What Can We Learn from an Accident?
Accident Recovery: The Perils of Success
By David Jack Kenny
The tension between managing weather hazards and responding to urgent need is intrinsic to both search-and-rescue (SAR) and helicopter air ambulance operations. Lost hikers, downed aircraft, and bloody auto crashes — the very weather that makes these accidents more likely will also impede the first responders trying to reach them.
While there are missions weather doesn’t complicate — a fallen rock-climber on a sunny spring afternoon, or an interhospital transfer above rush-hour gridlock — operators have to anticipate that when they’re needed most may be when they’re least able to respond. Standardized procedures for settling the most difficult decisions before dispatch can’t anticipate every contingency encountered in the field, and a long run of brilliant rescues may even blur the true location of the risk-benefit balance. Such was the case in the crash of an experienced helicopter pilot during a 2013 rescue mission in Alaska.
In Good Hands
At 7:35 p.m. Alaska Daylight Time on March 30, 2013, the MatCom dispatch center in Wasilla received a 911 call from a snowmobiler whose vehicle was stuck in a ditch under the InterTie high-voltage lines between Larson Lake and Talkeetna. He had suffered bruised ribs, was unable to free his machine, and was at risk of hypothermia if not rescued quickly.
The Alaska state trooper on duty at Talkeetna tried to coordinate a ground expedition, but could not. No Alaska wildlife troopers were on duty nearby, and local civilians with snowmobiles and SAR experience were put off by the combination of long distance, poor snow conditions, and deteriorating weather, including rain showers.
At 8:09 p.m., the trooper contacted his SAR coordinator to request dispatch of the Alaska Department of Public Safety’s (DPS) primary search-and-rescue helicopter: N911AA, a Eurocopter AS350 B3 operated under the call sign HELO-1.
The coordinator reached the pilot on call within 10 minutes. After checking the weather, the pilot called back a few minutes later to accept the flight. When his wife asked, he told her the weather was “good” and left for Ted Stevens Anchorage International Airport (PANC). At 8:51 p.m., he called the Signature fixed-base operator to request help pulling N911AA from its hangar. At 9:17 p.m. — just over two hours after the initial 911 call — he radioed dispatch that he was airborne.
The mission could hardly have been placed in better hands. In more than 12 years of flying for DPS, the pilot had received three commendations for exceptional performance during rescue operations. His personnel file contained half a dozen letters of gratitude from people he’d saved or their relatives, and his annual performance evaluations uniformly rated him “outstanding” or “high acceptable.”
Six months before, he’d rescued three people stranded on a gravel bar in a flooding river after an Air National Guard HH-60 Pave Hawk had to turn back for weather too low to traverse a mountain pass. The DPS pilot stayed up all night monitoring weather radar until he spotted a break, launching at 3:00 a.m. He followed a lower route to avoid the pass and succeeded in extricating all three.
His wife and colleagues attested to the great satisfaction these rescues brought him. However, his superiors, pilot colleagues and observers, and tactical flight officers who’d flown with him also agreed that he wasn’t a risk-taker. He flew “by the book” and could be “completely trusted.” His relief pilot, who’d flown with him frequently, described him as “a sound professional” and “the best helicopter pilot” with whom he’d ever flown.
To the Scene
At 9:42 p.m., the pilot advised dispatch that he was approaching Sunshine, a landing zone near the Talkeetna barracks, where he boarded a state trooper qualified to serve as flight observer. They lifted again almost immediately, found the snowmobiler a few minutes later, and landed on a frozen lake less than one-quarter mile away. By that time, the victim was severely hypothermic and too weak to walk.
It took the pilot and observer about an hour to reach him and get him back to the helicopter. At 11:16 p.m., the pilot radioed dispatch that they were returning to Sunshine and asked to have an ambulance meet them. That was the last transmission received from HELO-1.
The outbound flight from Sunshine to the landing zone had taken 11 minutes. Beginning at 12:39 a.m. on April 1, almost an hour-and-a-half after their reported departure, the dispatcher made more than 30 attempts to reach either the helicopter or the snowmobiler, alternating between radio and telephone.
A ground search was launched at 2:30 a.m., by which time the helicopter’s fuel supply would have been exhausted had it remained airborne. An Alaska National Guard Pave Hawk joined the search at 7:00 a.m. At about 9:30 a.m., its crew located the accident site roughly 3 miles south of HELO-1’s point of departure from the frozen lake.
Investigators found N911AA lying inverted and almost entirely consumed by a postcrash fire. All of the wreckage was located at the scene. The debris path was only 75 feet, and although the crash site was forested, only one tree was broken, indicating a near-vertical descent. Impact and fire damage were severe, but to the extent evaluation was possible there was no evidence of any mechanical failure before the crash.
In earlier decades, these circumstances — a remote location with neither witnesses nor survivors — might have made the cause of the accident impossible to determine. But data recovered from three pieces of equipment on HELO-1 allowed investigators to construct a detailed picture of the aircraft’s final moments.
