When the Evidence Disappears: Inside the Yvette Walsh–SAA Dispute

By Garth Calitz

An aviation safety investigation, a pilot’s career on the line, and an uncomfortable question: how do you challenge a finding when the evidence you want to examine is no longer available?

South African aviation has no shortage of regulations, procedures, committees and people whose job descriptions suggest that every conceivable eventuality has already been anticipated. Yet when something goes wrong, the system occasionally produces a result that leaves everyone asking a rather inconvenient question: who was watching the process?

The case involving South African Airways (SAA) Captain Yvette Walsh is one such example. What began as a serious turbulence encounter aboard an Airbus A330-300 in October 2024 has become a dispute over aviation safety findings, cockpit decision-making, the preservation of flight data, and the fairness of the process that followed.

Walsh, an experienced airline pilot with a reported 23 years of flying experience, is challenging the investigation and the effect of its findings on her professional future. She seeks an independent investigation and argues that the loss of important recording evidence has limited her ability to challenge the conclusions.

SAA opposes her application, arguing that another investigation is unnecessary and disputing the urgency of the relief sought. The High Court has removed the matter from the urgent roll, with the dispute expected to proceed through the ordinary court process.

None of this establishes that Walsh is right, or that SAA is wrong. It does, however, raise questions that deserve answers beyond the familiar institutional response that the appropriate procedures have been followed.

In aviation, confidence in the process matters almost as much as the process itself. Passengers trust that aircraft are safe, pilots trust that investigations are evidence-based, and the public trusts that regulators can distinguish safety analysis from assigning responsibility. That is a lot of trust to ask for without showing your working.

On 27 October 2024, SAA flight SA314 departed Cape Town International Airport for O.R. Tambo International Airport aboard an Airbus A330-300, registration ZS-SXJ. The flight carried 223 people: two pilots, 10 cabin crew members, and 211 passengers. Approximately 33 minutes after departure, while cruising at Flight Level 410, the aircraft encountered clear-air turbulence. The event was sudden, and the weather radar did not indicate significant weather in the immediate vicinity.

According to the South African Civil Aviation Authority’s serious incident report, the first officer was flying when the airspeed trend moved rapidly towards the aircraft’s overspeed protection range. She disconnected the autopilot and attempted to control the aircraft manually. The aircraft subsequently experienced pitch oscillations and rapid changes in airspeed.

The captain, who was monitoring the flight, was instructed to request a descent. The seatbelt sign was switched on, and the aircraft descended to Flight Level 390 without further incident. The consequences in the cabin were less comfortable. Four cabin crew members were injured, one seriously and three with minor injuries. Some were thrown against the ceiling or other parts of the aircraft, while hot beverages caused additional injuries. No passengers were injured, and a subsequent structural inspection found no damage. The incident was serious. The injuries were real. The safe landing should not be mistaken for evidence that nothing significant happened.

The SACAA report identified clear-air turbulence as the probable cause and cited the first officer’s disengagement of the autopilot without proper communication with the monitoring pilot, contrary to Airbus operating guidance, as a contributing factor. Poor crew resource management was also identified.

Those are the published findings, and they must be taken seriously. But a safety report does not necessarily resolve every question, particularly when its consequences for an individual pilot extend beyond the technical conclusions. Source: SACAA Aircraft Serious Incident Report CA18/3/2/1458, published in October 2025.

Flight recorders provide objective information about an aircraft’s behaviour and, through cockpit voice recordings, the crew’s communications. They allow investigators to test recollections against recorded evidence.

In the Walsh case, the SACAA report states that flight data recorder information for 27 October 2024 had already been overwritten by the time the occurrence was reported to the Accident and Incident Investigations Division. It also states that quick access recorder (QAR) data was available and downloaded at South African Airways Technical on 12 November 2024.

This distinction matters. Flight data recorders and QARs are not interchangeable in every respect; their capabilities, recording arrangements and accessibility can differ. The available QAR data was used in subsequent technical analysis, including an Airbus report dated 28 March 2025.

