story heights agree with the original LERA floor_heights.txt transcription. Floor geometry uses cumulative story elevations with floor 1 as its local reference. Roof elevations and the top of the floor-110 schedule band are separate quantities; the model is not based on uniform story spacing.
Remaining limit. Matching the source transcription verifies the implemented schedule, not every historical slab or finish elevation. Raised secondary plates, depressed machine rooms and other local datums require their own sources. Absolute drawing datums must not be replaced by a single story-count calculation.
Verification date: 2026-09-11. Accuracy-note reconciliation against current production data, existing source/mesh review records and the September 11 release checks. Not a new independent measurement of every source drawing.
GEO-002 · unresolved source limitReference section dimensions differ from analytical work points
Finding. The primary reference depth is 29 in chord out-to-out;28.05 in is centroid separation and about 30 in is a composite analytical depth. Bridging reference depth is 24 in.
Integrated result. Primary/bridging depths retain29/24in reference dimensions; modeled gaps and reused specimens are explicitly distinguished from individual assignments.
Remaining limit. Uniform 7.5 in/5 in pair gaps are modeling choices, and the reference section is reused without a full member schedule.
Verification date: 2026-09-07. Source audits plus integrated code review and CPU geometry/state regressions; no browser interaction certification.
INT-006 · partly addressed with recorded limitCore structure and section schedules remain incomplete
Finding. Core aperture profiles cover WTC 1 floors 92–100 and WTC 2 floors 77–85. The renderer retains 12 tabulated core-column deflection profiles, 66 damage-table rows and 658 story-column section assignments. Of those assignments, 216 use representative wide-flange dimensions and 7 are partly dimensioned; explicit box exceptions remain. Architectural opening detail does not establish complete structural framing.
Integrated result. Core aperture profiles cover WTC 1 floors 92–100 and WTC 2 floors 77–85. The renderer retains 12 tabulated core-column deflection profiles, 66 damage-table rows and 658 story-column section assignments. Of those assignments, 216 use representative wide-flange dimensions and 7 are partly dimensioned; explicit box exceptions remain. Architectural opening detail does not establish complete structural framing.
Remaining limit. Core beams repeat the floor-96 North and partial floor-77 South prototypes. Exact sections, joints, trimming around revised openings, reinforcement and core deck remain unresolved. The core slab retains a 4.5-inch/5-inch thickness conflict.
Verification date: 2026-09-11. Accuracy-note reconciliation against current production data, existing source/mesh review records and the September 11 release checks. Not a new independent measurement of every source drawing.
NCSTAR1-2B_final_ch1-8 · Fig. 3-6, 3-18 · NCSTAR1-2B_final_ch9-11 · Fig. 9-34 · NCSTAR1-5G_final · PANYNJ-12673
INT-007 · verified against integrated codeInterval clipping is geometric inference; it does not create supports
Finding. Executed interval checks preserve original chord/web slope and portions outside openings, with seats considered only at original endpoints.
Integrated result. Original-coordinate clipping precedes smooth deformation. Chords are subdivided and webs follow the same field. Ten explicitly traced representative retained runs replace matched original pairs, recovering visible steel inside summary openings. Building-side seats remain at original supports; no seats are created at cuts.
Remaining limit. Tests do not prove original web pitch, all seat failures, individual pair-member damage, or a surviving load path.
Verification date: 2026-09-07. Source audits plus integrated code review and CPU geometry/state regressions; no browser interaction certification.
EXT-006 · partly addressed with recorded limitNominal depth must not disappear from bent source stems
Finding. FEMA supplies retained elevation contours; NIST describes nominal 14-inch box columns with the spandrel replacing the inner web. Those sources do not provide a complete deformed section survey.
Integrated result. Default geometry now uses neutral steel, contour-following hollow column walls, open ends and recessed inner-face spandrels. Bent stems retain depth away from their original centerlines. A separate thin-profile reference preserves FEMA projection. Original 134 retained regions and apertures remain authoritative.
Remaining limit. Nominal 14-inch depth and 3/8-inch display plates are representative. Source-edge band classification and concealed column continuation are inferred. Plate-specific tearing, exact wall thicknesses and complete depth displacement remain unresolved.
Verification date: 2026-09-07. Independent source/construction review and CPU source-projection, bent-stem, hollow-end, spandrel-depth and cross-view checks.
NCSTAR1-1 · pp. 10-11 · FEMA403-CH2 · Fig. 2-16 · pp. 2-18 · NCSTAR1-1 · pp. 11 · NCSTAR1-2A · pp. 38
EXT-007 · partly addressed with recorded limitSpandrel datum and nominal thickness must be consistent
Finding. Fig 2-9 places the spandrel top 13.5 in above floor; a 52 in spandrel therefore has its center 12.5 in below floor. The NIST subsystem uses 52 in by 3/8 in plates.
Integrated result. As-built spandrel centers use12.5in below their named floor, with52in depth. Canonical impact-block spandrels use3/8in nominal thickness. Source contours remain schematic; connection/context proxies are not exact sections.