An installed Garmin 430 and a handheld Garmin 296 GPS logged the aircraft’s airspeed, ground track, and altitude. An Appareo Systems Vision 1000 cockpit recorder captured images of the windshield and part of the instrument panel; the lip light on the pilot’s helmet indicated his head movements. The helicopter also contained an internal attitude heading and reference system that tracked acceleration in all three axes as well as pitch, yaw, and roll.
As he had on the outbound flight, the pilot took off using his night-vision goggles (NVGs) and with the turn coordinator’s circuit breaker pulled (his habit of disabling the turn coordinator was known but not explained). No blowing snow was visible, but nothing could be seen outside after liftoff.
The helicopter initially climbed to about 250 feet above ground level (AGL) at a groundspeed of 60 knots but descended to 100 feet and briefly slowed to 20 knots as it maneuvered around hills, presumably beneath low ceilings. About three minutes into the flight, it slowed again and began a gradual climb as the lip light shifted from purposeful alternation between the flight instruments and windows to a focus on the Garmin 296. A moment later, it returned to the panel as the helicopter began climbing at 1,000 feet per minute but close to zero airspeed, simultaneously turning hard to the left.
A series of increasingly severe attitude excursions followed, with pitch attitudes of up to 30 degrees nose-up and 45-degree banks while the indicated airspeed remained zero. Forty-four seconds after initiating the climb, as the ship completed a 360-degree turn, the pilot caged the attitude indicator. Bank angles reached 85 degrees twice in the last minute before the recording ended.
An observer who’d teamed up with him on more than 300 flights told investigators that the pilot’s contingency plan for zero-visibility conditions was to climb and transition to instrument flight. Based on that information, archived weather radar images, and the similarity between the accident flight’s track and divergent oscillations documented in performance studies of helicopter pilots unexpectedly entering instrument meteorological conditions (IMC), the NTSB concluded that the pilot had been unable to maintain control after losing visual references.
With the turn coordinator already disabled, caging the attitude indicator — now giving increasingly extreme pitch and bank readings — left the aircraft’s fate wholly dependent on the hope of a return to visual conditions.
... And Why
In retrospect, the pilot’s escape plan seems more aspirational than practical. Although he held both a fixed-wing ATP certificate and an instrument-helicopter rating, almost all of his instrument time was in airplanes. He hadn’t flown in actual instrument conditions since before 2001, and his only helicopter time in actual IMC was 0.5 hours in 1986. While the AS350 did have basic attitude instruments, it lacked the stability augmentation or autopilot systems required for instrument flight rules (IFR) certification.
He did not contact Flight Service before accepting the flight. According to colleagues, his online weather check mostly likely relied on the area forecast and Talkeetna’s Terminal Aerodrome Forecast and current observations. Conditions at takeoff were consistent with Talkeetna’s forecast of broken layers at 1,000 and 1,800 feet AGL and more than 6 miles visibility in light rain. DPS imposed no weather minimums beyond statutory requirements, but the conditions met the pilot’s written personal minimums of 2 miles visibility and 500-foot ceilings for a visual flight rules (VFR) flight using NVG.
He didn’t seek an update before lifting off on the 16-mile return flight. During the hour they’d been on the ground, however, a line of showers had begun moving north from Palmer toward Talkeetna. Three witnesses in a 10-mile radius recalled that the rain began changing to heavy snow during the time the helicopter was on the ground. By 11:00 p.m., one said, it was “coming down like a son of a gun.” The fact that HELO-1’s initial altitude on the return leg was 500 feet lower than outbound suggests that ceilings had come down considerably in the interim.
The extremity of the snowmobiler’s condition isn’t known, but the NTSB concluded that the pilot’s powerful motivation to complete rescues and his impressive previous record likely figured into his decision to hazard a short flight under lowering ceilings. Once airborne, prospects for finding a suitable precautionary landing site were uncertain at best — and disappeared along with the visibility. Notable success can create perils of its own.
Success certainly beats failure, especially for rescue pilots, but it comes with hazards of its own. A history of overcoming challenges can make accurate evaluation of future missions’ risk/benefit balance more difficult. It may also become harder to remember that the weather you’ll deal with is the weather you get, not what was forecast. It pays to double-check conditions.
And once the patient’s safely on board the aircraft, waiting for the weather to pass may be a better option than immediate departure. Clouds and precipitation move, ground resources could be available ... and crashing doesn’t improve anybody’s prospects.
David Jack Kenny is a fixed-wing ATP with commercial privileges for helicopter. He also holds degrees in statistics from Stanford and The George Washington University. From 2008 through 2017, he served as the statistician for AOPA’s Air Safety Institute, where he authored eight editions of its Joseph T. Nall Report, multiple other research findings, and nearly 500 articles for popular audiences. He’d rather be flying.