UPS Cargo Jet Crashes Near Birmingham Shuttlesworth International Airport

[Alessandro]

not that it makes a difference, but i think the FO was a woman. 37 y/o.

True and the left seat was a man 58 y/o.

[3WE]

Please explain me how do you use your altimeter and VSI to judge your glide slope both in angle (2.5°, 3°, 3.5°) and in position (where you'll hit if you keep this path).

There are some indirect ways, but none of them is very straightforward and very accurate.

In a visual approach, your key inputs are the airspeed indicator and the sight out there (especially the VASI or PAPI if available).

Have those two nailed, and you don't need to worry about the altitude or the vertical speed. Not that you won't include them in your scan pattern.

Now the real question would be... Were they in a visual approach really? Calling "runway in sight" 4 seconds before impact well shy of the runway?

I don't thing we have an unability to arrest the descent as in "we tried but couldn't", but rather as in "we couldn't even try because the first news that we were really should have arrested the descent the sound of impact".

Regarding the "sink rate" warning, that's a point. But my main concern is what were they doing that low, 4 seconds before impact, I suspect well below minimums, with the runway NOT yet in sight (and hence the PAPI not in sight either).

Birmingham, AL has significant topography, it was oh-dark-thirty in the morning and there was light rain.

Although, the visibility is reported as good, in that environment (all three make for perfect conditions), random clouds and fog banks often occur, especially on slopes!

And, such things may not be reported by the automated weather machinery!

Now, all of that being said- if they were in some sort of IMC, then they are in an even more over the top, extremely unstabilized situation.

I've watched too many crop dusters and fighter pilots to know that in visual conditions, you can come come sweeping, diving, slowing and achieve a nice landing (while violating all semblances of a stabilized approach). (And choose your favorite jumbo jet and do it on MSFS- the only "stabilized approach gate" that truly matters is between 0 and 20 feet AGL, and the first few thousand feet of runway!)

Conversely, to follow an instrument approach- even a non precision one- You want awfully precise control of the aircraft- unless the weather is gentle.

[Evan]

Feel free to correct my math, but as I figure it, assuming you are aiming for a point 500ft down the runway, on a 3.24° glide path you want to be just above 300' above runway LDZ elevation at the point of impact.

The terrain where the first point of impact is placed (.8nm from threshold, about 5360' from the glidepath touchdown point) adds about 150ft elevation. Add a 75ft tree to that and it becomes about 225' so you are only about 75' clear of treetops ON the proper glidepath.

Fly it at 3° and that becomes only about 50'.

At 300' altitude, each dot deviation below subtracts 15-20'?

Does that seem about right?

[3WE]

Feel free to correct my math, but as I figure it, assuming you are aiming for a point 500ft down the runway, on a 3.24° glide path you want to be just above 300' above runway LDZ elevation at the point of impact.

The terrain where the first point of impact is placed (.8nm from threshold, about 5360' from the glidepath touchdown point) adds about 150ft elevation. Add a 75ft tree to that and you are at about 225' so you are only about 75' clear of treetops ON the proper glidepath.

Fly it at 3° and that becomes only about 50'.

At 300' altitude, each dot deviation below subtracts 15-20'?

Does that seem about right?

Figure that while flying 150 miles per hour with 2 minutes left to land, and don't forget to have a mental picture of the topography below you...and don't forget, since topography varies, we need about five of these distance-altiude reference points.

...and don't forget to monitor your airspeed and be sure the autothrottles are set right, and follow the localizer/R-nav inputs, and vertical speed and engine instruments, and contact the tower, and acknlowledge the landing clearance, and prep for a go-around- and the missed approach procedure if you exceed stabilized approach criteria.

By way, if it's a true LOC approach you already have some fairly random distance-altitude step downs to worry about and remember.

Dot's? No dots on this approach...

Sorry, I don't want my pilot worrying about trig (or handy trig short cuts), or previously figured distance-altitude points on short final with four big, bright and beautiful approach slope lights right out the window.

