Singapore AAIB Report - B777-300ER Loss of Separation Incident (Houston)

[WillowRun 6-3]

Air Accident Investigation Board of Singapore (Ministry of Transport) has issued its report regarding a loss of separation incident involving a B777-300ER aircraft executing a Standard Instrument Departure from Houston in early July 2014. Incident appears to implicate both CRM and TCAS advisory elements of safety.
Report link: http://www.mot.gov.sg/uploadedFiles/Ministry_of_Transport/Content_Blocks/About_MOT/Air_Transport/AAIB/SQ61%20loss%20of%20separation%203%20Jul%2014%20-%20FR%202015-11-11.pdf

[ATC Watcher]

Very interesting report , lot of small errors by everyone as usual . What I still find most worrying is that nearly 13 years after the collision in Ueberlingen , we still see crews ( and from established airlines ) still not following RAs when conflicting with an ATC clearance ! We all définitively and collectively failed to learn and train our people correctly, sad and discouraging.

[WillowRun 6-3]

ATC Watcher - indeed, the TCAS aspect of the report (and, of the incident itself) struck me as quite significant, possibly worthy by itself of a thread. The mid-air tragedy over Uberlingen led to a product liability lawsuit under EU legal principles; according to one published report, the absence of reversal Resolution Advisories was a defect in the TCAS system and/or its software as it existed at that time. At least, in this incident, reversal RAs were generated by the system, according to the AAIB report.
Fortunately the Houston incident will not become one of the subjects of study in a law school course, unlike the tragic mid-air and its nearly unbelievable aftermath. Still, the CRM aspects appear highly instructive. Not least, in situations where there was no calamity and the aviators can provide input, perhaps more lessons can be drawn, as compared to determinations of cause and effect where only things like wreckage, data recorders, and laws of chemistry and physics can do the talking.

[peekay4]

Maybe the new "climb via SID" phraseology should also be amended when there is a Top Altitude restriction.

E.g, "climb via SID, maintain 4000".

[JammedStab]

The obvious thing to say of course is, read everything on the chart. Easy to say but the question is, why did they not?

You can see what happened in this quote from the report which shows how easy you can miss something with EFB charts.

"According to the PF, after reading the first line in the ‘Routing’
section of the text box at the bottom left corner of the chart, he
scrolled to the pictorial portion at the right of the chart and zoomed
in to check on the track and distances. He ended his briefing on the
chart without returning to the ‘Routing’ section of the chart, which
contained an altitude restriction of 4,000ft"

So with EFB charts which may be a bit difficult to read, you can easily find yourself zoomed in to one area, then scrolling to another area. Then it turns out you missed something. So, it is probably best to read everything in the one area such as the initial climb portion and the routing section, then move to the pictorial/notes/climb gradient/etc areas. And at the end, zoom out and look over the chart to see if you can find anything you missed.

It was also strange that the report seemed to agree with the statement that the RA "Climb, Crossing, Climb"(I think this may no longer exist) had not been experienced before and was therefore the reason for not climbing. But the RA does say Climb so why descend. And pretty much everyone else has never had a "climb, crossing, climb" so would we all be expected to also not do it properly and not climb when it says climb. Plus you typically follow the commands of on the PFD in this situation so why put the vertical speed in the red even if you misinterpret what the ADI is saying to do. I would suspect that they likely had another reason for not climbing such as not hearing the RA as they were so focused on the ATC instruction. Audio warnings can be tuned out in times of stress.

[xyze]

Maybe the new "climb via SID" phraseology should also be amended when there is a Top Altitude restriction.

E.g, "climb via SID, maintain 4000".

Per the last paragraph on page 7 of the report, there is ALWAYS an altitude restriction or top altitude in the SID when this phraseology is used. The failure of the crew to recognise this at the time they received their clearance was another missed opportunity to catch their error of omission.

Curious that this factor is not listed in the discussion as it suggests that the operator's efforts to educate its flight crews regarding this new phraseology was ineffective, at least in the case of the crew involved in this incident.

[ATC Watcher]

JammedStab
I would suspect that they likely had another reason for not climbing
I though we had passed these "excuses" part, the way TCAS works you cannot have any form of excuse for manoeuvring against the sense of an RA. Under mode S , RAs are coordinated and the other aircaft is then very likely to follow its own RA and manoeuvre directly towards you.
I am constantly amazed and sad to see how many pilots and controllers still do not get it .
Pity this report does not go into what BOTH aircraft did regarding TCAS . I suspect the intruder at 6000 ft did follow his own decent RA and that is why they ended up 200 ft apart .
As to issuing a DEP SID clearance mentioning the restriction altitude ,as a Controller , that would make a lot of sense, especially if you have a traffic opposite .

