Dangerous Aerobatics : C-17 crash at Elmendorf (2010)
In 2010, a US Air Force transport plane crashed just one minute after take-off while practising for the Arctic Thunder Air Show at Elmendorf Air Force Base in Alaska.
The Boeing C-17 Globemaster III, a large military transport aircraft, was being flown by a highly qualified commander and three crew members, all of whom died on impact. The aircraft, valued at 184,570,581 US dollars, was destroyed.
The final report of the USAF Aircraft Accident Investigation Board Report doesn’t appear to be available online, although it seems the report was released to the public, as FlightGlobal and other aviation magazines are clearly based on the investigation results.
I have been emailed a copy of the report and I’m interested in doing an analysis here, because it is an interesting combination of human factors.
The fatal accident (mishap in military parlance) took place in a C-17 Globemaster III on the 28th of July 2010 at the Joint Base Elmendorf-Richardson (JBER).
Elmendorf Air Force Base is the largest military installation in Alaska, with 13,000 square acres housing over 10,000 people (6,000 military personal from US and Canadian armed forces and their family members) with another 15,000 civilian workers, retirees and their families living on site. There are three elementary schools, 150 miles of roads and two runways. It’s pretty clear that the airshow, Arctic Thunder, is the highlight of the year.
The C-17 Globemaster III, nicknamed Spirit of the Aleutians was owned by the 3rd Wing, which moved to Elmendorf in 1991 and became the host unit. The 3rd Wing is responsible for Alaska’s air defense and the support of the Pacific Air Forces. The C-17 was also operated by the 176th Wing, the Alaska Air National Guard. It is a large military transport aircraft developed by McDonald Douglas and built by Boeing for the US Air Force. The C-17 is popular for its ability to rapidly deliver troops and cargo to deployment areas, as well as being able to perform tactical airlift and airdrop missions. It is 174 feet long (53 metres) with a wingspan of almost 170 feet and a maximum take-off weight of 585,000 pounds (265,000 kilos). The C-17 is powered by four fully reversible Pratt & Whitney PW2040 engines, each producing 40,440 pounds of thrust. It cruises at 450 knots.
The required crew is a pilot, a co-pilot and a loadmaster. On that day, the crew was made up of three Air National Guard members (the commander, the load master and the safety observer) and one active duty member (the co-pilot, which I will refer to as the first officer). The commander was the Pilot Flying and sat in the left front seat. The first officer was the Pilot Monitoring and sat in the right front seat. The safety observer was in the right additional-crew-member seat, which gave him visibility of most of the flight deck displays and switches. The load master sat in the right-rear area of the cargo compartment.
The ‘mishap sortie’ to use the military term, was a practice flight for the Arctic Thunder Airshow, which would serve as an aerial demonstration proficiency and currency flight. An aerial demonstration flight involves the aircraft going through a series of practice demonstration manoeuvres (profiles). There are four documented profiles and, on that day, the crew planned to fly Profile 3, which is known as the 12-minute Profile.
The 12-minute profile starts with the following manoeuvres:
- A maximum performance climb to 1,500 feet above ground level,
- An 80/260° reversal turn,
- A high-speed pass at 500-foot above ground level
The maximum performance climb demonstrates the climb-capability of the aircraft using minimum control speed, in this case VMCO. V-speeds are used to define airspeeds of an aircraft; they work as a place-holder, where the actual speed is based on the aircraft, the model, the configuration and the indicated airspeed. VMCO is the speed required for the aircraft to clear an obstacle if only three of the four engines were operating.
Once the C-17 is 1,500 feet above the ground, the reversal turn can be initiated. This is an 80° turn away from the initial heading, which establishes the outbound leg. The pilot should then fly a safe distance away from the runway before performing the 260° reversal turn back towards the runway. The C-17 descends from 1,500 feet to 500 feet above the ground and then accelerates to 250 knots to fly past the spectators at the centre of the viewing area at high speed.
The commander, the pilot monitoring and the safety officer had practised several aerial profiles in the sim the day before, including the 12-minute profile 3.
That afternoon, the flight crew used Operational Risk Management to evaluate the mission risk. This is a decision-making process which evaluates possible courses of action and identifies the risks of benefits for each. The Operational Risk Management category for the planned display was in the “Caution” range based on the complex and demanding nature of the manoeuvres. All crew members agreed that it was safe and prepared for the flight.
The aircraft commander briefed the practice flight, which was planned to stay within 20 nautical miles of the airfield. The flight crew reviewed NOTAMS, the weather forecast and other relevant issues. Then the commander filed the VFR flight plan for the practice flight. The aircraft had just completed a test flight after unscheduled maintenance, so the flight crew did an Engine Running Crew Change, in which the pilot and co-pilot positions are swapped out one at a time, so that there is always a qualified pilot at the controls. The crew departing confirmed that the aircraft had no malfunctions and the flight crew took control of the aircraft.
The commander, as pilot flying, first flew around the local area to confirm that the wind and flight conditions were acceptable for their practice run. The weather was agreed to be within limits. They landed and waited half an hour to be able to start their sortie.
