Flights Gone Wrong

Real stories. Deeper understanding.

What Changed After Tenerife: CRM and Clearer Radio Calls

The 1977 runway collision at Tenerife killed 583 people and turned on a few words of radio and a cockpit where no one stopped the captain. This guide explains the two big changes that followed, standard radio phrasing and crew resource management, and where they have since saved lives.

Two things changed after Tenerife. Radio language became stricter, so that a route clearance could no longer be mistaken for permission to take off. And airlines began to train crews to work as a team, where the first officer and flight engineer are expected to speak up and the captain is expected to listen. That training became crew resource management, or CRM.

What went wrong in the cockpit and on the radio

On 27 March 1977 two Boeing 747s, KLM Flight 4805 and Pan Am Flight 1736, collided in fog on the runway at Los Rodeos Airport on Tenerife. Both had been diverted there after a bomb exploded at Las Palmas. All 248 people on the KLM aircraft and 335 on the Pan Am aircraft were killed, 583 in all. Sixty-one people on the Pan Am aircraft survived. The full sequence is on the Tenerife airport disaster event page.

The FAA’s summary of the Spanish investigation sets out the last minute:

  • At 17:05:53 the controller gave KLM its route clearance, the instructions it would follow after take-off. It was not a take-off clearance.
  • KLM’s first officer read it back and added “we are now at take-off” as the captain began the take-off roll. The controller seems to have understood “we are now at take-off position” and replied “okay”.
  • He then said “stand by for take-off, I will call you”. At the same moment the Pan Am crew transmitted that they were still taxiing on the runway. The two calls overlapped, and the KLM crew did not hear them clearly.
  • KLM’s flight engineer asked whether Pan Am was clear. The captain replied emphatically that it was.

The Spanish board found that the fundamental cause was that the KLM captain took off without clearance, did not obey “stand by for take-off”, did not abort when Pan Am reported it was still on the runway, and answered his flight engineer emphatically. The report also noted that the captain’s great prestige made it hard for the first officer and flight engineer to imagine an error of that size, and that they raised no further objection.

The runway side of the story, including taxiing on an active runway in low visibility and the lack of ground radar, is covered in our guide to runway incursions.

Standard phrasing: why “take-off” is now only said with a clearance

The Spanish board made three recommendations, as quoted by the FAA:

  1. Great emphasis on the importance of exact compliance with instructions and clearances.
  2. Use of standard, concise and unequivocal aeronautical language.
  3. Avoiding the word “take-off” in the ATC clearance, and adequate time between the ATC clearance and the take-off clearance.

The third recommendation is now written into radio manuals. The UK Civil Aviation Authority’s Radiotelephony Manual (CAP 413, paragraph 2.67) puts it plainly: an ATC route clearance is not an instruction to take off or enter an active runway. The words “take-off” are used only when an aircraft is cleared for take-off, or when that clearance is cancelled. At all other times the word “departure” is used.

So a pilot now reports “ready for departure”, never “ready for take-off”. The only time anyone says the word is in “cleared for take-off” or “cancel take-off clearance”. The paragraph before it (2.66) says controllers will never pass a route clearance to a pilot who is lining up or taking off, which addresses the second half of the Spanish recommendation.

Read-backs matter for the same reason. A read-back lets the controller hear what the pilot understood. At Tenerife, the read-back contained the warning sign, “we are now at take-off”, but it was not challenged in time.

From cockpit hierarchy to crew resource management

Tenerife was not the only accident of the 1970s in which a crew had the information to prevent disaster but did not use it. Two American accidents pushed the change along.

Eastern Air Lines Flight 401, Everglades, 29 December 1972. An L-1011 TriStar crew became absorbed in a landing gear indicator light and did not notice the aircraft descending. According to the FAA, 94 passengers and 5 crew were fatally injured, and two survivors died later. The FAA notes that after this crash the industry reviewed ways to address poor crew performance in emergencies.

United Airlines Flight 173, Portland, 28 December 1978. A DC-8 circled near Portland while the crew dealt with a landing gear problem, and ran out of fuel. Eight passengers, the flight engineer and a flight attendant were killed. The NTSB found the captain failed to monitor the fuel state and to respond to his crew’s advisories about it (NTSB-AAR-79-7). The FAA’s summary says the second officer did not express the right level of concern, and that the accident was pivotal in showing the need for a team approach and for more assertive junior crew members.

In June 1979 NASA held an industry workshop in San Francisco called “Resource Management on the Flight Deck” (NASA CP-2120). According to the FAA, United Airlines adopted the first formal CRM programme in 1981, following that workshop, and other airlines quickly followed.

The FAA defines CRM as the effective use of all available resources: people, hardware and information. The term began as “cockpit resource management”. It became “crew resource management” as training grew to include flight attendants, dispatchers, maintenance staff and air traffic controllers (AC 120-51E). In the US, the FAA notes, CRM training has since become a regulatory requirement.

