Cause & Circumstance: Accelerated Stall At Low Altitude Ends In Crash

An accelerated stall is one in which the G-load on the airplane is greater than +1G. One place an accelerated stall commonly occurs is in an overbanked turn to final approach. Excessive banking has been a factor in many approach-to-landing accidents, including three between 2017-21 involving business jets. The most recent of these took place at Gillespie Airport in El Cajon, California, on Dec. 27, 2021.

The crew of a Learjet 35A, N880Z, was attempting to land at their home base in nighttime instrument flight rules (IFR) conditions, and they were conducting a circling approach. They made a decision early in the flight that they would be able to make it in, and they never reconsidered that decision despite worsening conditions as the flight progressed. Their final act of overloading the airplane in a low-altitude turn should be understood not just as an isolated action, but as the result of a series of poor judgments and procedural violations.

The repositioning flight departed from Orange County Airport (SNA) at 1856 PST. They were only 68 nm (78 mi.) from Gillespie Field (SEE) in El Cajon, California, their destination. The crew had previously transported a patient to SNA from Lake Havasu City Airport (HII), and there were two flight nurses onboard.

The co-pilot was flying. The flight was cleared to climb to 11,000 ft. Just 4 min. after departure, the pilot-in-command (PIC) had the RNAV (GPS) 17 approach to SEE loaded and was listening to the current ATIS for Gillespie. Information Golf reported variable winds and mist with a 2,000-ft. broken cloud deck and a 2,600-ft. ceiling. Visibility was 2 to 3 mi., and they were landing and departing Runway 27R.

The crew set their landing reference speed bugs at 124 kt. and 134 kt. and ran the approach checklist. SoCal Approach then cleared them to cross ASIXTY intersection at or above 5,000 ft. and for the RNAV Runway 17 approach.

The PIC briefed the approach even though the co-pilot was flying. “Okay so inbound course is gonna be one forty seven we’re gonna inter-intercept at JUGAL at twenty five hundred feet ... our M-D-A is gonna be thirteen sixty. Airport elevation we know what that is three seventy nine. Missed approach fifteen hundred and three thousand ... our temperatures they’re * — heat’s goin’ back on.”

The pilots never intended to land on Runway 17. At 4,145 ft. long and with only 3,695 ft. available beyond the threshold, that runway was too short to safely land their aircraft. When the PIC mentioned that the Runway17 PAPI was out of service, the co-pilot said, “well we don’t care—we’re goin’ to two seven.”

At 1908, just 12 min. after they took off, the PIC contacted Gillespie Tower. The tower cleared them to land on Runway 17. The co-pilot then said to the PIC, “’kay we’ll have to cancel and then make a left … right?” He agreed.

Circling was not authorized at night or anytime northeast of Runways 17 and 27R, and night had already fallen. The only remaining option was to fly a visual flight rules (VFR) traffic pattern. The Jeppesen 20-9A approach information said the Runway 27R traffic pattern altitude from dusk to dawn was 1,588 ft., 1,200 ft. above the ground. But that is not what the PIC set in the altitude alerter.

JUGAL was the final approach fix (FAF), 6.6 nm from the Runway 17 threshold. Passing JUGAL with flaps 20 and gear down, the co-pilot called for the altitude window to be set to 1,360 ft., but the PIC replied, “nah, I’m gonna watch it that’s your minimums I’m gonna set fifteen hundred which is your missed.” Then he said, “actually I’ll do fourt—,” then “we do we go down to thirteen sixty so we got about three hundred feet to minimums.”

Just before arriving at the 1,360-ft. minimum, the PIC saw the surface and said “I got it—there it is.” The co-pilot still saw nothing, but instead of remaining at circling minimums, 1,440 ft., or traffic pattern altitude, 1,588 ft., he continued to descend. When the PIC reported the airport in sight to the tower and asked to squawk VFR, the airplane was at 950 ft., only 562 ft. above the airport elevation.

After the tower cleared the flight to land on Runway 17, the PIC requested 27, and the tower replied “Learjet eight zero zulu I-F-R cancellation received you can overfly the field and make left traffic runway two seven right ...***... change to runway two seven right cleared to land.”

The airplane was still north of the airport. About 1 mi. out from the field, the PIC asked the tower to turn the runway lights up, but they were already at 100%. They were at 700 ft., only 312 ft. above the ground, and their ground speed was 128 kt. Neither pilot had said anything about their passing altitudes or their intended level off altitude. Finally, the PIC said, “level off—level off.”

They were very low and very slow, as though they were feeling their way along in the dark. Just before crossing the airport, the co-pilot saw it and said he would head right down the runway. The ADS-B flight track shows the airplane flew to the left of the Runway 17 centerline, as though the co-pilot was looking down to keep it in sight. The airplane crossed the airport at 785 ft., 400 ft. above the ground, with an airspeed of 138 kt. The time was 1913.