At 18:22 local time with all four on board (Pilot Flying, Pilot Monitoring, Safety Officer and Load Officer), the commander initiated a maximum power take-off with a nose-high attitude of 40°. All four engines were set to maximum thrust (92.5%) and remained at maximum thrust throughout the flight.
The aircraft levelled off at 850 feet above ground level.
Here we already see two major issues within the first ten seconds of flight.
The steep climb with a 40° nose-high attitude meant that the aircraft never achieved the mandatory minimum climb-out airspeed. The C-17’s peak airspeed in the climb was 107 knots; the target climbout airspeed was 133 knots.
Second and more obviously, the commander levelled off at 850 feet above ground level, barely half the minimum altitude of 1,500 feet required for the manoeuvre.
They were low and slow.
The commander started the 80/260° reversal turn as they passed through 800 feet above ground level. He banked left 57° for the outbound leg at a heading of 340° and then levelled off at 852 feet above ground level. The US Air Force’s “prescribed bank limit” is 45°, known as a “steep turn”.
As they levelled off, the airspeed reached 151 knots. The minimum flap retraction speed (Vmfr) was 150 knots, so the Pilot Monitoring silently retracted the flaps.
The C-17 continued outbound for seven seconds while the flaps retracted. Then the commander initiated an aggressive right turn, banking right by 60° for the turn back towards the runway.
As they reached 188 knots, the first officer initiated the slat retraction, although the minimum slat retraction speed (Vmsr) was actually 193.
Five seconds into the right turn, the stall warning sounded. The C-17’s airspeed was 199 knots, six knots below stall speed. The stick shaker triggered while the stall warning repeated STALL STALL STALL over the intercom.
The Pilot Monitoring called out, “Temperature, altitude, lookin’ good!”
The commander continued the turn with full right rudder and constant control stick pressure. The angle of bank increased to 62°. Now the aircraft was not only beyond the 45° angle of bank recommended for the steep turn, they had exceeded the maximum allowable bank angle for the C-17, which is 60°. The commander continued to apply control stick pressure, increasing the g-force on the aircraft to 2.4 g.
The aircraft had ALS, an Angle of Attack Limiter System, which was meant to protect the aircraft by stopping the aircraft from attaining AoA attitudes which could result in a deep stall. The ALS limits commanded nose-up elevator positions. However the commander’s manoeuvres were so rapid and aggressive, the ALS was unable to protect the aircraft.
The C-17 went into a deep stall. As the airspeed decayed to 184 knots, they began to descend rapidly, as fast as 9,000 feet per minute. The Pilot Monitoring, still not aware of the gravity of the situation, said, “Not so tight, brother,” while the safety officer said “Watch your bank.”
He repeated it twice more, “Watch your bank,” and the commander moved the control stick in the opposite direction, full left. However, at the same time, he applied full left rudder, which increased the stall. He also continued to maintain constant control stick pressure.
The aircraft went into a minimal left roll. Two seconds later, the C-17 crashed into a wooded area. One minute had elapsed since take-off. The aircraft exploded into a fireball hundreds of feet into the air.
You can watch the tragic flight on this video taken by a spectator at the runway.
Emergency vehicles were dispatched immediately but when they reached the area, they found they were unable to approach, with debris and fire scattered over a large area. The captain of the Anchorage Fire Department watched a fireball rise 750 feet (230 m) into the air. It was already clear that there was no chance of survival.
All four crew members onboard the C-17 were killed on impact. The Safety Investigation Board was immediately notified and the team began its investigation on the 2nd of August.
Next week, I’ll go over the results of the investigation which attempted to make sense of why qualified crew allowed the aircraft to go into a stall with apparently no attempt at recovery until it was much too late.
I do not remember ever coming across a speed “VMCO”. New to me ! After take-off, with an engine inoperative, we used “V2” which should be calculated to give adequate climb gradient to clear obstacles, then accelerate to “VFTO” to clean up and either return or proceed to a take-off alternate. Speeds would be calculated according with aircraft type, actual weight, runway, weather (wind and temperature) and elevation.
All these factors would have been pre-calculated and kept on board (TL-tables) to assist the crew in their pre-flight preparations. A card with essential information (V-speeds, flap setting, runway, SID and other data, format depending on the operator) would be clipped to the control column or displayed prominently as quick reference for the crew. But of course, display flying requires different parameters.
I thought I’d replied to this, sorry! VMCO seems to be specific to the C-17.
There’s a Vmco for the E-2 Hawkeye as well. Many multi engine aircraft have have Vmco and Vmcg (minimum controllable airspeed on the ground)
Reading this I just wonder: Had it been possible to do an autopsy on the body of the pilot flying? His actions and reactions seem out of tune with those of any experienced pilot. Was he in the process of becoming incapacitated? And very strange that the other crew members acted so passively. They were not inexperienced themselves. Was the hierarchy in the air force so strict that nobody felt that they could take over?
I didn’t see any evidence that they tried to turn the rudder left to correct and regain lift, just before the crash. Unfortunately even at full screen viewing I’m not sure I’m seeing the control surfaces correctly. But to me it sure looks like the never tried to pull out of the turn.
Does anyone else see it that way?
