The FZ1073 upset reconstructed from open data

On 30 September 2026 a Boeing 737 MAX 8 (flydubai FZ1073, A6-FKF) went into a steep spiral dive over northern Saudi Arabia after a struggle on the flight deck, and landed at Tabuk with its rudder missing. Public reports describe a loss of 14,000 ft in under 30 s and a peak of Mach 0.96. We reconstruct the upset from two public Automatic Dependent Surveillance-Broadcast (ADS-B) feeds and the ERA5 reanalysis, and check every derived quantity against the airspeed, Mach number and wind that the aircraft transmitted through Mode S Enhanced Surveillance before and after the dive: over 164 points the reconstructed Mach number has a spread of 0.006 against the transmitted one, close to its resolution, and a systematic difference below 0.003. Each altitude segment is converted into the Mach number needed to fly it; jumps that would need more than Mach 1 or 3 g are rejected. What remains supports the 14,000 ft in 29 s, at a mean Mach number of 0.86, with a steep phase of about 200 m/s that needs Mach 0.91 on average and 0.95 over its first 10 s, beyond the limit of the ADS-B vertical-rate field. At the bottom of the pull-out the aircraft flew at Mach 0.88 (0.85 to 0.91) and 438 kt calibrated (422 to 454), against a maximum operating speed of 340 kt: the design dive Mach number, at least 0.87 by regulation, was likely reached in the steep phase, and the design dive speed at the bottom. A point-mass model, calibrated on the phugoid that followed and constrained by the surrounding data, fills the gap with 288 compatible histories: a bank near 90 degrees or a push to low load factor, a flight-path angle near -46 degrees, a peak near Mach 0.95 (0.91 to 1.00), and a pull-up of at least 2.7 g with the wings still banked. These are predictions, fixed before recorder data. A drift model places the rudder within about a kilometre of 29.82 N, 38.33 E. Changes in v4: a new section reconstructs the control histories compatible with the data inside the 30 s where the record is sparse (point-mass model with the OpenAP polar of the 737 MAX 8, calibrated on the observed phugoid; population sampling under twelve constraints; two seeds and two sensitivity runs), with a table of predictions to be tested against recorder data; the debris paragraph treats the rudder as the whole surface that the photographs show missing, under two release hypotheses; the viewer flags follow the consistency chain. Earlier versions: v3 corrected three statements, v2 restated the validation statistics. This record contains the preprint, the analysis and simulation scripts, derived data, the accepted histories and an interactive viewer. Raw ADS-B files are not redistributed; the scripts retrieve them from the original providers. Code: MIT licence. Preprint, derived data and figures: CC BY 4.0.

Authors

Institutions

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-10-14
DOI
https://doi.org/10.5281/zenodo.23357086
Primary Topic
Aerospace and Aviation Technology
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
preprint

The FZ1073 upset reconstructed from open data

Federico Toson
Zenodo (CERN European Organization for Nuclear Research)
Aerospace and Aviation Technology
preprint

The FZ1073 upset reconstructed from open data

Federico Toson
preprint en

Abstract

On 30 September 2026 a Boeing 737 MAX 8 (flydubai FZ1073, A6-FKF) went into a steep spiral dive over northern Saudi Arabia after a struggle on the flight deck, and landed at Tabuk with its rudder missing. Public reports describe a loss of 14,000 ft in under 30 s and a peak of Mach 0.96. We reconstruct the upset from two public Automatic Dependent Surveillance-Broadcast (ADS-B) feeds and the ERA5 reanalysis, and check every derived quantity against the airspeed, Mach number and wind that the aircraft transmitted through Mode S Enhanced Surveillance before and after the dive: over 164 points the reconstructed Mach number has a spread of 0.006 against the transmitted one, close to its resolution, and a systematic difference below 0.003. Each altitude segment is converted into the Mach number needed to fly it; jumps that would need more than Mach 1 or 3 g are rejected. What remains supports the 14,000 ft in 29 s, at a mean Mach number of 0.86, with a steep phase of about 200 m/s that needs Mach 0.91 on average and 0.95 over its first 10 s, beyond the limit of the ADS-B vertical-rate field. At the bottom of the pull-out the aircraft flew at Mach 0.88 (0.85 to 0.91) and 438 kt calibrated (422 to 454), against a maximum operating speed of 340 kt: the design dive Mach number, at least 0.87 by regulation, was likely reached in the steep phase, and the design dive speed at the bottom. A point-mass model, calibrated on the phugoid that followed and constrained by the surrounding data, fills the gap with 288 compatible histories: a bank near 90 degrees or a push to low load factor, a flight-path angle near -46 degrees, a peak near Mach 0.95 (0.91 to 1.00), and a pull-up of at least 2.7 g with the wings still banked. These are predictions, fixed before recorder data. A drift model places the rudder within about a kilometre of 29.82 N, 38.33 E. Changes in v4: a new section reconstructs the control histories compatible with the data inside the 30 s where the record is sparse (point-mass model with the OpenAP polar of the 737 MAX 8, calibrated on the observed phugoid; population sampling under twelve constraints; two seeds and two sensitivity runs), with a table of predictions to be tested against recorder data; the debris paragraph treats the rudder as the whole surface that the photographs show missing, under two release hypotheses; the viewer flags follow the consistency chain. Earlier versions: v3 corrected three statements, v2 restated the validation statistics. This record contains the preprint, the analysis and simulation scripts, derived data, the accepted histories and an interactive viewer. Raw ADS-B files are not redistributed; the scripts retrieve them from the original providers. Code: MIT licence. Preprint, derived data and figures: CC BY 4.0.

Zenodo (CERN European Organization for Nuclear Research)
University of Padua (IT)
Aerospace and Aviation Technology
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.