Remaining limit. Tower offset and exaggerated thickness differ from floor/detail views. Fractional-floor warping can stretch a schematic plate near non-12-foot stories.
Verification date: 2026-09-07. Source audits plus integrated code review and CPU geometry/state regressions; no browser interaction certification.
NCSTAR1-1 · Fig. 2-9 · pp. 25 · NCSTAR1-6C · Fig. Table 6-1a note · pp. 160
EXT-008 · unresolved source limitPanel grouping does not prove a universal splice schedule
Finding. The ordinary-panel stagger in the damage diagrams reverses about the central panel; NIST describes two-story corner modules and special mechanical-floor treatment.
Integrated result. FEMA source stagger retained inside coverage. Unverified corner splice markers removed; remaining generic joint patterns explicitly representative.
Remaining limit. A single marching modulo-three schedule at whole floors is not the verified erection schedule. The trace's near 0.4 fractional splice ordinate is not a survey datum.
Verification date: 2026-09-07. Source audits plus integrated code review and CPU geometry/state regressions; no browser interaction certification.
FEMA403-CH2 · Fig. 2-16, 2-27 · pp. 2-18, 2-30 · NCSTAR1-6C · pp. 6, 157-158 · NCSTAR1-1 · Fig. 2-10 · pp. 26
APP-001 · verified against integrated codeAll located views must consume the same selected case
Finding. At the audit snapshot Assembly/Plan/Block used the selected case; Tower used base footprints and left core/slab damage unchanged.
Integrated result. Actual production scene-controller transitions and shared explorer geometry are checked with the interior datasets loaded. Both towers, all cases, all five main floor views, nine sequence chapters, initiation transforms and all four standalone views retain matching member and surface geometry. Both stack routes now include the complete impact range whenever an impact floor is selected.
Remaining limit. A shared control value alone does not guarantee shared geometry, inventory or effective historical state.
Verification date: 2026-09-07. Independent source/code audit, actual shared-geometry/controller regressions, and separate live-browser interaction/visual checks.
NCSTAR1-2 · Fig. Tables 7-1, 7-4, 7-6, 7-9 · pp. 183, 205, 228, 251
APP-002 · partly addressed with recorded limitInspection state stays explicit during the tower sequence
Finding. The audited transition reset lastStage to -1 while relying on previous-stage thresholds to clear damage.
Integrated result. Selected inspection case remains authoritative across chapters. A persistent status explicitly separates inspection state from historical chapter timestamps. Chapters do not silently overwrite the selection.
Remaining limit. The selected inspection case is independent of historical chapter timestamps; it is not a timed physical simulation.
Verification date: 2026-09-07. Source audits plus integrated code review and CPU geometry/state regressions; no browser interaction certification.
APP-003 · partly addressed with recorded limitDetail views require selected identity or an explicit reference-only state
Finding. The generic perimeter detail ignored damage while the unlocated generic truss sample could inherit a footprint; picking did not select a detail member.
Integrated result. Post-impact panel detail selects a FEMA identity; truss detail selects a located model pair. As-built generic panel is explicitly a construction specimen.
Remaining limit. FEMA panel IDs and located model pairs are traceable. Historical fabrication identities and precise interior section assignments remain incomplete.
Verification date: 2026-09-07. Source audits plus integrated code review and CPU geometry/state regressions; no browser interaction certification.
APP-004 · verified against integrated codeStandalone tower-floor route needs the same damage service
Finding. At audit time tower-floors.html loaded no damage state, zones or exterior traces and used a separate as-built renderer.
Integrated result. Standalone route now consumes the same data and canonical FloorAssembly/FloorBlock renderer through explorer-model.js.
Remaining limit. Comments claiming identical or exact logic do not make independent renderers consistent.
Verification date: 2026-09-07. Source audits plus integrated code review and CPU geometry/state regressions; no browser interaction certification.
APP-005 · verified against integrated codeInitiation clones must follow case-driven geometry replacement
Finding. InitiationView built its cloned graph and reference links once in the audited snapshot.
Integrated result. Initiation copies rebuild with case changes and preserve component visibility, materials, exact split-floor pivots and source-coordinate bowing. Member picking follows the rendered bow and tilt and rejects clipped-away halves.
Remaining limit. Updating an existing instance matrix is not equivalent to replacing source groups or geometries; clones can retain an old case.
Verification date: 2026-09-07. Independent source/code audit, actual shared-geometry/controller regressions, and separate live-browser interaction/visual checks.
APP-006 · verified against integrated codeLayers and slab transforms must preserve structural scope and openings
Finding. The audited perimeter toggle hid core columns in Assembly but not Block; X-ray omitted replacement panels; Facade overlaid an uncut box.
Integrated result. Heavy core and perimeter batches belong to their own layers in every view. Ordinary spandrels and FEMA profiles share the inner column face. X-ray restores each material’s original opacity and depth settings after toggles and case changes; facade envelopes cannot cover impact openings.