Of course this raises some very serious questions, similar to how Rudy is thinking.
The video cuts off a few seconds before the crash; I’d want to correlate time stamps on the video and whatever produced the report of left rudder) before assuming the video has all the data.
Rudy — either the formal hierarchy or established personal hierarchy between the pilots may have been an obstacle. I’m also surprised that the PM didn’t say anything at the very first turn (at half the specified altitude) but that may be part of the hierarchy. OTOH, the report sounds like they were both out of it, with the PM reporting altitude as OK when it clearly wasn’t and retracting the slats when aircraft wasn’t up to speed.
The report says the pilot had been flying a modified version of profile 3 for some time (stall warning and all). I expect Sylvia’s going to go into more detail on this in the followup post, but that would explain why the crew was used to it.
I wonder which version they flew on the simulator while preparing for this.
C-17 is a fly by wire system. The pilot puts his request into the system he computer tries to figure the best way to accomplish it. It could be that at such a deep stall the computer knew the best thing was to break the AOA. But it was too little too late.
Something definitely not correct here!!!! I agree. This ALS system cut him out I believe. He lost all roll and pitch control.
You can find the report online at Wikimedia Commons and elsewhere. https://commons.m.wikimedia.org/wiki/File:2010_Alaska_USAF_C-17_crash_report.pdf
Thanks — I hadn’t spotted this, I’ll add it.
Mendel does make a point, but it still does not explain why the pilots initiated steep turns and retracting lift augmenting devices, still below anything even resembling safe speed AND at low altitude.
Anyone remembers the following:
Nothing is as useless as the airspace above you,
the runway behind you and
the fuel you did not put in the tanks ?
Maybe we could add:
the airspeed you did not each !
The steep turns were part of the modified routine with the aim of keeping the plane closer to the spectators. A wider turn would have put the plane at a greater distance from the crowd (and so would a longer climb).
Retracting the slats early would reduce drag and allow the plane to pick up more speed?
I think they might just have been a little slower than usual that day, possibly they had more fuel on board (from the test flight) than they were used to?
Rudy — the obvious guess is that he did this (and had been doing it for some time) because it made for a snappier show. The Blue Angels and the Thunderbirds have partly-separate planes that can fill in between formation passes; a solo C-17 does not, and might even have gotten complaints from ignorant show announcers about the gap between takeoff and low pass.
The report that Mendel links to is appalling, but the most appalling thing about it is what’s missing. This trimmer was not just a demo pilot but an instructor of demo pilots; he clearly misled them about what was safe, but there’s no recommendation that all his students be given new check rides and monitored to make sure they don’t follow his bad habits, let alone that instructors have regular check rides (as this induhvidual did not) to make sure that they’re not teaching bad habits. I expect Sylvia will be quoting the details of this misbehavior in her next column; I wonder whether there’s some other report with recommendations.
Seems like a classic case of turning too steeply at too low an altitude. Similar to the B52 incident at Fairchild airbase. I wonder, had any of them realised the danger in time, could the leading edge slats have been deployed fast enough to have saved their skins? It might seem callous but it seems to me that unusually, three of them contributed to their demise. Only the load master had no part in it!
Our AAIB would describe this as uncontrolled flight into terrain. Spinning in as the result of banking too steeply at too low an altitude and too low a speed, (often after engine failure) is probably the cause of more pilot deaths than anything else. In fact I suspect that a survey might find it to be the cause of more deaths than all other causes put together! Sadly far too many pilots seem to think they are immune from the effect of gravity!
I’m a 3 axis microlight pilot, not a commercial pilot. When I trained 8 years ago, the scariest thing I did by far was to induce (with an instructor) an incipient spin resulting from the combination of excessive bank and insufficient speed. It was drummed in to me in such a situation SET NEUTRAL AILERON, APPLY OPPOSITE RUDDER AND PITCH DOWN. ANYTHING ELSE WILL KILL YOU!
I note that on this flight, in a severe right bank when already stalling, “the commander moved the control stick in the opposite direction, full left.” Where is the evidence of basic training? Perhaps in the American military it is not considered necessary these days?
Sorry to bang on about this but for sheer incomprehensible stupidity (or ignorance) this makes my hackles rise!!!
Ouch, a painful lesson learned too late by the crew…
A note: Acres are a square measurement, so saying “square acres” is redundant – I’m assuming you are going off a press release and didn’t come up with that on your own!
Newbie pilot here who loves reading accident reports.
Shouldn’t there be mandated callouts by the PNF or annunciator callouts if the turn exceeds either the recommended max bank angle by the organization (in the case the military) or the maximum bank angle allowed by design?
Stevedo wrote: and he’s correct, left rudder and nose down but at that elevation that would amd could only happen for a few seconds to gain speed and stabilize. Ailerons then need apply to level and then slowly climb out. As in any flying, you get into trouble slowly, usually and must pull out of it a bit quicker and steadily. As in a lost engine, at take off w out enough altitude or airspeed, first reaction is NOSE DOWN or stall, spin and crash… Nose down, light in the seat and fly the plane and look for a area of a controlled landing Into Terrain. CAOA kills w the added feature of lack of air speed.