Why don't airliners have a navigation GPS as "back up" in case of a pitot tube malfunction? Aeropero 603 and Air France 447 plus many other im sure, could have been saved if these aircrafts would have had a simple Navigation GPS as a plan B. The GPS could have given the pilot the basic information needed to keep flying instead of leaving them confused about what the instruments were showing due to the blocked pitot tubes. having reliable information of ground speed, heading and altitude would have helped the pilots recognize the situation they were in, which was caused by Pitot tubes blockage. Most private pilots now use GPS why not the big airliners? There is many blogs about why airliners do not have GPS tracking so the planes could be found after a crash, but none talking about why GPS is not use to help prevent some of those crashes.

  • 17
    $\begingroup$ The pitot tube is not used for ground speed, but for indicated airspeed (and calibrated airspeed when corrected for instrument error, true airspeed when corrected for temperature and pressure). Knowing your groundspeed from a GPS is not going to get you your airspeed, and it's the indicated/calibrated airspeed that is telling you where you are within the flight envelope. The higher you go, the greater the difference between groundspeed and indicated airspeed. At 40,000 feet, you're going to have well over 100 knots difference between ground speed and calibrated airspeed as I remember. $\endgroup$
    – Terry
    Feb 1, 2015 at 1:13
  • 5
    $\begingroup$ @Terry Very true. The difference will be even larger when you're flying in a 180 mph jet stream. On airliners flying through the Northern polar jet stream, I've seen ground speeds that would have been well over Mach 1 in still air. $\endgroup$
    – reirab
    Feb 1, 2015 at 5:32
  • 2
    $\begingroup$ @Trebia "USA already darked Kosovo." That's interesting. Do you have a link for that? It seems like that would be really hard to do for such a small area (relative to the area that the signal from a satellite would hit,) unless, of course you just jammed the frequencies from the ground. $\endgroup$
    – reirab
    Feb 1, 2015 at 14:48
  • 1
    $\begingroup$ @reirab in response to pressure from the FAA, in 1996, Bill Clinton signed a bill precluding Selective Availability from being used after May 2000. It seems like Trebia's information is about 15 years out of date. Newer GPS Block III satellites (est. launch 2016) supposedly don't even have the feature. Here are some interesting error plots just before and after SA was terminated. $\endgroup$
    – Nick T
    Feb 1, 2015 at 21:55
  • 1
    $\begingroup$ Hmm, actually we already have this question: Could the GPS be used to aid the autopilot with speed? $\endgroup$ Feb 1, 2015 at 22:14

3 Answers 3


Airliners do generally have GPS receivers, but GPS doesn't give you the information that's needed in this case.

In AF 447, the primary consequence of the blocked pitot tubes was that the flight envelope protection system couldn't do its job properly and therefore automatically disengaged together with the autopilot. The subsequent crash was caused by the pilots' confusion and failure to realize which consequences this had for the aircraft's reaction to their control inputs.

However, what this system needs to function is not something a GPS system can possibly provide. GPS knows how fast the aircraft is moving relative to the earth, but for flight envelope protection you need to know how fast the aircraft is moving relative to the surrounding air, which GPS doesn't have any way of knowing. The differences between these values is caused by wind and can be pretty large. It would be a recipe for disaster to let the computer limit what the pilot can do based on an arbitrary assumption that there's no wind. (And, as Terry comments, there's an additional difference caused by air pressure).

If I remember correctly, the AF 447 investigation concluded that the pitot tubes had thawed and began functioning properly well before impact; it was not the loss of airspeed indication that prevented the pilots from salvaging the situation. By this time the pilots were confused and thought the airspeed indication was still malfunctioning because it showed a much slower speed than they thought they ought to have -- whereas in fact it was working perfectly well.

