Aftermath: Failure to Focus on the Overall Situation

"The entire normal operating checklist for the new Embraer Phenom 100 light business jet fits on both sides of a laminated card that you can slide into a shirt pocket," former Flying editor Mac McClellan wrote approvingly in his 2009 report on the airplane. "The Phenom is designed to cut pilot workload to a minimum so there is time to focus on the overall situation while the airplane and its systems take care of themselves."

Unfortunately, neither the Phenom nor any other airplane can ensure that the pilot actually does focus on the overall situation.

The overall situation presenting itself to the pilot of a Phenom 100 approaching Montgomery County Airpark near Gaithersburg, Maryland, in December 2014 was a tricky one. While there was ample visibility and a 2,800-foot ceiling at the airport, he would have to approach it through a cloud layer with a possibility of icing. Using anti-ice during the approach, however, would require a much higher approach speed (vref), and that would result in a landing distance almost exactly equal to the runway length.

The IFR flight, with two passengers aboard, had originated at the pilot’s home field of Chapel Hill, North Carolina, less than an hour earlier. As the jet leveled out at FL 230 in cloud, the pilot turned on anti-ice for the engines (which used compressor bleed air to heat the intakes) and, with a separate switch, for the wings and horizontal stabilizer (which used pneumatic boots to remove ice after it formed). Two and a half minutes later, presumably after emerging into the clear, the pilot switched anti-ice off and did not use it for the remainder of the flight.

The Phenom’s quick-reference handbook (QRH) provides a simple rule for using anti-ice: If the total temperature — which includes the temperature rise due to ram effect — is below 10 degrees C and visible moisture is present, engine anti-ice should be turned on. If the total temperature is below 5 degrees C, or at the first sign of ice accretion, wing and stabilizer boots should be used as well.

Before leaving FL 230, but after hearing the automated weather broadcast from Montgomery County, the pilot set vref for the approach. The value he selected, 92 knots, implied two things. First, he underestimated the airplane’s weight by about 400 pounds. Second, he forgot, or chose to ignore, that descending through clouds at the prevailing temperatures would require using anti-ice.

The discrepancy due to weight was small: 3 knots. The discrepancy due to the chance of icing was much larger. According to the QRH, vref should have been not 92 nor 95 but 121 plus 5 knots. The landing distance — defined as the distance from 50 feet above the threshold to a full stop — would then be 4,117 feet. The runway length at Gaithersburg is 4,202 feet — about the same, incidentally, as that of the pilot’s home airport at Chapel Hill.

The Phenom tracked the ILS on autopilot, with the human pilot controlling only power. Its airspeed diminished, gradually and somewhat irregularly, from 130 knots intercepting the glidepath, through the selected vref of 92 knots, to a minimum of 88 knots. The gradual deceleration was due to the pilot's keeping the power setting slightly too low so that the autopilot had to continually bleed off speed to maintain the calculated RNAV glidepath.

The airplane had emerged from clouds, and the pilot and the passenger occupying the copilot’s seat were discussing the fact that the runway was in sight, when the indicated airspeed reached 92 knots. The pilot added power but too late. The angle of attack continued to increase, and the wing began to stall. The airplane rocked from side to side with increasing violence — witnesses described “uncontrolled S turns” — and then rolled semi-inverted and crashed into a group of houses 4,000 feet short of the runway. The jet’s three occupants perished in the crash, along with a mother and two children in one of the houses.

The National Transportation Safety Board concluded that ice collected during the descent through clouds must have caused the wing, which would have been expected to stall at 73 kias, to stall at the much higher speed. The atypically large difference between the Phenom 100’s recommended vref with and without anti-ice suggests that the stall characteristics of its wing are strongly affected by ice accretion and possibly by operation of the pneumatic deice boots. The mere presence of a stick shaker suggests that the airplane's stalling behavior is not benign.

The NTSB pondered the question of why the pilot did not turn icing protection on for this approach. It suggested three explanations.

First, he may have been concerned about landing distance. The margin of safety with anti-ice turned on — 85 feet — was scarcely reassuring, and it happened that, several years earlier, he had experienced a runway excursion incident in a TBM 700 at this very airport. “He was likely highly motivated to avoid this scenario,” the NTSB wrote. If he left anti-ice off, on the other hand, he would be able to use a lower approach speed and have a much larger stopping margin.

Second, he may have become “task saturated” during the approach and forgotten to activate anti-ice. This hypothesis appears inconsistent, however, with the fact that the pilot selected a no-anti-ice vref during a period of very low workload, long before the approach began. The NTSB chose to overlook that discrepancy, instead enlarging on the theme of forgetfulness to suggest that single pilots of turbine aircraft need something more than broad environmental criteria to remind them to use anti-ice.

Third, the pilot perhaps “did not accurately assess the effect of icing conditions on airplane performance.” The NTSB noted that the pilot, 66, a 4,700-hour ATP, was “new” to the Phenom 100 — although he had flown it 136 hours, including at least 70 hours of flight instruction and a corresponding amount of ground instruction. The NTSB was concerned that neither the airplane’s documentation nor training curricula may have sufficiently emphasized that “not using the airplane’s wing and horizontal stabilizer deice system, with even small amounts of ice accretion, could result in extremely hazardous low-speed handling characteristics.” While the NTSB did not say so, this explanation logically implied the first — that the omission of anti-ice had been deliberate.

Intention being impossible to prove, the NTSB contented itself with finding simply that the pilot’s failure to use anti-ice resulted in a premature aerodynamic stall. It did not mention another factor that some might see as relevant. A certification standard that requires an increase of nearly 30 knots in vref when anti-ice is used, combined with a broad, nonempirical definition of when anti-ice is needed, renders a large number of normal runways unusable and creates a strong incentive for pilots to improvise. This may be a “deficiency in pilot performance” — or maybe a larger problem waiting to be solved.