[peekay4]

Per the last paragraph on page 7 of the report, there is ALWAYS an altitude restriction or top altitude in the SID when this phraseology is used.
The question we should be asking is, "how can we prevent or catch this type of mistake in the future?"

E.g., SQ vows to implement better training and procedures, but what about other flight crews? How do we ensure that the new training and procedures are 100% effective and always complied with?

Currently when there is a Top Altitude, the initial altitude is optional in the clearance phraseology. Simply including this altitude and verifying correct read back from the pilots solves this problem.

Besides, specifying the initial altitude is already mandatory in other situations. Currently:


If there is a Top Altitude, then the initial altitude is optional, and the phraseology is "climb via SID"
If there is no Top Altitude, then the controller will assign an initial altitude, and the phraseology is "climb via SID, maintain (altitude)"
If there is a Top Altitude, but a different altitude needs to be assigned, the phraseology becomes "climb via SID, except maintain (altitude)"

To me, always specifying the initial altitude makes the "climb via" procedure more consistent and safe, with very little extra effort from both controller and crew.

[JammedStab]

JammedStab

I though we had passed these "excuses" part, the way TCAS works you cannot have any form of excuse for manoeuvring against the sense of an RA. Under mode S , RAs are coordinated and the other aircaft is then very likely to follow its own RA and manoeuvre directly towards you.
I am constantly amazed and sad to see how many pilots and controllers still do not get it .
Correct.



As to issuing a DEP SID clearance mentioning the restriction altitude ,as a Controller , that would make a lot of sense, especially if you have a traffic opposite .

I believe the whole point of the SID is so that the controller does not have to mention these things. If the controller starts having to mention the altitude restrictions because of an error, then the controller starts having to mention the routing because of another error, then the speed restrictions because of someone else's error, then the controller might as well do it the old way where there were no SID's. And then after that, there will be a miscommunication between controller and pilot leading to an error requiring SID's to be issued again.

The better solution is to just read the chart properly. No system is foolproof. But printing important things more clearly, as has been done after this incident should help.

[peekay4]

The better solution is to just read the chart properly.
I don't really agree with that sentiment. Safety procedures are based on a balance of cost, benefit, and risk.

Omitting the top altitude seems of negligible benefit but at a potentially huge risk, as this incident clearly demonstrates. No one is calling for the entire SID to be confirmed.

We shouldn't have to wait for an accident before changes are made.

[JammedStab]

A change was made. The top altitude is now written more clearly on the chart.

[peekay4]

On FAA charts, which was not used by this crew.

And that change was only made following this incident after at least 10 years of proposals and debate within various FAA workgroups on whether or not to include it. :rolleyes:

I believe the Jeppesen proposal was to add a "Top Altitude" column but still within the departure routing text section.

Unfortunately this may not be the last incident re: top & bottom altitude confusion.

[Back2Final]

Dialling up FL 310 in the MCP seems like the big hole in the cheese. Any other crews do this when unsure of an altitude?

[fullforward]

It's funny to notice reading the posts here how this matter is being taken lightly: I just imagine what people would be writing here if the concerned airline was some of less prestige than SIA. For instance one from Africa, South America, some Korean or Chinese or even Air France.
There would be lots of gun firing!
IMHO, this is just a simply case of poor airmanship coupled with lots of complacency/negligence. No excuses, no empty blah blah blah...:=
It's simply not acceptable that a trained and experienced international crew with such an unprofessional attitude.

[JammedStab]

On FAA charts, which was not used by this crew.

And that change was only made following this incident after at least 10 years of proposals and debate within various FAA workgroups on whether or not to include it. :rolleyes:

I believe the Jeppesen proposal was to add a "Top Altitude" column but still within the departure routing text section.



Thanks,
I did notice that the chart example in the report showed an FAA chart while it was mentioned in another portion of the report that the crew used Jepps which did seem strange.

[Capn Bloggs]

What does the Jepp chart look like?

Setting FL310 on the MCP altitude window was not in accordance with the operator’s procedure
What is the company's procedure?

How big are the EFBs - why is zooming in necessary?

The Flight Crew Training Manual (FCTM) on standard radio communications phraseology requires pilots to articulate both passing altitude and assigned altitude in their initial contact...
This is standard RT practice anyway, without the FCTM, should have been learnt at basic flight school.

[xyze]

Quote:
The Flight Crew Training Manual (FCTM) on standard radio communications phraseology requires pilots to articulate both passing altitude and assigned altitude in their initial contact..

This is standard RT practice anyway, without the FCTM, should have been learnt at basic flight school.


