Accident Review: Aerodrome Works Safety & Singapore Airlines Flight 006

An aerodrome is a hazardous environment: lots of moving parts, competing objectives, humans being human, weather, etc. When we need to conduct airside works, we introduce even more hazards and more risk. And this requires a specific set of management activities. One of the worst aviation accidents involving aerodrome works was the 2000 crash involving Singapore Airlines Flight 006 (SQ006) at Taiwan Taoyuan International Airport.

The story behind this accident and its aftermath is complex but let’s look at it from an aerodrome works safety management point of view.

The Event

The accident occurred on 31 October 2000 with the inbound Typhoon Xangsane impacting the airport with heavy rain and poor visibility. SQ006 had stopped in Taipei en route from Singapore to Los Angeles and was departing for its second leg a little after 2300 local time.

The Boeing 747-400 aircraft was cleared for taxi to Runway 05L via a series of taxiways. This taxi route saw the aircraft taxiing southwest parallel to the runway system in the opposite direction to take-off. I say runway system here because between the aircraft on Taxiway NP and Runway 05L was Runway 05R. This is going to become very important in just a moment.

As the aircraft approached the end of the runway system, it turned right onto Taxiway N1. Unfortunately, it continued to turn right (effectively a U-turn) onto Runway 05R and commenced its take-off roll. Approximately 41 seconds later, the aircraft struck the concrete barrier and construction equipment associated with an aerodrome works project on the runway.

The aircraft continued through the construction area, colliding with more equipment, breaking apart and coming to rest about three-quarters down the length of Runway 05R. The airport activated a significant emergency response and brought the fire under control in under 20 minutes. However, 83 people were killed (including two in hospital), 39 were seriously injured, and 32 received minor injuries.

For a more extended presentation and excellent analysis, you can watch Mentour Pilot breakdown the event:

The Investigation

In line with international standards* and best practices, Taiwan’s Aviation Safety Council commenced and completed an investigation of the accident. Their report was extensive but not without controversy. Singapore objected to the report’s supposed focus on pilot error in its findings relating to probable causes.

However, as I dived into the report, I found plenty of detail and analysis on other factors, including aerodrome design, lighting maintenance, ground radar and, what I want to discuss today, aerodrome works management.

But the first thing I noted was that, officially, this accident is classified as a crash on a “partially closed runway during takeoff”. That makes sense given the information I laid out above, but with a bit of reading, you discover that “Runway” 05R’s status was in a state of transition at the time.

* For complex international political reasons, the Republic of China is not an ICAO Contracting State but follows or attempts to follow ICAO SARPs.

Taxiway NC

The Taiwanese authorities had long planned for Runway 05R to be redesignated** as a taxiway. However, the runway was already used as a taxiway regularly enough for green centreline lights to have been installed.

Although the report doesn’t appear to discuss it, the aerodrome works on Runway 05R were repair works. The CAA issued a NOTAM closing the runway on 31 August, with the project starting two weeks later. The NOTAM showed a completion date of 22 November. It is unclear from these details and the report whether these works were directly related to the redesignation decision.

On 3 October, two weeks into the works project, the CAA made the redesignation official. They advised, via NOTAM and AIP SUP, that its complete transition, including new markings, would take effect on 1 November. This was before the completion of the repair work and the day after the accident.

And yet, less than three weeks after this notification, on 23 October, another NOTAM was issued stating that the redesignation had been “postponed until further notice”. The reason behind this was a delay in the procurement of signage, but my concern is that for five weeks of this works project, five weeks of this runway closure, everyone was expecting that this runway would never be a runway again.

** They use the term “degraded”, but I try not to place value judgements on different aerodrome facilities.

Works Planning

I don’t want to judge a twenty-year-old investigation of this magnitude and importance. But, still, I would love to know more about the bimonthly Airport Operators Committee meetings that occurred at the airport. Firstly, this type of committee is a hallmark of a mature safety management system and safety culture. And it was happening at the very dawn of those concepts entering our standards and regulations.

And also because it runs a counter to a statement two paragraphs later that says that “airport users were notified about airfield construction that will affect operations through the issuance of NOTAMs”. And again, with twenty years of hindsight, I feel uneasy about an over-reliance on the NOTAM system.

In the absence of any information to the contrary, it is fair to assume that the aerodrome operator undertook a reasonable analysis of the impact of the works project before issuing the go-ahead to the works contractor. However, some of the subsequent analyses might contradict that assumption.

Filling in the Gaps

The ASC report breaks up its findings into three groups: probable cause, risk, and other. This second group is very interesting. This group has the potential to have the most significant impact on avoiding a repeat of this type of incident, which is the stated objective of this and all Annex 13 investigations. The first five of these findings relate to works safety, with the first four essentially saying that there were no standards to tell the airport how to mark the closed but open runway/taxiway. 

I can’t entirely agree with these findings, but I agree with some related analyses.

The absence of visual indications is a legitimate safety concern. I agree with the report that the aerodrome operator should have identified the risk of an inadvertent takeoff in its works planning analysis. But I don’t think the answer was more makings; it was less.

Today and at the time, ICAO SARPs require the obliteration of markings on permanently closed runways. And I believe the data and analysis suggests that, while the taxiway remained open, this runway was permanently closed. 

The works were started two weeks before the official announcement of the re-designation and were expected to continue through this process. So, why weren’t the runway markings obliterated in those five weeks leading up to the announcement of the delay? Maybe the answer lies in a different works plan that wasn’t analysed as part of the investigation, but these are some tight timeframes.

The report calls out Annex 14 for being “vague” on some of its terms and not specifically addressing the scenario at the root of this accident. As an aerodrome standards officer, I find this unfair. Safety standards can’t address every specific situation. Hence the emergence of safety management systems, risk management and outcome-based regulation.

And before you point out that this was over twenty years old, I understand, but this report already discusses risk analysis and other safety management concepts.

It is essential that aerodrome operators undertake a risk assessment before conducting aerodrome works. In Australia, we have several structured processes embedded in our standards, and I’ll be going over these in the coming days.

Backtracking

Obviously, it is easy to be critical of a report from the early 2000s. I acknowledge that this investigation team had an enormous number of factors to explore and analyse and an intensely political environment. Nevertheless, this report provides countless insights into complex accident causation, human factors, and aerodrome design and maintenance.

The fact that I can sit here and mine this report for even more lessons is a testament to the original authors.