Remaining limit. Cladding damage cannot be inferred wholly from structural damage, and an intact reference envelope must not read as reconstructed surviving facade.
Verification date: 2026-09-07. Independent source/code audit, actual shared-geometry/controller regressions, and separate live-browser interaction/visual checks.
APP-007 · verified against integrated codeVisible evidence and exported records must match the rendered case and scope
Finding. The audited severe-mode carousel still appended a base map; picked records and exports differed in case detail and displayed scope.
Integrated result. Main and standalone source links follow tower and selected case. Inspector records clear when the selection changes. Standard framing records carry tower/case. Exported ledgers contain the visible floor or pair scope; selected-case source policies remain separately available.
Remaining limit. Counts and exports can identify intersected modeled spans, not whole historical trusses proven destroyed.
Verification date: 2026-09-07. Independent source/code audit, actual shared-geometry/controller regressions, and separate live-browser interaction/visual checks.
NCSTAR1-2 · Fig. 7-12, 7-26, 7-42, 7-56 · pp. 187-189, 209-211, 232-234, 254-256
APP-008 · partly addressed with recorded limitMissing datasets must not produce a silently incomplete damage view
Finding. The audited boot path did not require exterior traces for post-impact availability; direct impact navigation could select a disabled mode.
Integrated result. Required source datasets load before geometry; missing fetches disable reconstruction and state controls. Network error handling checked by code inspection; no browser fault-injection test performed.
Remaining limit. Interior-only geometry under a caption claiming FEMA exterior reproduction is incomplete evidence presentation.
Verification date: 2026-09-07. Source audits plus integrated code review and CPU geometry/state regressions; no browser interaction certification.
APP-009 · partly addressed with recorded limitRoute changes and reloads need an explicit selection contract
Finding. The audited routes stored tower/floor/view/case separately and navigation discarded selections.
Integrated result. Validated damage case persists through localStorage and storage events across routes. Both boot paths synchronize storage again after source loading. Tower, floor and view selections remain route-local; the case remains explicit when sequence timestamps change.
Remaining limit. Tower, floor and view selections remain local to each route; the damage case is shared.
Verification date: 2026-09-07. Independent source/code audit, actual shared-geometry/controller regressions, and separate live-browser interaction/visual checks.
GEO-003 · partly addressed with recorded limitReference dimensions and survey placement do not establish full fabrication geometry
Finding. The model retains the 207 ft 2 in reference-line plan dimension and source-qualified typical modules. Shared survey-derived tower centers now align the towers and sublevels; reviewed lobby column trees and special zones supplement the repeated upper framing.
Integrated result. The model retains the 207 ft 2 in reference-line plan dimension and source-qualified typical modules. Shared survey-derived tower centers now align the towers and sublevels; reviewed lobby column trees and special zones supplement the repeated upper framing.
Remaining limit. Complete section schedules, centroid offsets, fabrication joints and floor-by-floor construction remain incomplete. Nominal exterior box envelopes and ordinary panel dimensions are references, not universal exact members. Differing source panel-width descriptions remain recorded.
Verification date: 2026-09-11. Accuracy-note reconciliation against current production data, existing source/mesh review records and the September 11 release checks. Not a new independent measurement of every source drawing.
GEO-004 · unresolved source limitTruss-pair lattice, bridging stations, and span endpoints need separate provenance
Finding. Sources describe typical 6 ft 8 in. pair-center spacing and 13 ft 4 in. bridging pitch. NCSTAR 1-2A gives analytical long spans of 59 ft 8 in. and 58 ft 10 in. for different span conditions.
Integrated result. Exact alternate-column lattice is retained. Existing footprint rounding, span endpoints and bridging station assumptions remain recorded; no new historical measurements invented.
Remaining limit. Pair centers must be registered to the exterior-column lattice, not recentered arbitrarily. Bridging edge offsets are inferred; the representative model span of 60 ft 1 in. does not reproduce both analytical lengths or a complete as-built endpoint schedule.
Verification date: 2026-09-07. Source audits plus integrated code review and CPU geometry/state regressions; no browser interaction certification.
NCSTAR1-6C · pp. 3-4 · NCSTAR1-2A · Fig. 3-16 · pp. 48-51 · FEMA403-CH2 · pp. 2-3
EXT-009 · partly addressed with recorded limitM-2 inward bend: observed direction and illustrative amount
Finding. NIST correlates M-2, columns 129–131 at floors 96–99, with pre-collapse photographs. Lower portions bent inward with the floor-97 slab acting as a fulcrum. No bend displacement or angle is supplied.
Integrated result. Only FEMA panel WTC1-N-C129-131-F96.4-99.4 receives an inward depth warp below floor 97. The explicitly illustrative 1:1 slope is 45 degrees and yields at most 4.57536 ft inward at the retained lower ordinate 96.61872. This is not a NIST measurement or a conserved-length reconstruction. Source x/y and the thin reference remain unchanged. Every view and export uses the same rule.
