Remembering a Christmas tragedy 50 years later

The 1960s heralded the dawn of the Jet Age for air travelers. And for regional carriers the step up to turbine power often took the form of the Convair CV-580, a conversion of the redoubtable twin-engine Convair series from P&W R-2800 piston engines to Allison 501 turboprops.

An early adopter was Allegheny Airlines, one of the storied regional operators plying the skies over the Northeast and upper Midwest during that decade, flying a route structure that encompassed some of the most challenging weather and terrain in America.

Allegheny ticket
A period ticket folder shows Allegheny’s obvious pride in their new “Jet Powered” airplanes.

All things being equal, the airline had a good reputation among aviation professionals and the flying public. But during the holiday season of 1968, in an isolated Pennsylvania community, Allegheny’s professionalism, safety culture and luck would abandon the airline to a sequence of events no fiction writer could invent. And the echo of those tragedies continues to resonate a half century later.

The Tragedies

Allegheny flight 736 to Washington, DC, originated in Detroit with intermediate stops in Erie, Bradford, and Harrisburg. The Christmas Eve departure from Detroit was delayed almost an hour by the late arrival of the airplane. But the Erie leg and turnaround had been routine as the CV-580 climbed into darkness for the short hop to Bradford. Fifteen minutes later, Erie Approach cleared flight 736 to the Bradford VOR in preparation for the runway 32 approach and relayed the Bradford Flight Service Station’s hourly weather observation: “ceiling broken two thousand one hundred, visibility one mile blowing snow… wind thirteen gusting twenty two.” The report would not be a cause for concern to this experienced crew. But they were unaware that conditions were rapidly changing.

At 20:06 the First Officer reported over the Bradford VOR outbound on the approach. Erie authorized a frequency change to the Bradford Flight Service Station and requested notification when the flight was on the ground. Moments after checking in with Bradford, the Flight Service specialist relayed a report of “poor” braking conditions on ice and packed snow given by a previous Convair crew. Flying outbound from the VOR both pilots discussed the landing conditions:

First Officer (FO) – “little bit of ground contact”

Captain – “Yeah, I saw the runway when we went over it”

FO – “I hope you see it when we get down to minimums”

Captain – “But that’s not the problem. It just means getting stopped.”

FO – “that’s it…. I don’t like, especially when they are reported poor”

Their task was to get the airplane down on the first 1000 feet of the 6500 foot runway to ensure there was adequate length to stop. By 20:09 they completed the procedure turn inbound and initiated a 600 foot-per-minute descent. During the next 30 seconds, the First Officer verbally briefed the missed approach, the step-down fix (2900 feet MSL) and the minimum descent altitude (2643 feet MSL, 500 feet AGL).

Two minutes prior to impact, they began configuring the Convair for landing. A minute later the First Officer told the Captain, “You’re about two and a half miles from the end of the runway.” Beyond that point, they stared through a shroud of blowing snow focused on making visual contact with the runway. The descent rate gradually increased to 1500 feet per minute. Ten seconds before impact, the Captain turned on the landing lights and began to extend the gear. Eight seconds later, the First Officer, staring into the reflected glare, said, “I don’t see a thing.”

As this was transpiring, the Flight Service specialist at Bradford recorded a special surface observation. It showed the ceiling had dropped to 800 feet and visibility to one-half mile. This was immediately radioed to flight 736 but there was no reply. As minutes passed both Flight Service and Erie Approach became concerned. An inbound flight from Jamestown, New York, was directed to hold over the Bradford VOR until the Flight Service specialist at Bradford and the controller at Erie Approach could sort things out.

Kinzua railroad bridge
Flight 736 impacted terrain just beyond the Kinzua railroad bridge.

Inside the shattered fuselage of flight 736, the surviving passengers hung upside down in pitch darkness while coming to the realization that the airplane had crashed. Those who were able to free themselves emerged into deep snow on a hellish night of 30-knot winds at 10 degrees F. Some of them managed to start a fire using Christmas packages, seat foam and anything else on the ground that would burn. And the able-bodied began pulling the injured they could reach out of the wreckage.

Flight 736 was nearly half an hour overdue when the call went out to responders. The State Police swung into action, but everything was working against them. Bradford and its surrounding communities are located in a remote corner of Western Pennsylvania abutting the Allegheny National Forrest. Clinics were either closed for Christmas or minimally staffed. Responders were occupied with family or at social functions. No one knew where the crash was, but every potential location was inaccessible. Still, with the information they had, calls went out to citizens who owned four-wheel drive vehicles and the response came to life.

