N722BD: back-of-the-envelope synchronized ATC / ADS-B timeline

Created
· Expires 2027-02-01 20:48 · 261103-rf

N722BD: back-of-the-envelope synchronized ATC / ADS-B timeline

Working reconstruction — November 1, 2026 (UTC)
Prepared November 2, 2026

STATUS: PROVISIONAL. This is not an official accident reconstruction. The radio transcript is partly automated and partly corrected by a listener. The UTC synchronization has an unverified offset. No cockpit voice recorder or flight data recorder information was used.

Shared in the hope that it might be found useful as a rough outline. Not a forensic analysis.

Time alignment

Assume the LiveATC archive KPBI1-ZMA01-Nov-01-2026-1730Z.mp3 begins at exactly 17:30:00 UTC. Add each VTT cue's elapsed time to 17:30:00 to obtain the tentative UTC times below. The archive filename alone does not prove that start time.

Independent check: in the same recording, ATC tells AA672 to descend to FL320 at VTT 08:01.68 (tentatively 17:38:01.68 UTC). The separately downloaded AA672 FR24 track shows it near FL360 at 17:38:12, descending by 17:38:44, and near FL320 by 17:42:26. This is consistent with the proposed offset, but the sparse AA672 data cannot validate it to the second.

Event sequence

For each event, altitude and vertical rate are taken from the latest available N722BD FR24 CSV observation at or before the stated UTC time, with separate fields carried forward independently. They are not cockpit instrument readings. Positive rate means climbing; negative means descending.

  • 17:48:02.220 (VTT 18:02.220): Crew requests climb to FL370. ADS-B: 35,100 ft; +1,536 ft/min.
  • 17:48:08.300 (18:08.300): ATC clears climb to FL370. ADS-B: 35,275 ft; +1,472 ft/min.
  • 17:48:12.780 (18:12.780): Crew reads back climb clearance. ADS-B: 35,350 ft; +1,088 ft/min.
  • 17:48:28.900 (18:28.900): Crew reports descending 8,000 ft/min. ADS-B: 34,525 ft; −10,752 ft/min.
  • 17:48:30.260 (18:30.260): Mayday; crew reports descending 10,000 ft/min. ADS-B: 34,375 ft; −12,544 ft/min.
  • 17:48:39.300 (18:39.300): Crew reports an overspeed condition. ADS-B: 31,500 ft; −20,864 ft/min.
  • 17:48:41.000 (18:41.000): Crew reports descending 20,000 ft/min. ADS-B: 30,975 ft; −20,928 ft/min.
  • 17:48:54.480 (18:54.480): “We're gonna crash into the ground.” ADS-B: 26,375 ft; −23,360 ft/min.
  • 17:49:01.600 (19:01.600): Crew reports descending 22,000 ft/min. ADS-B: 23,350 ft; −30,464 ft/min.
  • 17:49:05.000 (19:05.000): Crew says “we are banking ... sixty degrees” (the word “banking” is listener-confirmed). ADS-B: 21,675 ft; −32,640 ft/min. This is a reported bank angle, not a roll-rate measurement.
  • 17:49:17.180 (19:17.180): Tentative listener transcription: “Jack, look outside, Jack! Look outside!” ADS-B: 13,300 ft; −32,640 ft/min.
  • 17:49:20.700 (19:20.700): Final, truncated exclamation, tentatively “Holy fuck, we're outta tə—”. The completion “time” is not established. ADS-B: 10,550 ft; −32,640 ft/min.
  • ~17:49:34.4 UTC (estimated, NOT observed): Ground impact. Extrapolation of the final changing ADS-B barometric-altitude readings (17:49:12–17:49:22.955 UTC) gives a fitted downward rate of approximately 45,700 ft/min (761 ft/s) and a ground-level crossing near 17:49:34.4 UTC. This assumes the final measured trend continues without a significant change in vertical acceleration or trajectory. It is not a confirmed impact timestamp. The last changing altitude observation is 8,750 ft at 17:49:22.955 UTC; subsequent altitude reports remain stuck at 8,750 ft through 17:49:24.592 UTC, while vertical-rate reports continue until 17:49:32.166 UTC. Do not treat the frozen altitude as continued level flight or as a measured altitude at 17:49:32.