Remaining limit. The exact hinge axis, angle, displaced surface and torn-plate geometry are unmeasured. Figure 3-5 has a conflicting 94–97 caption; text on p103 and Figure 2-17 identify the 96–99 panel.
Verification date: 2026-09-07. Independent primary-source review; numerical checks establish only the declared representation.
EXT-010 · partly addressed with recorded limitM-30 retained fragments supplement the FEMA trace
Finding. NIST identifies retained lower fragments of M-30 columns 132 and 133, with local outer-web and spandrel folds near floor 95. The nearly 90-degree fold is a local plate observation.
Integrated result. Two lower hollow stems, a retained spandrel remnant and local inward-folded lips supplement the unchanged FEMA source polygons. The retained columns use the independently identified type 123 plate schedule.
Remaining limit. Lower splice registration, torn outline, lip lengths and exact fold angle remain representative within recorded ranges. Column 134 and a complete three-story panel are not restored.
Verification date: 2026-09-07. Primary-source review and registered source overlays where applicable; production geometry/state tests; targeted live browser inspection. Final integration: 253 checks passed, with no failures.
NCSTAR1-3Cv1 · Fig. 2-14, 2-16, 3-2 · pp. 39, 40, 98, 99, 101, 102, 121
INT-008 · partly addressed with recorded limitRepresentative bent trusses and retained source runs
Finding. NIST NCSTAR1-2 supplies plan and front views of calculated truss damage for WTC1 floors95–96 and WTC2 floors80–81 in both cases.
Integrated result. Fourteen profiles contain36 localized controls,10 retained truss polylines and3 retained beam paths. The same paths drive floor, block and tower geometry; truss pair inspection uses its matched truss identity. Explicit retained paths are not warped twice.
Remaining limit. Controls, vertical drops and pair assignment are representative fits to published views, not recovered solver nodes.
Verification date: 2026-09-07. Independent source audit and actual geometry comparisons across views.
NCSTAR1-2 · pp. 184, 185, 186, 187, 207, 208, 209, 229, 230, 231, 252, 253, 254
INT-009 · partly addressed with recorded limitSeparate damaged concrete from retained steel deck
Finding. NIST distinguishes modeled concrete failure from metal-deck rupture in Figs7-11/25/41/55; red concrete-failure regions need not be through-floor holes.
Integrated result. 32 coarse exposed-deck masks across eight cases/floors reveal corrugated deck. Concrete fragments show local breakup; original openings and core voids remain empty. Surface picking and exports retain the source record.
Remaining limit. Mask boundaries are generally5ft-scale estimates; fragment geometry and two-foot fractured shoulders are illustrative.
Verification date: 2026-09-07. Independent source audit and actual geometry comparisons across views.
NCSTAR1-2 · pp. 184, 185, 186, 187, 207, 208, 209, 229, 230, 231, 252, 253, 254
INT-010 · partly addressed with recorded limitRepresentative connection assemblies and separation
Finding. NIST1-6C pp39–44 documents two bolts per ordinary truss pair, shared seats, vertical stiffeners/stand-offs, horizontal exterior gussets and distinct connection-failure paths.
Integrated result. Building-side fixtures are omitted where FEMA entry-face steel or the documented WTC1 ejected south panel is absent. Retained truss-side bearing fragments remain. Building fixtures on M-2 follow the same illustrative inward displacement as its steel. Separation is measured relative to retained wall support. The existing representative shared seats, two bolts per pair and bottom ties remain consistent across floor, detail and stack views.
Remaining limit. Representative separation does not assign a historical weld/bolt/walk-off failure mechanism; detailed seat classes are not assigned outside the documented floor96 quadrant.
Verification date: 2026-09-07. Independent source/code audit, actual shared-geometry/controller regressions, and separate live-browser interaction/visual checks.
APP-010 · verified against integrated codeReloads must not mix model and data revisions
Finding. The existing browser had older modules loaded; a refresh exposed an incompatible scene/floor module pair and prevented initialization.
Integrated result. Both routes use content-fingerprinted module graphs and entry scripts. Source JSON requests bypass old cached versions. Browser reloads initialize successfully and the old module export error is gone.
Remaining limit. An already-open document must reload to apply changed application code.
Verification date: 2026-09-07. Live-browser reload and route changes; complete static module graph validation.
INT-011 · partly addressed with recorded limitTwelve core-column centerlines use NIST calculated deflection amplitudes
Finding. Tables 7-1, 7-4, 7-6 and 7-9 publish twelve heavy-column lateral deflections across the two towers and two impact cases.
Integrated result. Twelve source-identified retained columns bend continuously in absolute elevation with tabulated maxima of 10–32 inches; severed segments remain omitted. All 66 table rows, including moderate and light states, are preserved in case-scoped exports.
Remaining limit. Peak elevations, detailed curves and complete 3D displacement vectors are not established by the tables. Moderate and light columns receive their recorded classifications without invented centerline offsets.
Verification date: 2026-09-07. Primary-source review and registered source overlays where applicable; production geometry/state tests; targeted live browser inspection. Final integration: 253 checks passed, with no failures.
