Haneda Runway Collision

Derek Viner January 10, 2024 0 Comments

A sample formal Occurrence analysis and discussion of its application using the January 2004 Haneda runway collision case.

INTRODUCTION

For a representative press article, see https://www.theguardian.com/world/2024/jan/03/japan-plane-crash-jet-haneda-airport-clear-to-land-japan-airlines

Before Tuesday, Japan had not experienced a serious commercial aviation accident for decades. The crash came weeks after the global airline industry was given new warnings about runway safety.

Local media reports said police would conduct an investigation into whether possible professional negligence had led to deaths and injuries.

“There’s a strong possibility this involved human error’.

I’m saddened that negligence is mentioned as it does not at all help ‘investigation’.  It’s a bit like looking for witches:  if you look you’ll find them, but to what avail?  Also, it is indeed highly likely you’ll find human error as the ATC and flight control system relies on people.  But will that help us control the risk?

FORMAL OCCURRENCE ANALYSIS

Before jumping to subjective judgement, it is valuable to use formal analytical techniques to identify all possibilities.  Below is an example to illustrate application to this case of the formal method.  To understand this see also Chapters 3 and 4.

Event: Aircraft 1 lines up on the duty runway when another aircraft (2) is on final approach.  [The possibility of collision is evident and the energy sources are kinetic energy and gravitational potential energy.]

(possible) Mechanisms of (Reasons for) the Event

1. Air Traffic Control (ATC) erroneously issues a “line up and hold” command to aircraft 1 AND this command is not heard/understood by the crew of aircraft 2

OR

2. ATC issues a clearance to aircraft 1 to “taxi to the holding point” of the duty runway (after which pilots would expect an “line up and hold” or “enter and take off” clearance), but the crew of aircraft 1 understand this as “line up and hold” AND ATC fail to recognise the error in the ‘read back’ of this misunderstood clearance. 

AND

c)  ATC (possibly erroneously) issue a landing clearance to aircraft 2 but neither this nor the ‘read back’ acknowledgement by the crew of aircraft 2 is heard or implications recognised by the crew of aircraft 1.

(possible) Outcomes (results) of the Event

1. The crew of aircraft 1 realise the conflicting situation and broadcast their position AND the crew of aircraft 2 hear and understand this and initiate a go-around. (a Null Outcome)
2. ATC realise the conflicting situation and issue a go-around command to aircraft 2. (a Null Outcome)
3. The crew of aircraft 2 observe aircraft 1 stationary on the runway and initiate a go-around. (a Null Outcome)
4. Aircraft 2’s landing trajectory is such that it touches down beyond aircraft 1. (a Null Outcome)
5. Neither ATC nor the crew of aircraft 1 recognise the conflict AND the crew of aircraft 2 do not see aircraft 1 lined up at the runway threshold AND their approach trajectory is precise and hence results in a collision with aircraft 1.

(possible) Consequences

For Null Outcomes:  investigations

For Outcome 5: damage to both aircraft 1 and aircraft 2 and occupants

Likely Worst Consequence:  hull loss and all occupants killed for both aircraft.

Relevant circumstances

Night time – affects visibility for both ATC and pilots, particularly for pilots on final approach

Relevant conditions

There are two runways in use, having the same direction and called ‘Left’ and ‘Right’.  ‘Right’ is where the collision occurred.

The ‘line up and hold’ point on runway Right is much the same as the touch down point.

In Japan, as in the majority of other countries, English is the language used by ATC.  This may influence the probability that ATC instructions and read-backs are understood. 

All ATC clearances are repeated (“read back”) by flight crew to enable errors in understanding to be identified and rectified.

One of the significant known tasks of flight crew is to maintain situational awareness (who is where?).  This applies to both aircraft’s crews.

The ability to identify aircraft lights from above and to the rear when surrounded by a large number of light sources on runways, taxiways, aprons, roadways and buildings.

Discussion

I’m a light aircraft pilot not a commercial airline pilot. I have some familiarity with practices as I am licenced to fly in controlled airspace.  I’m also a risk engineer, so am familiar with formal analysis techniques.  

I’m aware of technology that can affect the simple analytical structure above – one is ground radar that shows ATC ground controllers (using a different frequency from the Tower controllers who would be controlling aircraft 1 and 2) who is where.  Another is ADSB-Out and ADSB-In, which (if fitted) could be another source of information to the crew of both aircraft about who was on the runway and who was approaching it.

My point in showing this analysis is to illustrate the use of formal analysis techniques and make it evident that when proper effort (people with the required detailed knowledge) is put into developing the analysis it can be used for both pro-active (which improved controls are required where?) and reactive (‘accident investigation’) purposes.

It will be evident to the reader that a significant knowledge of human factors (in particular concerning perception and cognition applied to pilots and controllers) is necessary to properly understand the nature of errors made.