These times mark the start of the relevant VTT phrase or cue, not necessarily the instant a complete spoken value was uttered. Word-level wording comes from a listener's markup, which is not published here.

What the alignment suggests

The aircraft climbed briefly to approximately 35,450 ft near 17:48:15 UTC, then entered a rapidly accelerating descent. Around 17:48:41, the crew's spoken 20,000 ft/min descent rate is close to ADS-B's −20,928 ft/min. By the later spoken 22,000 ft/min report, ADS-B shows roughly −30,000 ft/min. The difference might involve instrument filtering, verbal approximation, lag, or imperfect audio alignment; it should not be silently reconciled.

The repeated −32,640 ft/min ADS-B values represent the 1090ES ADS-B vertical-rate overflow code, not a physically constant descent rate. The nine-bit vertical-rate field uses 64 ft/min increments; its highest code denotes a magnitude greater than 32,608 ft/min. The displayed −32,640 ft/min is therefore a saturated/overflow indication, not a measured exact rate. Similarly, the spoken 60° bank does not establish whether the aircraft was rolling or holding a fixed bank; ground track is not attitude.

The final radio exclamation maps to approximately 10,550 ft reported altitude, conditional on the assumed recording start. The independent extrapolation from the final changing altitude observations yields ~17:49:34 UTC as a tentative impact estimate, about 13.7 seconds after the final exclamation began. Barometric altitude is not exact height above ground, and the extrapolation does not prove the actual impact time.

Methodology and limitations

  1. Audio clock: Use VTT offsets from the named 17:30Z LiveATC recording; tentatively add them to 17:30:00 UTC. The VTT runs beyond 30 minutes, so its absolute clock requires further verification.
  2. Transcript: Prefer the listener's ear-checked wording to automated recognition, while preserving ambiguous words and speaker identities as uncertain. All ATC-side transmissions heard in the recording are not necessarily present in this selected event list.
  3. ADS-B: Sort the N722BD granular CSV by UTC timestamp. For each event, take the latest non-null altitude and vertical-rate observations at or before that instant (a zero-order hold, not interpolation). ADS-B fields update asynchronously, so their individual measurement ages may differ.
  4. Independent timing check: Compare the AA672 descent clearance with AA672's separately obtained FR24 track. The match supports a near-17:30 start but does not establish a precise offset; AA2122's handoff is less useful because it need not produce a visible maneuver.
  5. Impact extrapolation: Fit a straight line to changing barometric-altitude observations from approximately 17:49:12 to 17:49:22.955 UTC, excluding the repeated 8,750-ft readings after 17:49:22.955. Extrapolate to approximately zero feet barometric altitude as a ground-level proxy. This gives ~17:49:34.4 UTC, but assumes constant final vertical velocity and neglects local pressure/ground elevation corrections, possible trajectory changes, and ADS-B latency. The repeated −32,640 ft/min values are a protocol overflow indicator, not the true vertical speed.
  6. Interpretation: Altitude is in feet; rates in feet/minute. Spoken vertical rates, ADS-B rates, bank angle, and apparent attitude in crash video are different observations and should not be treated as interchangeable.

Sources used

  • KPBI1-ZMA01-Nov-01-2026-1730Z.mp3 and .vtt — LiveATC recording, relative audio timestamps and machine transcript.
  • A listener's corrected, word-timed transcript, not published — key wording and unresolved final phrase.
  • Flightradar24_N722BD_Granular_Data.csv — N722BD ADS-B altitude and vertical-rate observations.
  • AA672_42635cba.csv and AA672-42635cba.kml — independent descent timing check.
  • AA2122_42633503.csv and AA2122-42633503.kml — second aircraft, but no useful maneuver yet identified.

Companion documents: N722BD: back-of-the-envelope speed estimate from dashcam video and N722BD: back-of-the-envelope check of the radar data.

The next verification step would be another independent, unmistakable altitude or heading clearance matched to an aircraft's ADS-B maneuver, or a reliable absolute-time marker from the original LiveATC stream.

Additional technical references