INT-012 · partly addressed with recorded limitImpact core openings follow tower-specific plans within reviewed areas
Finding. WTC 1 floors 92–100 and WTC 2 floors 77–85 have differentiated aperture profiles. Reviewed hoistways, stair openings and service areas replace prototype geometry only within their source masks. Tower B uses independent original drawings, including raised floor-80 secondary plates and depressed floor-81 machine-room plates. The North 80-span and South partial 158-span beam prototypes remain separate from this architectural review.
Integrated result. WTC 1 floors 92–100 and WTC 2 floors 77–85 have differentiated aperture profiles. Reviewed hoistways, stair openings and service areas replace prototype geometry only within their source masks. Tower B uses independent original drawings, including raised floor-80 secondary plates and depressed floor-81 machine-room plates. The North 80-span and South partial 158-span beam prototypes remain separate from this architectural review.
Remaining limit. North elevators 6/7/50 clear faces and adjacent services, smaller South openings, floor-81 machine elevation and floor-82 transfer headroom remain unresolved. Typical plan-reading uncertainty is approximately 0.75–1 ft for the reviewed circulation geometry; each source registration retains its own uncertainty. A small control-fit residual is not a construction tolerance. Aperture polygon counts can include clipped fragments and are not counts of historical openings.
Verification date: 2026-09-11. Accuracy-note reconciliation against current production data, existing source/mesh review records and the September 11 release checks. Not a new independent measurement of every source drawing.
EXT-011 · partly addressed with recorded limitRecovered-panel column plate schedules
Finding. NIST NCSTAR1-3B Tables 4-1 and 3-5 identify M-2 as type122, M-30 as type123, and M-27 columns129/130/131 as types123/123/124 (inside-view ordering).
Integrated result. M-2 uses 3/8-inch sides and 1/4-inch webs. M-30 uses 7/16-inch sides and 1/4-inch webs. M-27 columns129/130 use 7/16-inch sides, column131 uses 1/2-inch sides, all with 1/4-inch webs. Spandrels are 3/8 inch. The 25 independent M27/M2 geometry checks pass.
Remaining limit. Nominal 14-inch box envelopes, concealed continuations, torn surfaces and exact local deformations remain representative. These schedules are not assigned to unrelated panels.
Verification date: 2026-09-07. Independent final report table readback, inside/outside orientation review and rays measuring the actual emitted plate gauges.
NCSTAR1-3B_final · Fig. 3-5(b) · Table 3-5, 4-1 · pp. 13–14, 23, 51
APP-011 · verified against integrated codeBowed geometry and lighting share the deformation field
Finding. The normals used for lighting must remain perpendicular to the displayed deformed surfaces.
Integrated result. Bowing now transforms normals with the inverse-transpose deformation Jacobian, in the same tower coordinates as vertices and CPU picking, including instanced members and tilted clones.
Remaining limit. This corrects geometric lighting. It does not create a structural simulation or add historical deformation evidence.
Verification date: 2026-09-07. Primary-source review and registered source overlays where applicable; production geometry/state tests; targeted live browser inspection. Final integration: 253 checks passed, with no failures.
APP-012 · verified against integrated codeInitiation display clones preserve shared geometry and evidence
Finding. Repeated tower initiation must retain the selected damage state without duplicating the full per-instance evidence graph or disposing shared source geometry.
Integrated result. The recursive evidence-copy allocation was removed. In the scoped diagnostic, immediate post-onset heap fell from 1735.39 MiB to 70.80 MiB. Live browser chapter and case switching completed without further crash or console errors.
Remaining limit. Shared evidence is treated as immutable. CPU measurements exclude GPU memory and do not establish performance on every device.
Verification date: 2026-09-07. Clone ownership, identity, disposal and case-restoration tests; full 253-check integration; nine live chapter transitions plus twelve bow/onset case combinations.
EXT-012 · partly addressed with recorded limitM-27 above-floor 95 inward bend
Finding. NCSTAR1-3C’s pre-collapse photo and text identify inward-bent columns above the floor 95 spandrel of M-27; the lower portion appears intact.
Integrated result. The retained FEMA M-27 geometry receives an inward depth fold only above the floor 95 hinge. Shared support queries use the same above/below-hinge rule. Source projection and removed upper material remain unchanged.
Remaining limit. The source supplies direction, not measured bend angle or hinge coordinates. Display slope 1 is representative; only the existing short retained lip bends. Recovered splaying and broken lower connections are not assigned to impact.
Verification date: 2026-09-07. Independent source transcription, renderer inspection, production geometry and browser verification.
INT-013 · partly addressed with recorded limitSix additional floor/case deformation profiles
Finding. The detailed final NCSTAR1-2B plates include WTC 1F94 and WTC 2F79/F82 in both impact cases.
Integrated result. The shared interior profile set now covers 14 floor/case combinations. Ten added plan/front assets are selected by floor and case in both application routes. WTC 2F82 base has no assigned vertical displacement.