[Evan]

Figure that while flying 150 miles per hour with 2 minutes left to land, and don't forget to have a mental picture of the topography below you...and don't forget, since topography varies, we need about five of these distance-altiude reference points.

...and don't forget to monitor your airspeed and be sure the autothrottles are set right, and follow the localizer/R-nav inputs, and vertical speed and engine instruments, and contact the tower, and acknlowledge the landing clearance, and prep for a go-around- and the missed approach procedure if you exceed stabilized approach criteria.

By way, if it's a true LOC approach you already have some fairly random distance-altitude step downs to worry about and remember.

Dot's? No dots on this approach...

Sorry, I don't want my pilot worrying about trig (or handy trig short cuts), or previously figured distance-altitude points on short final with four big, bright and beautiful approach slope lights right out the window.

Wow, it's almost like you need some kind of co-pilot.

But I'm suggesting that they would have figured this. I'm just trying to get an idea of the margin of error on this approach (I used dots as a recognizable standard deviation). It doesn't seem very forgiving.

[N-ONE]

There has been a bit of chat about PAPI and visual slope indication, however, this “ LOC only” runway is a “Non-Precision” approach. A non-precision approach for the A300 would have been initially done with the autopilot, and set-up for either a selected or managed approach.
For the non-industry folks:

Selected approach; the pilots tell the autopilot how to fly by adjusting individual knobs on the glare shield for Speed, Heading, Altitude, and Vertical Speed (V/S).

Managed approach: the Flight Management Computer (FMS) tells the autopilot how to fly the approach as long as that particular approach for that runway is in the FMS database, including non-precision approaches.

Chances are that UPS does not carry this particular non-precision approach in its FMS database as it appears that it is not a runway they normally use.

A selected non-precision approach has the crew change their visual references on the Primary Flight Display from the normal pitch and roll bars to the Flight Path Vector / Flight Path Reference or better known as Bird in the Cage (bird on the Perch for you A320/330/340 drivers). The vertical position of the cage is adjusted by a separate dial in the EFIS control that can also be used for setting the Decision Height.

The Vertical Speed window is a two digit display on the A300 unlike the later Airbus’s where it is a four digit display as modified after the unfortunate Air Inter crash several years ago (ref the link below).

http://en.wikipedia.org/wiki/Air_Inter_Flight_148

At the final approach fix they are to set an appropriate V/S and adjust the cage for a 3 degree approach angle.

What if the wrong knob was selected to “30” setting a descent rate of 3000 ft/min instead?

Sink Rate, I can hear it.

[Evan]

What if the wrong knob was selected to “30” setting a descent rate of 3000 ft/min instead?

The PFD would show VS as '-30'?

[Gabriel]

Feel free to correct my math, but as I figure it, assuming you are aiming for a point 500ft down the runway, on a 3.24° glide path you want to be just above 300' above runway LDZ elevation at the point of impact.

The terrain where the first point of impact is placed (.8nm from threshold, about 5360' from the glidepath touchdown point) adds about 150ft elevation. Add a 75ft tree to that and it becomes about 225' so you are only about 75' clear of treetops ON the proper glidepath.

Fly it at 3° and that becomes only about 50'.

Well, that was my thinking, without doing the maths, after playing with Google Earth and its simulator.

At 300' altitude, each dot deviation below subtracts 15-20'?

Does that seem about right?

No, no, NO!

Please stop treating this as an ILS approach. THere was no ILS available, functional, installed, or planned to be installed in the foreseeable future.

Let me quote myself:

If (again: IF) they were using an FMS generated glide slope, those ones don't follow the conic geometry of the ILS.

[N-ONE]

Evan,
Agreed, and there would have been a bunch of other indications as well, Rad Alt, the inertial trajectory (bird ) dropped out of the box, EGPWS Sink Rate, any of these lines of defense should have prevented this.

Human Factors, fatigue and a dark and stormy night. I wish I had those kind of answers.

IMHO

[Gabriel]

Was there a PAPI or VASI for RWY 18 indeed?
Look at the lower-left box in the approach charts below. It doesn't look like there was.

