FSX-SE NDB "R" ("ROMEO" or "ROZO"?) - Tragedy of AA 965

[P_7878 6,594]

For a recent while, I've been meaning to fly my CS-757 again (have not flown it for about 4 years since my break - I did, however, fly and post an entry with my QW 757 here). For B757, the subject-flight of this post caught my attention due to its uniqueness. This is especially because, there was a time, I was scrupulously bringing up my (FMS) Navigraph database up-to-date for PMDG/iFly planes, but, frankly, I have been lately content with not-up-to-date databases for use with the FMS (including a couple of Carenado aircraft, I purchased recently, that, by default, come equipped with not the latest navigational data for the FMS). I just happen to clear the FMC message and proceed with my flight. This is after all simulation! Of course, kudos to those simmers, who faithfully maintain their navigational database! However, the topic of this post, for the first time, made me realize how critically significant this is in RL, where actual lives depend on it! Also, I now clearly recall the young aviatrix, who went around the world in a PC-12, and had said, "...I realized that when I was in flight, I was responsible for my own life and anyone else that was in the aircraft...". Not so in simulation!

American Airlines Flight 965 was a regularly scheduled flight from Miami International Airport in Miami, Florida, to Alfonso Bonilla Aragón International Airport in Cali, Colombia. The aircraft was a Boeing 757-223. On December 20, 1995, at 5:14 pm, Flight 965 pushed back from Gate D33 in Miami, and then taxied to runway 27R, but seasonal congestion caused the 757 to actually take off at 6:35 pm, 1 hour 21 minutes late. The Flight mainly carried people returning to Colombia for the Christmas holidays, vacationers and business-people. The cockpit crew consisted of Captain Nicholas Tafuri, age 57, and First Officer Don Williams, age 39: both pilots considered to be highly skilled airmen. Captain Tafuri had more than 13,000 hours of flying experience and First Officer Williams had almost 6,000 hours. All cabin crew personnel were born in Colombia and were veterans from Braniff International Airways who had moved to Eastern Air Lines and then to American Airlines, when the routes were transferred from one airline to the other. They had voluntarily chosen this flight, as a prerogative awarded by seniority, to spend Christmas time with their families in Bogotá.

At that time, Cali Airport's air traffic controllers had no operative radar (due to its destruction by prior terrorist-activity) to monitor such aircraft. Instead, Cali's approach used several radio beacons to guide pilots around the treacherous (high) mountains and canyons that surround the city. So, a sense of mutually shared situational-awareness between ATC and Crew was of paramount importance.

On approach to Cali, since the wind was calm, the controllers asked the pilots whether they wanted to fly a straight-in approach to runway 19 (from the north) rather than coming around (an essential U-turn) to ILS runway 01 (from the south). The pilots agreed to approach straight-in to RWY 19, hoping to make up some time. At that point, they decided to clear the pre-programmed approach waypoints from the FMS, and, therefore, when the controller (subsequently) asked the pilots to check back in while over Tuluá VOR (ULQ), north of Cali, this navaid was no longer programmed into the FMS. So, the pilots had to consult their maps to find it. By the time the pilots found Tuluá's coordinates, they had already passed over it. In response to this, they next attempted to program the FMS for the next approach waypoint NDB "ROZO" (south towards RWY 19). However, the "ROZO" NDB was identified as "R" on their Jeppesen approach plates. But, Colombia had duplicated this identifier ("R") also for the "ROMEO" NDB (near Bogotá), and the computer's list of stored waypoints did not include the "ROZO" NDB as "R". The key point here being that the FMS-sourced navigational information used a different naming convention from that published in the navigational charts they carried. In cases where a country allowed duplicate identifiers, it often listed them with the largest city first (Bogota over Cali, here). By picking the first "R" from the list, the captain caused the autopilot LNAV to start flying a course towards Bogotá NDB "ROMEO", resulting in the airplane turning east in a wide semicircle (and, unfortunately, directly towards the mountains on the east side across the Cauca Valley, please see FSX MAP screenshot of the nearby terrain topology). "ROMEO" is a navaid 150nm east from "ROZO" (and 130nm east from Tuluá VOR (ULQ) on 80deg radial - consistent with the CS757 LEGS data shown in the FMS screenshot). Ironically, both ROZO/ROMEO NDBs have the same frequency (274).

Coincidentally on my (currently outdated) CS-B757 FMS, typing in "R" indeed inserted (the incorrect) NDB "ROMEO" into the (Cali) route just as it did for the Crew (please see ND screenshot of approach to Cali), but mine is mere simulation!). So, the 757 in my final flight-segment, here, is also turning east (from Tulua VOR) towards "ROMEO".

This turn caused confusion in the cockpit since ROZO 1 was to be a straight-in approach (directly south). Therefore, sensing something wrong, 87 seconds after commencing the turn, the crew activated Heading Select (HDG SEL), which automatically disengaged LNAV and now they started a right turn intended for back to RWY 19 (BTW, I've terminated my flight at this point right after initiating a similar right-turn - the last image). Shortly, thereafter, the Crew heard the Ground Proximity (GPWS) warning sound, but they were already too close to the high mountains. With increased engine power and nose-up, the crew tried to climb. However, aircraft's speed brakes were previously deployed to slow it down for descent. At 9:41pm EST, the aircraft crashed into a mountain at about 8900 feet not too far below the crest - 151 out of the 155 passengers and all 8 crew members lost their lives. There were 4 survivors. It is presumed that, without the speed brakes been deployed, the aircraft may have cleared the crest of the mountain (which observation naturally (later, during analysis) begged the question: Should the speed brakes be automatically disengaged when full-throttle is applied for climb...?)

The crash was investigated by the Special Administrative Unit of Civil Aeronautics (Aeronáutica Civil) of the Republic of Colombia, with assistance from the U.S. National Transportation Safety Board (U.S. NTSB) as well as other involved parties. Among the many probable causes, were cited: lack of situational awareness, and not reverting back to basic radio navigation (when the FMS-assisted navigation produced confusing results - along with excessive cockpit workload in the critical phase of the flight especially in dangerous terrain and after nightfall).

The Cali (Colombia) crash resulted in (at least) three aviation safety/awareness improvements. They are succinctly captured by the following three items:

1. Where is it Taking Us? (Critical/active cross-check of FMS-based navigation)
2. Greater Situational Awareness (Closer knowledge of aircraft position w.r.t the surrounding terrain: Here drifting off too far from the narrow valley/corridor of safety)
3. Enhanced GPWS (Newer system to give pilots warnings of terrain-conflict, well in advance: The GPWS system on AA 965 did warn the pilots of danger but not until it was too late (= typical case of CFIT))

If you wish, please refer to miscellaneous available sources for additional details. Thanks for viewing/reading this story that comprises a rather unique set of contributory circumstances - tragic nonetheless. [CS(757)/REX]

Edited March 28, 2019 by P_7878

[rightseat 277]

Really impressive recreation. This clearly took a lot of time to plan, write, fly, and take pics. It was a very interesting and informative post. I appreciate the effort. 🙂

[P_7878 6,594]

Much appreciated the thoughtful comments! Thanks!

The trip was long mostly over the ocean..but then it quickly got dark and finally rather cloudy towards the end...for the fateful segment...quite a learning experience overall...