Remaining limit. Topology and direction follow calculated raster figures; control amplitudes/radii and vertical interpolation remain representative. No additional slab openings or truss fracture identities are asserted.
Verification date: 2026-09-07. Independent source transcription, renderer inspection, production geometry and browser verification.
INT-014 · partly addressed with recorded limitFloor 79 retained framing must remain beam framing
Finding. Three connected lanes remain visible in WTC 2 severe floor 79 despite the coarse removed rectangle. The registered Type8 plan places these lanes in a beam substitution region.
Integrated result. Three explicit retained beam paths replace the six original primary-model IDs representing three beam lanes. They bypass the coarse member mask without restoring slab or adding supports, dampers or a second deformation.
Remaining limit. The south-side interpretation of FigG2 is registered; that figure has no north arrow. Lane registration residuals are below 0.5 ft. Vertical drops, individual sections, connection behavior and end fixity remain representative.
Verification date: 2026-09-07. Independent source transcription, renderer inspection, production geometry and browser verification.
NCSTAR1-2B_final_ch9-11 · Fig. 9-85 · pp. 305 · NCSTAR1-2A · Fig. G-2 · Table G-1 · pp. 192, 194
GEO-005 · partly addressed with recorded limitCore-column section schedule and historical dimensions
Finding. NCSTAR1-5G AppendixB tabulates47 core columns across WTC 1F92–99 and WTC 2F78–83: 658 story-column assignments. Historical AISC tables resolve 25 scheduled WF sizes; NIST sample drawings partly resolve 2 others.
Integrated result. Every view uses the same source-selected core section and world-axis orientation, partitioned at story boundaries. The 658 assignments select 144 box profiles, 291 complete historical WF profiles, 7 partly dimensioned WF profiles and 216 representative WF profiles. The 144 boxes include 2 explicit borrowed-dimension exceptions. C801F77 adds a separately identified box family with borrowed adjacent dimensions.
Remaining limit. NIST thermal-model sections are not a complete fabrication survey. Remaining WF dimensions, rolled fillets, unequal-flange mirror choice, box lips/corner fabrication, workpoint versus centroid and exact within-story splices are unresolved. Supplemental reinforcement dimensions are not substituted for parent-section gauges.
Verification date: 2026-09-07. Independent source transcription, renderer inspection, production geometry and browser verification.
NCSTAR1-5G_final · Table B-4, B-5, B-6, B-7, B-8 · pp. 62, 115, 284–293 · AISC-DG15-2002 · Table 2.2.1 · pp. 28 · NCSTAR1-3A_final · Fig. 2-6 · pp. 10
GEO-006 · partly addressed with recorded limitExplicit core section transition conflicts
Finding. NCSTAR1-2A impact-reference transitions disagree with thermal AppendixB for C1008F81–82 and C508F83. Recovered C801 pieces independently establish its box family atF77.
Integrated result. C1008F81–82 retains the impact-reference box family with its ownF80 box dimensions explicitly extrapolated; thermal14WF730 remains as an alternate. C508F83 retains its explicitly tabulated box with a transition conflict. C801F77 uses its ownF78 dimensions as a disclosed extrapolation.
Remaining limit. The evidence does not resolve exact splice heights or actual dimensions at the disputed elevations. Generic rolled sections are not substituted for known box families. Added flange reinforcement is documented but not placed without enough dimensions.
Verification date: 2026-09-07. Independent source transcription, renderer inspection, production geometry and browser verification.
NCSTAR1-2A · pp. 17 · NCSTAR1-5G_final · Table B-7, B-8 · pp. 290–293 · NCSTAR1-3B_final · Table 3-3 · pp. 10
EXT-013 · partly addressed with recorded limitApply exact sampled exterior schedules with recovered-panel precedence
Finding. The1119 resolved entries among1120 sampled thermal-model rows are eligible only for their exact tower/column/story; one type23 conflict and unlisted columns retain explicit fallbacks. Side, outer-web and inner-web gauges are applied separately. M2/M27/M30 nominal recovered gauges take priority in reference and damaged states.
Integrated result. The1119 resolved entries among1120 sampled thermal-model rows are eligible only for their exact tower/column/story; one type23 conflict and unlisted columns retain explicit fallbacks. Side, outer-web and inner-web gauges are applied separately. M2/M27/M30 nominal recovered gauges take priority in reference and damaged states.
Remaining limit. TableB3 t3 is the inner column web, not a spandrel assignment. Thermal dimensions are rounded meters. Actual fabrication splices and universal facade coverage remain unresolved; identified panel bounds follow the diagram registration.
Verification date: 2026-09-07. Independent source readback and current production geometry/state checks; exact test scope and source registrations are linked.
NCSTAR1-5G_final · Table B-3 · pp. Appendix B · NCSTAR1-3B_final · Table 3-5, 4-1 · pp. 14, 51
EXT-014 · partly addressed with recorded limitLater NIST observed contours revise the FEMA entry-wall baseline
Finding. NIST’s later observed schematics revise local retained stems and opening boundaries shown by FEMA. Gray regions mean damage cannot be determined; dashed cladding marks do not establish structural steel failure.