Thirty-five minutes after the crash, the flight from Jamestown (Allegheny flight 734) remained in a holding pattern above Bradford. Erie Approach cleared it for the VOR approach and asked them to look for evidence of flight 736. On the inbound descent, the crew spotted the survivors’ fire and reported its distance and bearing form the Bradford VOR. Geographically, it was just north of the Kinzua railroad bridge, a local landmark.

A young physician was one of the first to get a call from the State Police. Still wearing a shirt and tie, he navigated his Jeep along a snowbound railroad grade toward Kinzua gorge. In time, a fire came into view through the streaming snow storm. It was surrounded by stunned survivors. The first arrivals tried to open access to the twisted wreckage using axes and crowbars but timely evacuation of the casualties might prove impossible. If ambulances could get in at all, they would need to back down the railroad grade one at a time (there was no room to turn around). To further complicate rescue, the remains of the Convair rested in an icy marsh, its crust easily broken by the weight of a person.

The State Police had made a crucial decision early on to ask each responder to contact anyone they knew who owned a snowmobile. Now these began arriving to pull the injured, hypothermic passengers to ambulances on toboggans. Ultimately they were taken to the community hospital in Bradford or a nearby clinic in Kane.

Rescuers toiled more than four hours to extract the living from the mangled fuselage. Frozen and exhausted, they were wrapping up when the doctor and another responder decided to reenter it for a final check. They heard faint cries coming from the badly crushed forward area near the wing. There, a teenage girl was pinned in her seat. The twisted seat was cut out of the airframe using flashlights and a hacksaw. Decades later, her rescuers would recall that as she was being taken away she had wished them “Merry Christmas.”

The crash site
The fatally injured passengers were recovered on Christmas Day. One survivor had apparently died from hypothermia.

The NTSB investigating team started arriving on Christmas Day to begin the forensic recovery effort. Thirteen days later, they would achieve the distinction of being the only investigation in history to arrive at the scene of a crash before the airplane!

The second accident aircraft was Allegheny flight 737. And it was a frightening mirror image of the Christmas crash of flight 736. In fact, 737 was Allegheny’s return flight from Washington, DC, to Detroit. The crew also was flying the Bradford VOR approach but in the opposite direction (to runway 14). On descent, flight 736 had initially impacted trees, one of which severed the starboard wing rolling the airplane inverted prior to ground impact. Flight 737 suffered a similar fate. More remarkable was the absence of a post-crash fire in both mishaps.

Allegheny 737 proceeded uneventfully from Harrisburg when, at 20:23, Erie Approach cleared them to descend to 4000 feet and then cleared them for the Bradford VOR approach to runway 32. The crew was given the current weather observation: “ceiling eight hundred overcast, visibility one and one half and light snow showers, wind one seventy degrees at ten [knots].”

They were instructed to contact the Bradford Flight Service Station. Calling Bradford, the crew reported the flight’s position as 10 miles (southeast) from the VOR. Then they requested a change in their approach clearance to runway 14 to accommodate the wind and runway conditions. The Flight Service specialist coordinated an amended clearance with Erie Approach and relayed it to the flight.

By then, they were just one minute from the Bradford VOR. The flight continued outbound from the VOR and started the published course reversal. By 20:32 the captain had radioed “procedure turn inbound” and started a 1000 feet-per-minute descent. That descent would continue down to 2,500 feet MSL – a mere 350 feet above the airport elevation and 800 feet below the published minimum altitude for that segment of the approach.

The flight briefly leveled off (for about eight seconds) then resumed the descent. At 20:39, an anxious Allegheny gate agent walked the short distance to the Flight Service office and inquired whether they had any further communication with flight 737.

On the evening of January sixth, five miles northwest of the Bradford airport, a group of friends was watching college basketball on television. Behind the set was a picture window overlooking a golf club fairway. The sharp crack of splitting trees drew their attention to a dark shape materializing out of the night and drifting across the window. In a split second, the remains of flight 737 slid to a stop on its back. The debris trail had passed within 100 yards of the house.