The co-pilot was flying completely blind. He asked the PIC to tell him when to turn left when passing the airport, and again when he was on downwind. The airplane descended to 700 ft., then climbed to 950 ft. in the turn to base leg. When the PIC spoke, the co-pilot said “yep I will” and “’kay.” As he turned in toward the runway, the copilot said he saw “the little mountain,” then said “woah woah woah speed speed.” They each said “go around the mountain,” then “this is dicey.”

The PIC took control of the airplane and said “watch my speed for me.” The NTSB’s performance report showed he was steepening his bank in an attempt to turn toward the runway. The co-pilot immediately began calling out “speed speed,” “more, more” and “faster, faster.” The last recorded ADS-B target was at 1914:09, at an altitude of 875 ft. and only 100 ft. from the accident site. A witness said the airplane made a “very hard steep turn to the left” and the wing was “basically pointing straight down to the ground” before it crashed.

The airplane struck power lines and impacted the yard of a residence about 1.43 nm east of the approach end of Runway 27R. There were no survivors.

The Investigation

The NTSB conducted an extensive investigation and published its final report about two and a half years after the accident. Representatives from Bombardier, Honeywell, Aeromedevac (the operator), NATCA (the controller’s union) and the FAA participated. An NTSB team launched to the accident site.

A crater at the point of impact contained evidence that the airplane’s descent had been steep. The nose cone, instrument panel and windscreen were in the crater. The right wing was adjacent to the crater, and other wreckage was spread beyond the crater to a distance of 186 ft. Engine RPM indicators and airspeed indicators were recovered, but were impact-damaged. The thrust reversers were closed and locked, and flight control continuity from cockpit to airframe was established. Evidence at the scene showed both engines had been running.

When the airframe was examined at a remote site in June 2022, the ground proximity warning ( ) and both digital electronic engine controls (DEEC) were found, but no flight management system (FMS) could be located. The recovered instruments were found to be too damaged to produce useful information when they were later examined at Honeywell’s labs.

The airplane had a stall warning system, including stall warning lights, stick shaker and stick pusher. The investigation did not determine if it functioned.

The airplane was equipped with a Fairchild A-100A 30-min. cockpit voice recorder (CVR). The magnetic tape was recovered and transcribed at the NTSB’s labs by a CVR group. While three of the four channels produced good-quality recordings, the cockpit area microphone (CAM) was only fair. That channel is normally most useful for recording engine sounds, cockpit alarms and alerts and control movements. The CVR specialist reported “numerous cockpit alerts” near the end of the flight, but few of them were documented.

There were many witnesses to the accident, and their statements were gathered by email. Several witnesses noted the airplane was in a steep left bank. One heard two unusual noises which may have been compressor stalls. A person who lived in an elevated location north of the airplane’s flightpath said a heavy bank of fog had moved in after 1830 and the visibility at his location, about 800 ft. above sea level, was virtually zero. He saw a bright flash when the airplane crashed. A person in the parking lot of the Home Depot on Fletcher Avenue, just south of the airport, reported there was a hard rain there as he heard the airplane fly overhead. He saw a “blue pop” in the sky about 15 sec. later.

Since the airplane was not equipped with a flight data recorder (FDR) and no onboard flight data survived, the NTSB performance specialist relied on ADS-B data from the FAA to produce a depiction of the aircraft’s flightpath. He was able to estimate or calculate airspeed, rate of climb, angle of attack and bank during the final 4 min. of flight. When the airplane stalled, the angle of attack approached 20 deg.

Using the Learjet 35A Airplane Flight Manual (AFM), he determined that the +1G stall speed of the airplane at its estimated weight and 40-deg. flap setting was 96 kt. The airplane’s airspeed at 1914:04, just before the loss of control, was 119 kt.

The AFM data showed that the stall speed increased to 119 kt. in a 50-deg. level banked turn and to 134 kt. in a 60-deg. level bank. Based on his estimation of the bank angle exceeding 60 deg., the performance engineer found that the airplane experienced an accelerated stall.

Gillespie Field is terrain-constrained for approach category C and D aircraft. There is a 1,624-ft. hill southwest of the field that allows only category A and B approaches from the west, and a 1,273-ft. hill east of the field that is close enough to the Runway 27R final approach course that only circling minima are provided for the localizer approach to that runway. The hill to the east, known as Rattlesnake Mountain, was the “little mountain” the co-pilot was referring to just before the airplane stalled. The flight was 400 ft. below the crest of that hill at the time, and its proximity may have been what prompted the PIC to enter the steep turn.

An NTSB meteorologist found that although the METAR issued at 1855 still reported the airport visibility to be 3 mi., the visibility fell below 3 mi. at 1901, and the cloud decks began to lower. The scattered clouds at 1,400 ft. above the ground became broken at 1901, and the visibility dropped to 2 mi. at 1903. The visibility remained below the VFR minimum of 3 mi. for the remainder of the flight. Some clouds were reported as low at 700 ft. above the ground.