Not with this phraseology, at least not when the SID has a published top altitude. See page 5 of:

https://www.google.ca/url?sa=t&source=web&rct=j&url=http://www.faa.gov/about/office_org/headquarters_offices/avs/offices/afs/afs400/afs470/pbn/media/rnav1_rnp1_rnav2/climb_descend_via_faq.pdf&ved=0ahUKEwiVscL6vbLJAhUM6iYKHTuTA8IQFggjMAM&usg=AFQjCNHoZC0J_hxUFpcUudCIwfdeW_EUKw&sig2=lHcPCbGJ0aA9X3MR9Wx9Og

Which is why the crew's error was not detected by ATC until they busted the published top altitude.

Seems that this sort of error is inevitable if the crew are not familiar with the phraseology AND don't read the chart thoroughly. That it doesn't happen more frequently begs the question 'why this particular crew?' (no slander intended).

[Capn Bloggs]

I knew I'd shoot myself in the foot. I originally wrote "isn't this standard R/T?" but changed it! :) :{

Anyway, I agree. On the face of it, pretty dumb not stating the level you are climbing to... The procedure is probably OK for the locals that are familiar, but the "occasional" international driver? Set up yet again. I'm not letting the crew off the hook; a few big stuffups on their part, but... Perhaps that contributed to the incident being waved off as a non-event by the American authorities... How close do they have to get before the investigations start??

The investigation should have picked up the lack of the "climbing to..." requirement; it seems the PM actually did what they were supposed to do: not announce "climbing to 4000ft".

The report said they didn't follow SOPs WRT the Altitude Select:

Such action is not in accordance with the operator’s Flight Crew Operations Manual (FCOM) on “Flight Deck Preparation – Before Start” checklist. The checklist called for the assigned or SID limit altitude to be set in altitude window on the MCP.
But the crew didn't have (because they didn't see) the SID limit of 4000ft, so FL310 was the best thing to comply with the SOP...

[captainsmiffy]

Since the first stop off altitude is so important, why do we relegate it to a a small box on a SID chart and assume that, because it is there, the crew will have read it? What on earth is so wrong about giving the first assigned level/altitude along with the SID designator when the crew checks in with ATC for their clearance? Eg "you are cleared the........SID, maintain 4000'." clear and unambiguous and no chance of it being missed. KISS, methinks.

[philbky]

Whatever the rights and wrongs of the way information is displayed on physical and electronic charts, isn't it the case that there was a distinct lack of common sense displayed by an experienced crew? Even if this was the first time all of them had ever visited IAH, it must have occurred to them that in a busy TMA, with two major and many minor airfields on their charts, plus the almost continuous chatter on the frequencies when inbound a couple of days before and as they departed, a climb to cruise altitude with no restriction would be unusual, if not impossible, especially at that time of day. This isn't thinking outside the box, just common sense self preservation.

[zonoma]

Everyone is focusing on the charts & the crew, ATC aren't blameless either. The controller should have checked the cleared level immediately after the crew failed to report it, especially with only 1'500ft to go with traffic descending on top. Nothing mentioned about Mode S, do the Houston radars show it? If not then this case must highlight how safety could be vastly improved with it.

[Uplinker]

Without seeing the actual chart, it is hard to see why this crew got it wrong.

However, in my opinion, something as important as the cleared altitude/level should be clearly and unambiguously shown. In the AAIB report, the box from the plate shown has the cleared altitude of the SID tucked away in the explanation of the routing - in the same font, and font size as the rest of the narrative. The box above is titled initial climb, the box below is titled routing. The departure is unusual in that the actual first cleared altitude is 600', and then the SID cleared altitude is written in a box titled routing. I think this is ambiguous, and as a pilot I think I would expect to find all reference to cleared altitudes in that box - in bold type - and not in the routing box.

Pilots in modern commercial flight operations need such important things to be very clear, and they should not expect to have to read every word of a chart to find something as basic and as safety critical as this. Some pilots prefer the visual representations, others prefer the text explanation. On all the Jepps and AERADS I have ever used, the first cleared altitude/level is either shown very clearly on the pictorial diagram of the routing at the appropriate point, or clearly written as a bold heading above the departure explanation box.

This is something that we make sure we brief for every departure - "what do you make the first cleared altitude?",and then we confirm that that altitude is set on the FCU and on the PFD.

So, on the face of it, two reasons for this loss of separation: Poor plate layout and design, and poor pilot briefing. This was then compounded by an incorrect TCAS RA reaction.

[wiggy]

However, in my opinion, something as important as the first cleared altitude/level should be clearly and unambiguously shown. In the AAIB report, the box from the plate shown has the first cleared altitude tucked away in the explanation of the routing - in the same font, and font size as the rest of the narrative.