Local authorities must have been in disbelief when they received the call, but good site access, and recent experience, contributed to a timely and efficient response. As the witnesses to the crash struggled through deep snow, they encountered survivors emerging from gaps in the torn fuselage and led them to shelter in the golf course pro shop. There, they removed an interior door and pressed it into service as a stretcher to carry the more severely injured. A fourth friend, just arrived, used his station wagon to transport the most critical to the hospital in Bradford. And the rescue continued in good order as more resources arrived.

The investigation

Although these accidents were treated as separate occurrences by the NTSB, both investigations were essentially carried out by one team. It would need to find an explanation for how four experienced pilots could have deviated so fatally from published instrument procedures.

Convair cockpit
The CV-580 cockpit was well equipped by 1960s standards with weather radar, a flight director and an ad-hoc coat rack.

No defects in the airplanes’ mechanical systems or navigation equipment were uncovered that could have precipitated either mishap. But, was there any way such a profound loss of awareness could have affected all four pilots at a critical phase in the approach? The investigators didn’t think so.

So began an extensive engineering and flight test program to expose any hidden flaw in the CV-580 static system that could have delivered erroneous pressure signals to the cockpit altimeters. In the end, the Allegheny engineers and test pilots, and the NTSB investigators, only proved the original Convair certification engineers had done their job properly. And, one by one, other factors like the autopilot mode, VOR facility calibration, flight instrument error and crew fatigue were eliminated.

Ultimately, the investigation centered on human factors and Allegheny’s crew procedures. In that respect, the Christmas accident provided more information to work with. From the cockpit voice recorder they knew the crew was concerned about braking conditions on the runway and, consequently, not being too high at the touchdown zone. From the flight data recorder, they determined the copilot had misreported the airplane’s position to the (flying) captain as 2.5 miles from the runway when, in actuality, it was 2.5 miles (DME) from the VOR. The runway lay a mile farther beyond – a small error, but made at a crucial point in the approach.

Allegheny’s flight procedures required the pilot not flying to call out the altitude, airspeed and descent rate upon reaching 500 feet above the airport elevation. None of the calls was made by the first officer as flight 736 broke through the 500 foot level. Allegheny procedures would also require him to observe the outside conditions and call out ground references no later than 100 feet above the published Minimum Descent Altitude (MDA). But, on the Bradford VOR approach, the MDA was also 500 feet.

That would require the first officer to complete the two tasks simultaneously! A fact that was not lost on the NTSB who opined, “It is possible, therefore, that his attention was focused outside the cockpit in an attempt to comply with the former duty, with the result that heoverlooked the latter.” There is every reason to believe if the “500 feet” call had been made, the captain would have arrested the descent.

Bradford approach
The approach plate used by Flight 736’s crew. The Bradford VOR is situated 0.9 miles from the end of the runway. “A small but critical navigation error.”

In its final report, the NTSB gave the probable cause of the crash of flight 736 as: “[a descent] through the Minimum Descent Altitude and into obstructing terrain at a time when both flight crew members were looking outside the airplane.” Contributing factors were “minimal visual references available at night on the approaches to the Bradford Regional Airport” and “a small but critical navigational error during the later stages of the approach.” The rapid change in visibility was cited also.

Determining a cause for the January 6th crash of flight 737 would prove more challenging. The cockpit voice recorder revealed little about the crew’s activities. There was no approach briefing (a plate for the VOR Rwy 14 approach was located in the cockpit) and none of the required call outs were recorded. Similarly, the flight data recorder showed a steady descent through the MDA (interrupted by a brief level off) and into terrain. Both ceiling and visibility were well above minimums.

The NTSB examined 13 possible explanations. Ten were dismissed after analyzing the available flight data, engineering studies and human factors data. Of the three that remained no determination could be made as to which represented the accident’s probable cause: “(1) misreading of the altimeter by the captain, (2) a malfunction of the captain’s altimeter after completion of the instrument approach procedure turn, and (3) a misreading of the instrument approach chart.”

The Board cited an Air Force study showing that visual interpretation of a standard three-pointer altimeter is susceptible errors (in 1000 foot increments).

Afterwords

Within months, a new precision instrument landing system (ILS) approach was installed at Bradford. The airport had been eligible for it years earlier, but FAA funding shortfalls had delayed its implementation.

A number of the Safety Board’s recommendations went on to become cornerstones of a new methodology called “Crew Resource Management.”