Integrated result. 10 panel identities now incorporate reviewed, bounded NIST contour revisions. Original FEMA polygons remain archived, each operation records its source pixels and scope, and all other panel contours remain unchanged. 264 independent NIST source-witness rays pass through the emitted geometry at two datums and two representations.
Remaining limit. This is a mixed FEMA/NIST schematic reconstruction, not the calculated finite-element exterior mesh or a complete measured 3D survey. Only reviewed local differences are changed. Unresolved gray regions and smaller endpoint differences remain recorded; line thickness and registration limit precision.
Verification date: 2026-09-07. Original report readback, registered source comparison and production geometry checks.
NCSTAR1-5A_final_ch1-8 · Fig. 6-6, 7-25 · pp. 65, 111 · NCSTAR1-2 · Fig. E-28(a), E-46(a), 7-5(a), 7-36(a) · pp. lxxviii, xciv, 181, 226
EVID-003 · partly addressed with recorded limitObserved exterior geometry and calculated exterior cases are distinct
Finding. E-28(a) and E-46(a) are observed schematics; parts(b) are base-case calculations. E-36 and E-54 are more-severe-case calculations. The report describes local calculated exterior failure differences between cases.
Integrated result. Both routes explicitly state that the rendered exterior follows observational evidence in both cases, while the selector changes calculated interior damage. Original observed and case-matched calculated exterior figures are available separately in the shared evidence controls and exports record the policy.
Remaining limit. The viewer does not reproduce case-specific calculated exterior meshes. NIST’s schematic observations and numerical predictions are not interchangeable historical records.
Verification date: 2026-09-07. Original report readback, registered source comparison and production geometry checks.
NCSTAR1-2 · Fig. E-28, E-36, E-46, E-54 · pp. lxxviii, lxxxv, xciv, c, ci
APP-013 · partly addressed with recorded limitIsolated panel inspection displays a cropped source extent
Finding. The shared exterior emitter clips source panels to WTC1 floors92.4–99.4 or WTC2 floors78–85.4, including in isolated inspection. Some identified assemblies extend beyond that extent.
Integrated result. The selected-panel explanation labels the displayed impact-band crop. The camera frames retained steel and excludes nonphysical uncertainty outlines from its bounds. A cropped edge does not imply a historical fracture or a complete retained assembly.
Remaining limit. The isolated view still shows the shared cropped geometry. Source schematics must be consulted for portions beyond the displayed extent.
Verification date: 2026-09-07. Original report readback, registered source comparison and production geometry checks.
GEO-007 · partly addressed with recorded limitNumeric core workpoints have explicit source-level limits
Finding. The lower model now uses all 47 numeric column workpoints from the LERA 97th–100th-floor worksheet for each tower. Separate source grids apply at 107, 108, 109 and 110. The South core beam trace was re-registered to its numeric workpoints while preserving its source topology.
Integrated result. The lower model now uses all 47 numeric column workpoints from the LERA 97th–100th-floor worksheet for each tower. Separate source grids apply at 107, 108, 109 and 110. The South core beam trace was re-registered to its numeric workpoints while preserving its source topology.
Remaining limit. The lower grid is a prototype on floors 1–96 and 101–106. The exact transition to the changed upper workpoints, within-story splices, centroid offsets and fabricated beam-column joints remain unresolved. Registered drawing geometry is not an as-built survey.
Verification date: 2026-09-11. Accuracy-note reconciliation against current production data, existing source/mesh review records and the September 11 release checks. Not a new independent measurement of every source drawing.
INT-015 · partly addressed with recorded limitRetained core-column pieces follow projected NIST calculations
Finding. All 24 tower/case/column entries classified as severed in the selected NIST cases have reviewed retained geometry: 56 separate pieces. Source samples, section assignments and piece identities remain in the records. Touching or overlapping pieces in projection do not establish physical load continuity.
Integrated result. All 24 tower/case/column entries classified as severed in the selected NIST cases have reviewed retained geometry: 56 separate pieces. Source samples, section assignments and piece identities remain in the records. Touching or overlapping pieces in projection do not establish physical load continuity.
Remaining limit. The source elevation resolves one projected displacement component. Unmeasured transverse displacement, exact fracture surfaces and local plate tearing remain unresolved. Some visible ends are limited by obscuration rather than established fractures. These paths are calculated impact-case results, not an observational interior survey.
Verification date: 2026-09-11. Accuracy-note reconciliation against current production data, existing source/mesh review records and the September 11 release checks. Not a new independent measurement of every source drawing.
NCSTAR1-2 · Fig. 7-6, 7-21, 7-37, 7-51 · pp. 182, 204, 227, 249
EXT-015 · partly addressed with recorded limitExit-wall support follows surviving inner steel
Finding. WTC2 north-wall support queries prioritize retained inner spandrels and50in splice plates, including panel seams, before local column fragments. The retained outer web ofC254 does not establish an inner-face seat. The possibleF82 tear betweenC251–252 is recorded as localized evidence without deleting the whole bay.