Agreed, certainly in the early days of our use of electronic charts (not that long ago) there were SIDs where a hard height was hidden away exactly as you describe. Fortunately that has now been changed

Despite that I'd like to think that, as philbky has said, the thought of a straight climb to cruise level right off the ground in the States might have rung a few alarm bells with somebody, but then again I wasn't there.

[xyze]

Capn Bloggs

The problem is even more diabolical than you suggest. As the crew obviously did NOT know the phraseology, the PM's R/T WAS incorrect as he should done as you suggested and stated that they were planning to climb to FL310 until told otherwise by ATC, which clearly would have sounded odd so he omitted this. ATC on the other hand heard what they expected to from a crew familiar with the procedure. That ATC did not query the lack of a planned altitude likely confirmed in the crew's minds that what they were doing was ok. From this point on there was no way of detecting the error until the top altitude bust.

captainsmiffy

Agree with your thoughts completely and yet this whole issue (i.e. does the new phraseology leave a little bit too much unsaid between ATC and flight crews?) was not mentioned in the discussion of the report. Ironic considering the AAIB took it on 'to highlight safety lessons' the NTSB thought were not there to learn!

[Capn Bloggs]

The controller should have checked the cleared level immediately after the crew failed to report it
No. Have a look at page 5 of the link provided by xyze:

Climb Via/Descend Via Speed Clearances Frequently Asked Questions (https://www.google.ca/url?sa=t&source=web&rct=j&url=http://www.faa.gov/about/office_org/headquarters_offices/avs/offices/afs/afs400/afs470/pbn/media/rnav1_rnp1_rnav2/climb_descend_via_faq.pdf&ved=0ahUKEwiVscL6vbLJAhUM6iYKHTuTA 8IQFggjMAM&usg=AFQjCNHoZC0J_hxUFpcUudCIwfdeW_EUKw&sig2=lHcPC bGJ0aA9X3MR9Wx9Og)

The pilots are not supposed to report the top level.

(Disclaimer: that document may be out of date).

[suninmyeyes]

I am a little puzzled.

SID heights restrictions are normally programmed into the FMS automatically when the SID is selected. Departures are normally flown in VNAV. Fortunately I have never tried it but if you were to make that same error say out of Heathrow and put FL310 in the alt sel the aircraft would level off automatically at the SID altitude of 6000. Even if hand flying the flight directors would level you at 6000. So my question is did they push the alt sel to clear the initial level off altitude or do IAH SIDS not have level off altitudes programmed into the FMS? If the latter is the case then why not as it would prevent this exact scenario.

[JammedStab]

I am a little puzzled.

SID heights restrictions are normally programmed into the FMS automatically when the SID is selected. Departures are normally flown in VNAV. Fortunately I have never tried it but if you were to make that same error say out of Heathrow and put FL310 in the alt sel the aircraft would level off automatically at the SID altitude of 6000. Even if hand flying the flight directors would level you at 6000. So my question is did they push the alt sel to clear the initial level off altitude or do IAH SIDS not have level off altitudes programmed into the FMS? If the latter is the case then why not as it would prevent this exact scenario.

Heathrow's altitude restrictions for their SID's are all at specific waypoints. The one used in Houston is not related to a specific fix. Do all the SID's with altitude restrictions not specific to a fix such as the one in Houston have the actual altitude restriction in the FMC?

[WillowRun 6-3]

The comments posted about the report and incident, including those which take a stab at stating larger questions raised by the incident in a general sense, impel me to ask:
1) If it is accepted as fact that the NTSB gave the incident too little emphasis - given the several potential extra-holey Swiss cheese scenarios implicated - do those knowledgeable about such investigatory safety boards think that the NTSB's broad mandate which includes railroads and pipelines is a root cause of its limited attention span? As the forum community includes a pretty extensive non-U.S. contingent, and other countries appear to have safety agencies for aviation as such . . .
2) I had always thought that take-off, climb-out and ascent toward cruise Flight Levels - while immensely mysterious to such as an SLF - were less challenging, to the aviators, than executing a successful and safe approach, at least in many instances. But here, the front-end of the operation was out of alignment . . . is it correct that a bust of a cleared altitude on ascent toward cruise Flight Level is not typical, even highly atypical?
What all the factors involved in this incident say about the need to change the nuts-and-bolts of R/T and charts and ATC procedures, I don't know. What these factors say about perceptions about a need to tighten up on training standards and procedures of air carriers, I also don't know. But hoping those who do know what they're talking about might comment further.
WR 6-3