FlightAware Discussions

QFA2906 YPPH to Antarctica


#1

There is a flight currently flying YPPH to 66S100E and then back to YPPH.

How long will this be tracked? Will ADS-B track it beyond normal radar range?


#2

Seems to have dropped off tracking completely on leaving ground based ADS-B tracking… Now sitting south of Perth as a pending arrival. I’m seeing ADS-C positions on Planeplotter of QFA64 Jo’burg to Sydney well south over the ocean, but nothing on 2906.

EDIT: Just popped up on my PP feed about 600 nm S of Perth


#3

Good catch!

The tracklog shows the last real position reported out of Bunbury at 1146 (AEDT) but then Flightaware has estimated it turned around and is almost back to Perth.

If it had turned around it would have been tracked by all the ADS-B receivers south of Perth so that is probably incorrect. If it is true there will be a lot of irate passengers.


#4

It appears to be a 13 hour flight, started only a couple of hours ago, so I guess it is still on the way southwards.

The plotted estimated position is obviously incorrect. I guess the originating and destination airports being the same place has badly confused things.


#5

The flight is shown as having returned to the original airport but is shown as an estimated position. The timeline shows estimated arrival time as being almost four hours early. In reality the flight is still somewhere over Antarctica.

I think this exposes a bug in the software that does the estimating.

It will be interesting to see if the real arrival some four hours later will actually be tracked. The flight status is “landing shortly” and has been that way for quite a while.

EDIT: Sometime later, the tracker said “estimated landing 4:58 pm” and the track log also said this but didn’t say “estimated”.


#6

The flight is landing at 2309hrs according to Qantas website


#7

The original timeline showed it to be landing in an hour and a half, total flight time 13 hours 4 minutes (at one point, timeline showed 7:54 ? am to 8:58 pm australian time, but then it actually took off at 8.05 am australian time with estimated landing at 9:09 pm). I guess it will be landing in about two hours time.

Based on what happened after departure, it seems that it will be in range of ground-based tracking for only the last 30 to 40 minutes of flight.


#8

Flight reappeared on the tracker with live data almost an hour ago but insists in saying it landed three hours earlier. Once live data is received again, it would be useful for the earlier false claim of landling be removed. In the track log, the 4:58 false landing should have said “estimated” alongside for avoidance of doubt.


#9

Data stopped at 12:35 (8:35 australian) while still several miles from the runway.


#10

Having displayed an estimated landing time that was four hours early, the estimate was not removed when live data reappeared from the flight several hours later, nor has it been updated after it actually landed. Additionally, the real landling is not recorded in the tracking file.


#11

A same-airport destination (or more generally, a flight where the distance to destination doesn’t generally decrease as the flight progresses) is an unusual edge case. Some of our estimation logic looks at the distance from the last known position to the destination, which is going to be pretty misleading in this case.

FWIW the version of the flight including Aireon data had full coverage:

Screenshot%20from%202019-01-26%2013-34-18


#12

I guess the Aireon tracked part of the flight isn’t yet public data?

On the flight history, as it currently stands, the fake 4:58 landing means that the final leg of the journey, several hours later, will not replay.

Additionally, the altitude and speed graphs cut off at 4:58 too, some four hours early.


#13

Can we at least get the tracking log fixed by moving the landing from 8:58 am (4:58 pm) to 12:38 pm (8:38 pm), almost 4 hours later?

Currently, the landing is noted in the middle of the tracking log and is not marked as an estimate. In reality, the landing should be the very last line of the log.


#14

This Antarctic flight has been most interesting in exposing some issues with the tracking.

  1. I expected the flight to be untracked (on the public system) for the majority of the flight. What was unexpected (but quite logical after thinking about it a bit more) was that the estimated position did not follow the planned track but instead drifted slowly back to the airport and then “hovered” there for a very long time.

  2. After “hovering” over the airport for many hours, the system then decided that the plane had landed 4 hours early. However, the entry in the public tracking log merely says “arrival”, i.e. it fails to say that is an “estimated arrival”. Perhaps the system should wait longer before adding an “estimated arrival” entry? Perhaps it should wait until after either the originally planned arrival time has passed or the originally planned flight duration has elapsed? But, either way, flagging such events as “estimated” in the tracking log would seem to be a priority.

  3. When live data started to be received again several hours later, the earlier (estimated) “arrival” should perhaps have been removed from the tracking log (or the software should ignore that false entry from that point on).

  4. The real landing time has still not been added to the tracking log. It seems the fact that there is already a “landing” earlier in the track log, albeit only an “estimate”, one that was badly incorrect by many hours, prevents the real landing from being added. This means the flight duration shown on the tracking map page is incorrect.

  5. Somehow, these errors mean that replaying the flight fails part way through, the final leg (final hour) does not play. Additionally, on the same page as that tracking map, the speed and altitude graphs are cut off almost four hours early.

This looks like several interesting challenges for your programming team!

In the meantime, can the landing time issue be fixed in the track log for this flight?


#15

Looks like the Antarctic sight-seeing flight.
Qantas resumed them in 1994, after the TE901 disaster.


#16

It is, and the track can be seen on this map: https://uk.flightaware.com/adsb/coverage#data-coverage

Select the “Aireon space-based ADS-B (only)” option and then add the “Navigation” option to it to see the full track.