  • $\begingroup$ On the other hand, one could theoretically mix in satellite weather and GPS-based altitude to get an idea of indicated airspeed. Though I don't know how accurate that satellite data is for wind speed at different altitudes. $\endgroup$ Feb 1, 2015 at 4:30
  • 3
    $\begingroup$ @raptortech97: I don't think satellites have the foggiest idea about wind speeds and directions inside a thunderstorm. $\endgroup$ Feb 1, 2015 at 4:32
  • 2
    $\begingroup$ @raptortech97 Satellite weather data doesn't directly provide wind information at all, let alone with enough accuracy to use for an aircraft's airspeed indication. You can get some idea of winds by seeing how fast the clouds are moving, but this is only helpful for determining averages over a large area, not for giving an accurate reading for a particular position/altitude combination. If we could get accurate wind information from satellites, Hurricane Hunters wouldn't exist. $\endgroup$
    – reirab
    Feb 1, 2015 at 5:26
  • 1
    $\begingroup$ Wind would cause true airspeed to differ from ground speed, but it is vital to know indicated airspeed, which you can get from true airspeed only by knowing the pressure altitude. Given an altitude from a GPS (if you can trust it) you could estimate pressure altitude, but with an error that depends (again) on current local atmospheric conditions. There may be additional complications I haven't mentioned. This makes for a very complex calculation with lots of opportunity for error. $\endgroup$
    – David K
    Feb 1, 2015 at 13:52
  • 2
    $\begingroup$ @DavidK IIRC, airspeed is one of those nice things where people thought they should be measuring one thing (TAS) that was very hard to measure, but it turns out that the approximation that is easy to measure (IAS) is actually the more important number. $\endgroup$
    – cpast
    Feb 2, 2015 at 0:16

The short answer to this question is that commercial airliners do have GPS and they have for a long time. AF447 almost certainly would have been equipped with GPS. While what Terry said about the difference between airspeed and ground speed is very true, AF447 could have seen from their GPS unit that they were moving much slower than what they should have been.

Unfortunately, judging from the cockpit voice recorder transcripts, it seems that they were panicking and failing to think logically or communicate well with each other. It was almost entirely a case of pilot error, not lack of accurate information. The pilot in the right seat was almost continually pulling back on his side stick. This caused the aircraft to climb and bleed off airspeed until the wings stalled. Even after the aircraft entered a stall condition, the pilots never lowered the nose, which would have recovered from the stall. Instead, the pilot in the right seat actively held the nose high for the entire descent to the ocean.

Also, the pitot tubes thawed several minutes before the actual crash. All of the instrument indications were correct for quite a long time before the aircraft hit the ocean.

For more information on AF447, see Popular Mechanics' annotated cockpit voice recorder transcript of the incident.


Airliners do have GPS units. As the other answers and comments have indicated, GPS does not provide airspeed (or angle of attack) information.

It would seem reasonable to use the GPS as a backup in the following way: Under normal operation, the difference between TAS and the GPS estimate (allowing for some computational latency) should be fairly constant (the difference being the wind speed). When the difference exceeds some noise threshold for a period of time it issues a warning.

This wouldn't help in the case of Aeropero 603 but may have helped in the Air France 447 incident. (To clarify what I mean by help, I mean to reduce cockpit confusion.)

In the former case, a similar tracking system between the radar altimeter and the barometric altimeter may have helped.

  • $\begingroup$ It's not clear how you think that would have helped AF 447. The flight computers did realize the airspeed indication was unreliable -- that's why they went into alternate law and disengaged the autopilot. From that instant onward, the pitot tubes were completely irrelevant to the accident sequence. The pilots were not prepared to be given manual control at that point in the flight, and in their confusion ended up flying the fully functional aircraft into a stall so it dropped into the sea. $\endgroup$ Feb 1, 2015 at 21:23
  • 1
    $\begingroup$ Also, AF447 was flying through a thunderstorm at the time of the accident. Thunderstorms are characterized by strong turbulence, that is, dramatic local variation in wind speeds and directions, strong updrafts and downdrafts close to each other, etc. So your assumption that the wind velocity would be "fairly constant" wouldn't necessarily hold in those conditions anyway. $\endgroup$ Feb 1, 2015 at 21:32
  • $\begingroup$ In addition to what @HenningMakholm said, pilots use indicated airspeed, not true airspeed. At the altitude airliners fly, the difference between true and indicated airspeed is very large because of the much lower atmospheric pressure at those altitudes. $\endgroup$
    – reirab
    Feb 1, 2015 at 21:49
  • $\begingroup$ It may have helped to have an unreliable airspeed indication which may have given the crew some idea of what was happening. This might have lessened their confusion. Knowing that the icing incident was over may also have helped. $\endgroup$
    – copper.hat
    Feb 1, 2015 at 21:50
  • 2
    $\begingroup$ The only thing that it could have saved the flight to tell the pilots was: "The controls are now in alternate law and will let you fly into a stall. Take care not to do that". Telling them that doesn't depend on having a fanciful backup airspeed determination system. $\endgroup$ Feb 1, 2015 at 21:56

Not the answer you're looking for? Browse other questions tagged .