CRM working: Sioux City, Gimli and the Azores

United Airlines Flight 232, Sioux City, 19 July 1989. A DC-10 lost its normal flight controls after its tail engine failed. An off-duty DC-10 check airman in the cabin came forward and worked the throttles while the crew steered the aircraft with engine power alone. Of the 296 people on board, 111 died and the rest survived the crash landing. The NTSB said the crew’s interaction during the emergency was “indicative of the value of cockpit resource management training, which has been in existence at UAL for a decade.” United 232 will get its own event page on this site.

Air Canada Flight 143, the Gimli Glider, 1983. When the 767 ran out of fuel, First Officer Maurice Quintal worked out that they could not reach Winnipeg, and Captain Robert Pearson, with air traffic control, turned for the disused base at Gimli. All 69 people survived. See Air Canada Flight 143.

Air Transat Flight 236, 2001. After both engines stopped over the Atlantic, the crew glided the A330 to Lajes in the Azores, and all 306 people on board survived. The same flight also shows the limits of CRM: the crew had opened the fuel crossfeed from memory, without the checklist, and sent their remaining fuel out through the leak. See Air Transat Flight 236 and our guide to planes that ran out of fuel.

CRM today: managing error, not only avoiding it

Modern CRM accepts that pilots will make mistakes. The FAA’s guidance says it is now understood that pilot errors cannot be entirely eliminated, so crews need error management skills: preventing errors, catching them, and recovering from them (AC 120-51E). The same guidance says CRM training has three parts: an introduction, recurrent practice with feedback, and continual reinforcement.

The FAA also notes that investigations have shown human error contributing to 60 to 80 percent of air carrier incidents and accidents, and that many of the problems are about decision-making, communication, leadership and managing tasks, not flying skill. That is why investigators rarely stop at “pilot error”. Our guide to the Swiss cheese model explains how investigators look past the last mistake to the conditions that allowed it. The pilot error topic page lists the events on this site where crew actions were part of the finding.

What has not changed

Some of the conditions at Tenerife still exist. Busy airports still sometimes taxi aircraft on active runways. Radio is still a shared channel where two calls can block each other. Fog still rolls in faster than forecasts say.

The lessons are now in procedures rather than in hope: standard words, read-backs, a crew culture where “Is he not clear, then?” gets a proper answer. Both aircraft at Tenerife were Boeing 747s; our Boeing 747 topic page lists the other 747 events on this site, and Is flying safe? shows how the accident rate has fallen since.

For how investigators reach findings like these, see how to read an accident report.

Sources

  1. FAA (includes the Spanish board’s findings and recommendations). Lessons Learned: KLM Flight 4805 and Pan Am Flight 1736, Tenerife (PH-BUF) Source
  2. Subsecretaría de Aviación Civil, Spain. Spanish accident report on the Tenerife collision (copy hosted by the FAA) Read the report
  3. UK Civil Aviation Authority. CAP 413 Radiotelephony Manual, Edition 24, paragraph 2.67 Source
  4. FAA (summarising NTSB-AAR-79-7). Lessons Learned: United Airlines Flight 173, Portland (N8082U) Source
  5. NTSB. NTSB-AAR-79-7, United Airlines DC-8-61, N8082U, Portland, Oregon, 28 December 1978 Source
  6. NASA (NTRS record). Resource Management on the Flight Deck, NASA/Industry Workshop, San Francisco, 26–28 June 1979 (NASA CP-2120) Source
  7. FAA. Lessons Learned: United Airlines Flight 232, Sioux City (N1819U) Source
  8. FAA. Advisory Circular 120-51E, Crew Resource Management Training (22 January 2004) Source

More guides

All guides
  • Runway Incursions: How Tenerife and Near Misses Shaped Rules

    Causes

    A runway incursion is any aircraft, vehicle or person wrongly on a runway, and the worst ones end with two aircraft on the same strip of concrete at the same time. Tenerife in 1977 remains the deadliest accident in aviation history, and later collisions at Detroit, Los Angeles and Milan led to the radio rules, surface radar and automatic warning lights in use today.

  • Why Planes Run Out of Fuel: How Fuel Exhaustion Happens

    Causes

    Airliners rarely run out of fuel, and when they do it is almost never one mistake: a wrong conversion factor, an undetected leak or a crew unable to act has to get past several checks. This guide explains how it happens, using the official findings on the Gimli Glider, Air Transat Flight 236 and Helios Airways Flight 522, and what changed afterwards.

  • The Swiss Cheese Model: Why Air Accidents Need Many Failures

    Investigations

    The Swiss cheese model, developed by psychologist James Reason, explains why an air accident almost never has a single cause: several layers of defence each have to fail, and their weak points have to line up. Two real accidents, the Gimli Glider and Helios Airways Flight 522, show how it works.