Integrated result. WTC2 north-wall support queries prioritize retained inner spandrels and50in splice plates, including panel seams, before local column fragments. The retained outer web ofC254 does not establish an inner-face seat. The possibleF82 tear betweenC251–252 is recorded as localized evidence without deleting the whole bay.
Remaining limit. No actual modeled truss anchor lies in theC251–259 exit region in the current inventory. The query fix does not prove a historical seat failure. TheF82 tear’s impact attribution and exact contour are uncertain.
Verification date: 2026-09-07. Independent source readback and current production geometry/state checks; exact test scope and source registrations are linked.
INT-016 · partly addressed with recorded limitStair quantities distinguish reviewed flights from repetition
Finding. The primary impact ranges contain 48 source-supported main-stair stories, 96 flights and 921 modeled risers: North departures 93–99 and South 77–85. The North Stair B direction and intermediate landing height are corrected. Across both towers, 54 of 644 displayed stair stories have source support; 590 use reference flights. The complete display contains 1,306 flights and 12,488 modeled risers.
Integrated result. The primary impact ranges contain 48 source-supported main-stair stories, 96 flights and 921 modeled risers: North departures 93–99 and South 77–85. The North Stair B direction and intermediate landing height are corrected. Across both towers, 54 of 644 displayed stair stories have source support; 590 use reference flights. The complete display contains 1,306 flights and 12,488 modeled risers.
Remaining limit. These are model quantities, not a historical stair inventory. Shared footprints are corroborated across drawings rather than independently measured on every floor. Equal risers and some dimensions remain assumptions; rails, doors, landing thickness, wall assemblies, stringers and structural trimming remain incomplete. Post-impact damage and passability are not reconstructed.
Verification date: 2026-09-11. Accuracy-note reconciliation against current production data, existing source/mesh review records and the September 11 release checks. Not a new independent measurement of every source drawing.
INT-017 · partly addressed with recorded limitElevator identities and source footprints do not certify clearance
Finding. All 99 physical elevator-car identities per tower remain represented. Source-reviewed impact-zone hoistways distinguish clear shaft walls from car outlines. South cars 12–23, including freight 17, have reviewed runway boundaries on 77–80; the model also distinguishes their termination and the separate machine/secondary levels.
Integrated result. All 99 physical elevator-car identities per tower remain represented. Source-reviewed impact-zone hoistways distinguish clear shaft walls from car outlines. South cars 12–23, including freight 17, have reviewed runway boundaries on 77–80; the model also distinguishes their termination and the separate machine/secondary levels.
Remaining limit. Other floor bands still use reference envelopes. Fifteen South 75–76 overlaps with those envelopes remain recorded; they do not establish clashes in the historical building. Doorways, equipment, wall construction, local pits and complete clearances remain unresolved.
Verification date: 2026-09-11. Accuracy-note reconciliation against current production data, existing source/mesh review records and the September 11 release checks. Not a new independent measurement of every source drawing.
INT-018 · partly addressed with recorded limitSouth terminal and transfer levels retain an explicit datum conflict
Finding. Floor-80 secondary plates use concrete top +1 ft 10 in; shuttle overhead limits use +1 ft 6 in. The bounded floor-81 machine-room plates use provisional concrete top −1 ft 3¾ in from the floor-specific plan. The typical section instead implies −1 ft 9 in.
Integrated result. Floor-80 secondary plates use concrete top +1 ft 10 in; shuttle overhead limits use +1 ft 6 in. The bounded floor-81 machine-room plates use provisional concrete top −1 ft 3¾ in from the floor-specific plan. The typical section instead implies −1 ft 9 in.
Remaining limit. The floor-81 difference remains unresolved with 0.5 ft vertical uncertainty; plate thickness is provisional. Main-area plates do not include complete access, ramps, equipment or support construction. Floor-82 transfer headroom, doors and soffits remain unresolved.
Verification date: 2026-09-11. Accuracy-note reconciliation against current production data, existing source/mesh review records and the September 11 release checks. Not a new independent measurement of every source drawing.
INT-019 · partly addressed with recorded limitOnset motion illustrates source states without solving collapse
Finding. The tower sequences distinguish observed exterior damage, selected calculated interior states, published floor/wall results and reconstructed transitions. The animation stops at a chosen 2° tilt. It is an educational geometric illustration.
Integrated result. The tower sequences distinguish observed exterior damage, selected calculated interior states, published floor/wall results and reconstructed transitions. The animation stops at a chosen 2° tilt. It is an educational geometric illustration.
Remaining limit. The sequence does not solve coupled structural or thermal response, calculate stresses or connection failures, calibrate elapsed time, or calculate the subsequent collapse. The chosen 2° endpoint is not a measured frame; holding lower floors in place is a display choice. Observed and calculated results must not be treated as interchangeable.
Verification date: 2026-09-11. Accuracy-note reconciliation against current production data, existing source/mesh review records and the September 11 release checks. Not a new independent measurement of every source drawing.