Can someone please explain those approach charts?
What are the different MDAs?
What are the numbers in parenthesis?
What does "radar required" mean and the several times "radar" is written along the vertical profile? Does it mean "radio altimeter"?

http://155.178.201.160/d-tpp/1308/00050L18.PDF

http://155.178.201.160/d-tpp/1308/00050R18.PDF

[Evan]

No, no, NO!

Please stop treating this as an ILS approach. THere was no ILS available, functional, installed, or planned to be installed in the foreseeable future.

Yes, Gabriel, I realize that there was no GS. I am merely using dots as a unit reference to illustrate margin of error. For example, Airbus allows up to one dot variation in a stabilized approach. Convert that to whatever a dot represents (15-20' @ 300' I think...). I am only indicating a standard amount of deviation, not implying that they had ILS.

[Gabriel]

Yes, Gabriel, I realize that there was no GS. I am merely using dots as a unit reference to illustrate margin of error. For example, Airbus allows up to one dot variation in a stabilized approach. Convert that to whatever a dot represents (15-20' @ 300' I think...). I am only indicating a standard amount of deviation, not implying that they had ILS.

But you can't expect the same precision, standard deviation, margin of error, and safety margins, on a precision and a non-precision approach. That's why:
one has the name "precision" and the other "non-precison", and
one has minimums in the order of the 200ft and the other more like 500.

How is a pilot supposed to keep a +/- 1 dot precision in an approach with no vertical guidance?

What keeps you safe in non-precision approaches is NOT busting the minimums of each segment (step-down, dive'n drive).

If you have an on-board equipment that can generate a virtual glide slope, then it's safer than the step-down method AS LONG AS:
the slope doesn't bust the minimums of each segment, and
you respect the MDA until you have the runway in sight.

[Evan]

But you can't expect the same precision, standard deviation, margin of error, and safety margins, on a precision and a non-precision approach.

Exactly. You're not following me here. I am trying to determine how far from a precise 3.24° glidepath an aircraft can be (because there is no precision guidance) without striking the trees at the reported impact site. Forget dots. Call it feet. A two-dot deviation represents about 30-40' here, and a quarter degree variation on glidepath costs you another 20' or so. If you are 75' clear of the trees at a precise 3.24° then you are just skimming trees with those plausible deviations. IF my math is anywhere near correct...

Because of the terrain, I think maybe it's a runway in dire need of ILS.

[Gabriel]

Well, this info is more updated than the approach charts. Maybe they isntalled a PAPI when they had to close the main runway (06-24)

http://www.airnav.com/airport/KBHM

RUNWAY 18
Latitude: 33-34.392837N
Longitude: 086-44.831402W
Elevation: 644.3 ft.
Gradient: 0.2% UP
Traffic pattern: left
Runway heading: 183 magnetic, 180 true
Markings: nonprecision, in good condition
Visual slope indicator: 4-light PAPI on left (3.20 degrees glide path)
Runway end identifier lights: yes
Touchdown point: yes, no lights
Instrument approach: LOC/DME
Obstructions: 16 ft. gnd, 615 ft. from runway, 26:1 slope to clear

Relevant note: 26:1 = 2.2°
However, at least in Google Earth the point where I've taken the "photo" (see below), which is about 3800ft from the runway, seems to impose a higher gradient at some 2.9°.

[Gabriel]

Exactly. You're not following me here. I am trying to determine how far from a precise 3.24° glidepath an aircraft can be (because there is no precision guidance) without striking the trees at the reported impact site. Forget dots. Call it feet. A two-dot deviation represents about 30-40' here, and a quarter degree variation on glidepath costs you another 20' or so. If you are 75' clear of the trees at a precise 3.24° then you are just skimming trees with those plausible deviations. IF my math is anywhere near correct...

Because of the terrain, I think maybe it's a runway in dire need of ILS.

Forget it.

75' from an obstacle (especially an invisible one like a tree at night) is unacceptable (well, I guess, hope, it is), no matter ILS or what.