Week 5: Tenerife, 1977

583 people, one runway, and a sequence in which the phraseology failed at four separate points — read as evidence, not as a story

22 March 2027

Open the slides

On 27 March 1977, at Los Rodeos on Tenerife, a KLM 747 began its take-off roll in fog without a take-off clearance. A Pan Am 747 was backtracking the same runway. Both aircraft were destroyed. 583 people died. 61 survived, all from the Pan Am aircraft.

You already hold the rule from week 4. Today the accident is evidence.

Why both aircraft were there at all

Neither was scheduled into Tenerife. Both were bound for Las Palmas on Gran Canaria, which closed after a bomb detonated in the terminal. Los Rodeos took the diversions, and a small airport filled with widebodies parked on its taxiway.

Which is why the Pan Am aircraft was on the runway: with the taxiway obstructed, departing aircraft had to backtrack the runway and turn off part way down. And Los Rodeos sits at about 2,000 ft, where hill fog arrives quickly. Visibility was falling through the sequence.

Plan of Los Rodeos: a parallel taxiway blocked by parked aircraft at its apron end, and one runway carrying a 747 lined up for departure at the far end while a second 747 taxis toward it, past the third of four exits.
  1. The taxiway was full. Las Palmas closed after a bomb in the terminal and Los Rodeos took the diversions. Parked widebodies filled the apron and the stretch of parallel taxiway leading out of it.
  2. So the runway was the only way out. The KLM 747 entered at the far end, taxied the full length of the runway against the direction of departure, turned 180 degrees and lined up. That manoeuvre is a backtrack.
  3. The Pan Am 747 was still on it. It had entered behind, backtracking the same strip toward the same end, and was told to leave the runway at the third exit — counted from where it entered.
  4. The third exit. Taking it meant turning back through about 148 degrees, which in a 747 is not a turn-off so much as a reversal.
  5. The fourth exit. About 45 degrees, further down, and where the crew were heading instead. The collision happened before they reached it.
Schematic: relative positions and the two turn angles, not distances or a survey. And nothing in it was visible to anyone in it — hill fog, no ground radar, and a tower that could not see its own runway.

This matters for the course’s argument. The phraseology was operating outside the conditions it was designed for — non-routine ground movement, no ground radar, aircraft that could not see each other, and a crew under commercial pressure over duty hours. A designed language has an envelope, and this was outside it.

The four points where the language failed

One. A non-standard progress report. The Pan Am crew were asked to report when clear of the runway. The KLM crew, holding at the end, transmitted a report of their own status using the construction week 4 dissected:

KLM      We are now at takeoff.

Ambiguous between positioned at the takeoff point and commencing takeoff. The controller, expecting the first, replied with a phrase that in the phraseology of the day was not a clearance and not a prohibition:

TWR      OK. Stand by for takeoff, I will call you.

OK has no defined meaning. It arrived first, and it arrived as acknowledgement of whatever the KLM crew believed they had said.

Two. A simultaneous transmission. As the tower transmitted “stand by for takeoff, I will call you”, the Pan Am crew transmitted that they were still taxiing on the runway. The two overlapped. In the KLM cockpit the critical portion was obscured — a heterodyne where the prohibition should have been.

This is week 1’s simplex channel, doing exactly what week 1 said it does. No transmitting station could know it had happened.

Three. The reserved word, used outside its clearance, by ATC. “Stand by for takeoff” contains the word. In fog, partly stepped on, a transmission containing TAKEOFF reached a crew already committed to going.

Four. A challenge that was not pressed. The KLM flight engineer asked whether the Pan Am aircraft was clear. He was answered in the affirmative by both pilots and did not ask again. That is a crew-resource-management finding rather than a phraseology one, and this course is careful about the boundary: there was no phrase available to him for I believe the clearance is invalid. Week 12 comes back to whether there should be.

What changed afterwards

The recommendations emphasised standard phraseology in all normal communications, and specifically the avoidance of “take-off” in departure clearances. That is the direct ancestor of the rule you learned last week.

What to be careful about

Tenerife is the most-told accident in aviation, and the telling has smoothed it. Two cautions:

Read the ALPA human factors report alongside the findings. It disagrees with the Spanish report in places, and the disagreement is instructive.

Teaching team