It would therefore be inaccurate to suggest that investigators had no recorded flight data. They had QAR data, and Airbus used it to analyse the aircraft’s behaviour and surrounding wind conditions.

Walsh nevertheless alleges that critical recording material, including cockpit voice and flight data recordings, was not retained, leaving her unable to examine important aspects of the incident independently. The precise status of the cockpit voice recording and the evidence available to each party require clarification from the underlying records.

That is the distinction between a fact and an allegation. The confirmed loss of the original flight data recorder information raises questions about reporting and preservation procedures; the allegation concerning the cockpit voice recording raises another. Neither proves that the investigation was improperly conducted, but both warrant clear answers.

• When was the incident formally reported?

• Who was notified, and when?

• What preservation measures were considered?

• What material existed, what was retained, and what was made available to the investigator and the pilot concerned?

These are basic questions about the integrity of an aviation investigation. Evidence preservation should be an obvious priority, preferably before the relevant data disappears.

The SACAA report examined the aircraft’s movements, speed excursions and crew response. Airbus analysed QAR data and found wind variations consistent with the crew’s account of a clear-air turbulence encounter. The report also identified the autopilot disengagement without proper communication as a contributing factor.

That provides a technical foundation for the published conclusion, but it does not automatically settle whether the available data was sufficient to reconstruct events reliably, whether alternative explanations were considered, or whether the conclusions were tested against the applicable Airbus procedures and operating context.

Clear-air turbulence is difficult to predict and can occur without obvious weather cues. The aircraft was operating near a strong jet stream and Airbus found significant variations in wind components across three axes.

That context does not excuse inappropriate control inputs if those occurred. Equally, a control input during an emergency does not by itself establish that it was unreasonable. The relevant questions include what information was available, how quickly the situation developed, what procedures were required, and how the crew communicated. Investigators should explain those matters using the available evidence. Where important evidence is missing, they should also explain how that affects confidence in the conclusions.

An Airbus analysis and a regulator’s report are important, but neither is beyond scrutiny merely because it carries an official logo. If the conclusions are robust, transparency should strengthen them; if uncertainties remain, acknowledging them is more credible than pretending aviation reports descend from the mountain engraved on stone tablets.

There is an important distinction between investigating an aviation occurrence and deciding what should happen to an individual pilot’s career. The SACAA report states that its purpose is to promote aviation safety and reduce the risk of accidents and incidents, not to apportion blame or liability. An employer may nevertheless have separate obligations concerning operational standards, training, fitness for duty and disciplinary procedures. A safety investigation can inform those processes, but it does not resolve every employment or procedural question.

When safety findings are used in a process with serious professional consequences, the fairness questions are straightforward: Was the pilot given access to the material relied upon? Could she challenge the technical interpretation? Were technical conclusions distinguished from individual culpability? Were procedures applied consistently?

Those questions cannot be answered merely by pointing to the existence of an investigation report.

Neither does Walsh’s experience automatically settle the matter in her favour. Experience is relevant context, not immunity from error. Conversely, an adverse technical finding does not eliminate the need for a fair process. Aviation safety and procedural fairness are both necessary if the system is to command confidence.

SAA is opposing Walsh’s application for an independent investigation. According to reporting published on 18 October 2026, the airline argues that another investigation is unnecessary and that the urgency claimed by Walsh is of her own making. It has also maintained that the dispute does not present an ongoing safety concern relating to the aircraft. These are the airline’s reported positions, not findings that settle the dispute.

SAA is entitled to defend its position and insist that legal procedures be followed. But arguing that another investigation is unnecessary is different from demonstrating that every material question has been answered.

The airline should be able to explain what evidence was preserved, what was relied upon, how Walsh could respond to the findings, and what process governs decisions affecting her future. If those matters are documented, making the relevant information available through the proper legal process should help establish that fact.