Today, most scheduled airlines do not allow non-precision approaches. But for the rest of us, they remain a fact of life. Non-precision approaches won’t disappear in my lifetime and probably not in yours. For the record, I have been guilty of every single error and oversight committed by the Allegheny crews on those tragic flights. The only difference was my outcomes were better.

13 Comments

  • Interesting article. I’ve read about these two accidents before and am always dumbfounded by them. Although, I can understand the ‘hows’ of them both: Years ago, we used to do non-precision approaches all over the place as commuter airline pilots flying those 19-seat pencils around. We attempted three consecutive circling approaches off the NDB one night at Pitt-Greenville, N.C. before finally heading back to Raleigh. We spotted the runway very briefly a few times but were never in a good position to continue the descents below the circling MDA when we did. One pilot would do nothing, absolutely nothing but stare at both altimeters and squeal like a stuck pig if either one dropped below the MDA while the other guy was horsing the airplane around trying to get into a position to dive and drive. We were flying all over the countryside at 400 feet or so at night in the clouds and rain; tragedy could have struck anytime. But that’s the way it was back then before GPS approaches – especially before RNP precision. Glad we don’t do that kind of craziness anymore – especially with those really weak NDBs. No wonder we called them Now Dumb Blundering approaches…

  • In the same general time frame as the Bradford accidents a CV 580 airliner conducting a night VOR non presision approach into Benton Harbor Michigan prematurely descended through the minimum descent altitude on the final approach segment. The aircraft continued descent thru minimums until the landing made contact with the surface of the lake. A missed approach was conducted successfully with the aircraft landing safely. Upon inspection of the aircraft it was discovered that the horizontal stabilizer was damaged from the water spray when the main landing gear touched Lake Michigan.

  • Tom, all I can say is “good grief!” I had an old Beech-18 freight dog buddy of mine, Bill, get hired by one of those Lake Michigan area commuters a long time ago after they had had several crashes. They appointed him ‘chief pilot’ of hiring and training because he was a junk yard freight dog. One night he had a large group prospective young new-hires on board the YS-11 going through their initial observation portions of their preliminary evaluation rides – which consisted of two non-precision approaches as well as hold entries and radial interception and tracking. It was, in all fairness, a pretty demanding itenerary. After they landed and shut down Bill got up, went to the cabin, and went down the aisle slowly while pointing at about every third recruit telling them to “get off my airplane, get your things and don’t come back.” With old Bill you either ‘had it’ when you arrived or you didn’t. If you didn’t he didn’t want to see your face ever again. But that company never had another accident after that time.

  • Dec ’68-Jan ’69 was a real jinx time for airline travel in the US. In less than 7 weeks from early Dec through mid Jan, there were 7 fatal crashes on scheduled airlines in the US…Wein Consolidated & Pan Am in Alaska, the 2 Allegheneys at Bradford, North Central at Ord, SAS & United at Lax. In addition, a Pan Am JFK departure was lost with all hands in Venezuela & Ozark lost a DC-9 with no fatalities in Iowa. There were also 7 hijackings of US airlines to Cuba during this same period. I don’t believe there has been a worse period for air travel in America before or since.

  • Bob,

    Thanks for that bit of research. You almost had to live through that period to appreciate the level of safety we take for granted today.

    By the 70s the NTSB was pushing CRM hard. The FAA, on the other hand, had to find a way to regulate crew coordination without eroding command authority and responsibility. Success was far from guaranteed given the practices and attitudes of the time.

  • Charles,

    I am very saddened to hear this.

    In preparing the article, I read several accounts by survivors and relatives whose loved ones perished in these accidents. There was not enough space to do justice to their stories.

    Now your comment will speak for them.

  • Sal,

    No. The NTSB classified the cockpit area as non-survivable in both mishaps. Surprisingly, they were able to restore many of the flight and navigation instruments to working order to aid the investigations.

    Another Allegheny pilot deadheading on flight 736 was also killed. There is an image of his gravestone with a beautifully rendered Convair on the web. Search “Allegheny flight 736 grave sites”.

  • I stand corrected.

    NPAs are allowed by many airlines when a vertical descent profile is available e.g. LNAV+V, Baro or FMS generated. My statement was imprecise and incomplete. Thanks for prompting me to take a closer look.

    But for those of us in the low-performance world, many NPAs remain where vertical guidance isn’t provided.

Leave a Reply

Your email address will not be published. Required fields are marked *