air safety

Mar 2, 2017

SingaporeAir flew too close to mountains near Canberra last month

Singapore Airlines loses track of altitude near Canberra while the ATSB gets lost in its own backyard

Ben Sandilands — Editor of Plane Talking

Ben Sandilands

Editor of Plane Talking

[caption id="attachment_62023" align="aligncenter" width="610"] An old bushwalking photo of Bimberi Peak, Wikipedia[/caption] Updated Had you been atop Bimberi Peak (1912 metres or 6273 feet above mean sea level) on February 22 this year you might just have noticed a  270 or so seater Singapore Airlines 777-200 slicing through the air not far above or to one side of the wind blasted summit on its way to an otherwise uneventful landing at Canberra airport. The  Boeing should not have been there.  At that point of its descent toward Canberra it should not have been below its designated safe minimum altitude of 7500 feet but it was at about 6600 feet. Having been apparently advised of this by 'the tower' to use a good old fashioned phrase, flight SQ 291 climbed back to 7500 feet for the short period before it had cleared the beautiful but very high and often snowy hills and continued on its descent to its landing at the national capital at the first stop on its journey from Singapore to Wellington, New Zealand. The ATSB has this week launched an inquiry into the procedural incident, as it calls it on its notification page. If the media had relied at that official notification it would have ignored it, because it contained almost zero relevant information, and what it did publish was wrong and no doubt inadvertently misleading. In fact this totally but no doubt accidentally useless notification did result in no major general media coverage (as of this hour) apart from a good story in Singapore's Straits Times. The ATSB said the 'procedural incident' occurred 20 kilometres south of the airport when in fact it took place 33.3 kilometres to the southwest.  The difference between distance and direction turns a pretty boring and gently undulating part of the Monaro plains into a cluster of eroded very, very post glacial granite tops that are higher than the tree line in places and are often described geologically as the northernmost peaks and ridge lines of the Snowy Mountains. Updated March 4:  As pointed out by the comment posted by reader 'Jukebox'  the ATSB notification begins to look deliberately misleading when the exact coordinates of the 20 kms south reference on its notification page are expanded and placed over a map. The ATSB has so far declined to respond to questions about how it arrived at that location for the incident. [caption id="attachment_62025" align="aligncenter" width="555"] Compare this with graphic further down page[/caption] How do we know the ATSB notification was inadvertently wrong to the point of rendering the procedural incident un-newsworthy?  Because of the Aviation Herald, which nailed it as shown in the graphic below and which overlays the available data on the approach path over a Google Earth map. [caption id="attachment_62024" align="aligncenter" width="610"] Aviation Herald plot of descent path[/caption] It's one thing for an airliner to lose its altitude awareness, and not a very good thing either when it involves big hills, and another for a safety investigator to fail its general knowledge test of terrain which is only a short distance from its own offices in Canberra. PS. And Yes, we did seek clarification from the ATSB yesterday and No, there was no reply. PPS. It is unclear how close SQ291 came to Bimberi, or other nearby high peaks, but it came too close, which is what caused the ATSB inquiry.

Free Trial

Proudly annoying those in power since 2000.

Sign up for a FREE 21-day trial to keep reading and get the best of Crikey straight to your inbox

By starting a free trial, you agree to accept Crikey’s terms and conditions


Leave a comment

15 thoughts on “SingaporeAir flew too close to mountains near Canberra last month

  1. comet

    The Singapore 777-200 might have been only 227 feet above the terrain.

    That’s 69 metres. Marginally more than the length of a swimming pool.

    1. Mick Gilbert

      Comet, I can’t see how you’re getting such a close call. According to the AV Herald track, SQ291 was at around 9,500 passing over Mt Bimberi, a clearance of some 3,000 feet. Where it temporarily breach its LSALT it was at 6,600 feet over terrain of about 4,600 feet, a clearance of some 2,000 feet. Its next closest approach to high terrain was as it passed over the Rob Roy Reserve at an altitude of 6,300 feet, a clearance of around 2,300 feet.

  2. 40years

    I suspect that the distance error rose from confusion between miles, nautical miles and kilometres. The original incident report from approach to ATSB probably said 20 miles, which is the unit that is used. A transcriber in ATSB probably made the error.

  3. Ben Sandilands

    Agree that is a plausible explanation for the distance error, but not the directional error, which totally changes the terrain risks.

  4. comet

    Yet as far as we know, it didn’t set off any terrain warnings aboard the aircraft.