The NTSB found that the visibility had gone below VFR visibility minimums 7 min. before the controller accepted the pilot’s IFR cancellation. The current AWOS visibility was displayed in the tower, but the display was located in the back of the cab, and it was not equipped with an aural warning that would sound when conditions dropped below VFR minima. The controller said he would not have accepted the cancellation if he had known the visibility was less than 3 mi.

The controller said he was able to maintain visual contact with the airplane as it overflew the airport, but lost sight of it before it crashed. He said there was nothing remarkable about the flight. He “did not remember” the airplane’s altitude as it passed and could not remember if it was raining.

The two pilots had been employed at Aeromedevac for about 2.5 years. The 42-year-old PIC upgraded to captain on the Lear in July 2021, five months before the accident. He reported 2,200 total flight hours to the FAA on his last Class 1 medical certificate application, also five months before the accident. He had an ATP with a LR-Jet type rating. The co-pilot, who was 67 years old, had a commercial certificate with LR-Jet SIC privileges and a Class 2 medical certificate with a limitation to use corrective lenses. He had 1,244 hr. total flight time.

Investigators did not obtain the two pilots’ Learjet or turbine flight experience or their previous flying histories. The director of operations said all of the company’s pilots went through simulator training at FlightSafety in Tucson, Arizona. Both pilots’ last proficiency checks took place on Oct. 11, 2021, about two and a half months before the accident. The chief pilot said the co-pilot was rated unsatisfactory on that last simulator check due to flying an unacceptable departure procedure. He was approved on his second attempt.

The company had no safety reporting system or safety management system (SMS). The director of operations said he relied on a lot of personal one-on-one training with all the pilots.

Conclusions

The first idea to be disposed of is that this crew was flying a VFR traffic pattern. A look at the flight path shows they were making a close-in circling approach, even though doing so was prohibited. The second is that the tower controller was not aware of the hazard that the Learjet posed when it crossed the field at low altitude despite his approval of their request to fly a VFR traffic pattern.

The Safety Board mostly acknowledged these truths in their probable cause statement. In that statement, they said the flight crew descended below the MDA, conducted an unauthorized circle-to-land approach in night IFR and exceeded the airplane’s critical angle of attack. They said the tower crew contributed to the accident, in part, by not monitoring the AWOS display in the cab. What they did not discuss was the tower controller’s indifference to the airplane’s low-altitude, low-speed pass across the field in dark, foggy conditions.

It is true that air traffic control’s (ATC) job is to provide separation between aircraft, but no responsible controller just ignores patently unsafe actions by pilots. When the controller lost sight of the airplane after it turned east, he could have checked the AWOS and warned the LearJet crew of the declining visibility and prepared to issue them a new clearance. One reason he might not have done this is that he was accustomed to seeing such pilot actions in marginal conditions, as the circle or VFR pattern were the main ways for LearJets to get to Runway 27R. A frank conversation between the ATC chief at Gillespie and the Aeromedevac DO might have prevented this accident. An SMS at both operations might have helped to enable that conversation.

At the top of the list of safety issues to me was the PIC’s apparent lack of understanding of how swept-wing jet airplanes react to accelerated stalls. They just fall out of the sky. Unlike small trainers and many general aviation straight-wing airplanes, the Learjet and similar airplanes do not just recover quickly. Once they are headed down, they need a lot of altitude to recover. A big push over is necessary, and there is no room for that in the traffic pattern. You cannot let yourself fly into that box.

Related to this is the difference between a level turn and a descending turn. If you are descending on base leg at Vref plus 10 or 20, you can afford to use a little back pressure to line up with the runway. When you are at 300 ft. 2 mi. from the runway, you cannot. You have to make a level turn, and the G load increases. That is one of the hazards of flying a too-low traffic pattern.

This behavior of high-performance airplanes is hard to demonstrate. Airline simulators have now been set up to demonstrate stalls, but not all business jet simulators have. One thing the FlightSafety Tucson director of training said in an interview prompts me to comment. He said “the Buffalo crash” prompted a shift to more scenario-based training. While that may be true, there is a more important lesson about stalls to be learned from the Colgan 3407 accident. It is that you must not pull back—that is, load the airplane—when you are operating close to the 1G stall speed of the airplane.

That is what the Colgan captain did. That is what the Challenger crew did at Truckee, and that is what the Learjet crew did at Teterboro.

As more pilots, young and old, are drawn into the industry and upward to high-performance airplanes, their understanding of those high-performance airplanes must improve. One good place to start is the Airplane Flying Handbook (FAA-H-8083-3C), Chapter 16, “Transition to Jet-Powered Airplanes,” and in particular the discussion of stalls.