Legitimate confidentiality or legal constraints can be identified and addressed appropriately. Transparency does not mean publishing every cockpit conversation online; it means ensuring that those entitled to scrutinise the process can do so. For a national airline, credibility is an asset. The public should not have to assemble an investigative jigsaw puzzle from newspaper reports and fragments of court papers.

Walsh’s urgent application has been removed from the urgent court roll. The matter is expected to proceed through the ordinary High Court process in 2027, according to the latest reporting. That does not establish that Walsh’s account is wrong, nor that SAA’s position is right. It means only that the urgent route she sought was not the route through which the matter would proceed.

The substantive dispute remains unresolved in the material reviewed for this article. Until the court considers the relief sought and the arguments presented, it would be premature to declare that Walsh has been vindicated, that SAA has been cleared of every criticism, or that an independent investigation is necessarily the only appropriate outcome.

The point is to identify questions that should be answered by the parties and, where appropriate, the court, without turning an unresolved dispute into a verdict by headline.

The following questions arise from the published incident report and reported legal dispute. They seek clarification, not proof of misconduct.

1. When was the incident reported, and when was preservation action taken?

The SACAA report says the flight data recorder information had been overwritten by the time the occurrence was reported. A clear chronology should establish when the operator, regulator and investigation division became aware of the event and what preservation steps were taken.

2. What happened to the cockpit voice recording?

Walsh alleges that important cockpit recording material was not retained. The published report confirms the loss of the original flight data recorder information, but that alone does not establish the status of the cockpit voice recording. The record should be clarified.

3. What QAR data was available, and who could examine it?

The SACAA report confirms that QAR data was downloaded and used in technical analysis. The parties should clarify what data was retained, whether it was complete for the relevant period, and what access Walsh had to the underlying material and analyses.

4. How were the technical findings communicated and challenged?

The published report identifies control inputs and crew coordination as contributing factors. The separate process affecting Walsh’s career should establish how she was informed of the findings, what opportunity she had to respond, and how technical disagreement was considered.

5. What is the basis for rejecting the request for an independent investigation?

SAA reportedly argues that a further investigation is unnecessary. The substantive question is what existing investigation and evidence are considered sufficient, and why the issues raised by Walsh do not justify further independent scrutiny.

6. What lessons have been drawn for safety management?

The event involved clear-air turbulence, rapid aircraft movements, injuries to cabin crew and cockpit coordination. What practical lessons have been incorporated into training, procedures or safety communications? Any changes should be described accurately and supported by evidence.

This case should not be reduced to a personality contest between a pilot and an airline. The broader issues are the preservation of evidence, the credibility of safety investigations, the distinction between technical findings and disciplinary consequences, and an affected pilot’s ability to challenge conclusions through a fair process.

Aviation depends on disciplined reporting. Crews must report occurrences promptly; operators and investigators must follow applicable procedures; evidence must be preserved when required; and technical conclusions must be grounded in the available material. When questions arise about whether those systems worked properly, the answer should be evidence, not institutional defensiveness.

If pilots believe that challenging an adverse finding leaves them without a meaningful opportunity to examine the evidence, confidence in the safety reporting culture can suffer. That is not a conclusion about Walsh’s case, but it is a risk aviation organisations should take seriously.

Equally, an independent investigation should not become a reflexive remedy whenever someone disagrees with a finding. It must have a clear purpose, sound legal and technical grounds and a defined scope. The objective is to establish what happened as accurately as possible and ensure that the process is fair enough to withstand scrutiny.

So does the report’s statement that the original flight data recorder information had been overwritten, even though QAR data was available and used in subsequent analysis. Walsh’s allegations about other missing recording material and the fairness of the process remain matters to be tested, not facts to be assumed.

The central question is whether the evidence and procedures are sufficient to support the conclusions reached and the consequences that followed.

Flightline Weekly will follow the case as further court documents, official responses and verified evidence become available. Until then, the appropriate position is simple: take the safety findings seriously, take the allegations seriously, and insist that the evidence be allowed to speak for itself.