    “Pull up, Pull up…”

  5. PeterP.Syd

    To fly below LSALT in your current sector is an inexcusable navigational error (although many years ago as a new CIR pilot, I made the same mistake myself). It was and probably still is, Canberra Approach’s practice to assign a lower altitude than area/sector LSALT once positively identified on radar, and they have an alternate LSALT for that assigned track. That track LSALT can be below area/sector LSALT. Its possible that Approach may have made an error, not the flight crew. Around CB, if assigned an altitude below area LSALT, my practice was to question the assigned altitude.

  6. Rais

    That report in the usually quite reserved Straits Times is surprisingly frank. Singapore readers got information that readers of our daily press didn’t get.

  7. JW (aka James Wilson)

    Is there a bit too much willy-waving going on here? The ATSB says the incident occurred 20km S, the Aviation Herald says 18nm SSW and Ben says 33.3km SW, but does it really matter? It obviously matters in terms of deciding if the aircraft was likely to collide with terra firma, but does it matter in the context of an incident summary that is only intended to provide an approximate description of the location where an incident occurred? Plenty of ATSB summaries only say that an incident or accident occurred ‘near’ some place or other, without specifying a precise direction or distance. Surely the phrase ‘descent below lowest safe altitude’ should be enough to pique the interest of anyone who cares.

    The only thing that really matters is finding out why the aircraft descended below the minimum safe altitude (MSA). The MSA to the south of Canberra is 7,500 ft, however, the STAR steps the aircraft down initially to 7,000 ft until waypoint HONEY, then to 6,000 ft and finally to 5,400 ft to start the approach. It seems likely the aircraft descended towards 6,000 ft before passing HONEY, as a Qantas B737 did back in 2012. The ATSB found that incident was caused by Qantas procedures. Something similar here perhaps?

    1. Yes No Maybe

      Was the aircraft on a visual approach?

  8. comet

    I’m surprised nobody has invented an electronic guidance system on board aircraft that holds a 3D image of the terrain, and warns the pilot if he/she is doing something that might result in an impact.

    Something more sophisticated than the last-minute “pull up” terrain warning.

    Satellites have generated 3D plots of all the terrain on earth. Events like this Singapore Airlines error should never happen.

    1. JW (aka James Wilson)


      Most modern airliners (including the B777) are already equipped with such systems. The same Ground Proximity Warning System (GPWS) that provides the “pull up” terrain warning also provides look-ahead terrain alerts. The GPWS features a terrain data base with detailed terrain data near most major airports and lesser detail for areas between airports. Terrain within 2,000 ft of the aircraft’s current altitude is displayed on the pilots’ navigation displays in different colours, depending on the height of the terrain in relation to the aircraft altitude. I don’t know about SQ, but it is common practice for one pilot to have the terrain information displayed on the ND during the latter stages of descent and approach.

  9. Tango

    Agreed with James, its not the lcoaiton of the bust, its that it happened at all.
    The whole system is intened to keep clear of terrain and its the violation that’s the issue. The terrain can then become the ufnrovien, but tergian is irreleave as long as the aproahc is not comproiskiedm

    On the other hand I did not know Austril had such hansome moutnains.

    Even by Alaska standards those are credible. Himalaya maybe not, but out the East side of our house is a range of mountains that many are 7500 ft high (with a lot of snow on the right now)

    As we are at sea level they are impressive bumps.

    And I ddi not know Austrial had real moutnians!

    1. PeterP.Syd

      Europeans still refer to our mountains as “hills” 🙂

  10. jukebox

    Ben – when you open the ATSB page, and click on “Show Map” , then zoom in, it gives very specific co-ordinates – 35°29’24.0″S 149°12’00.0″E – which lies East of the Monaro Highway. That may be worth adding to your post, to clarify your point that the location being alluded to by them is not consistent with the description (and therefore prompts the question: where does that ATSB quoted specific location come from?).

  11. Rufus

    As a layman who has a vague (and doubtless misguided) familiarity to the procedure after many hours in the right hand seat of a light aircraft while my mate has been vectored for ILS approaches, I’m wondering how this happens. I thought every stage of the descent right until you intercept the localiser is controlled? Was he cleared too low, too soon – an ATC error? Or, once cleared for a Razzi 3 arrival (or whatever) is he free to descend in accordance with the approach plates (which he might have got wrong)?

    And on a different note, will a modern jet alert the pilots if they’re deviating from one of the standard approach plates? Could he have been on a Polli 5 arrival?

Share this article with a friend

Just fill out the fields below and we'll send your friend a link to this article along with a message from you.

Your details

Your friend's details