For what it's worth. There has always been active debate about whether autothrottle should be left engaged when landing in marginal conditions of crosswind and turbulence. There has been, as there always is with these "technique" questions, always two schools of thought. When the Operating Handbook doesn't specify (or even make recommendations) obviously there's a question extant i.e. are there hazards involved in either method? are there any benefits? I think this accident answers that question. 

I can recall that I always had a problem convincing pilots under training not to utilise the Flt Eng in those conditions but to take "pilot's power" (i.e. total control). At least you could trust the FE not to close the throttles without a command, but you cannot trust the autothrottle in those conditions. It simply relates to the bug speed and cannot relate to rapid onset of sink rates nor gust factors. The autothrottle is programmed to do that power reduction at 50 feet. In conditions of strong headwind when you are a little low and "powering into" that headwind component, it's the last thing you need (the autothrottle to be fore-closing on you at 50ft), because, in a strong headwind situation, high sink rates can result in a no-flare fall-through. 

So what, you might ask? You should have your hand on the throttles and stop any unwanted autothrottle reductions in thrust. Unfortunately my observations of pilots over the years was that at about 50 feet, entering the flare in strong crosswind conditions, they were always much more preoccupied with getting the decrab/kickstraight and imposition of the appropriate (into wind) wing down for touchdown i.e. they always went for two hands on the yoke at that stage for better roll axis control - and left the power-lever closure to the FE [on command]. If there was no obvious speed or glideslope problem this (the power) always seemed to be a low priority afterthought. 

In some aircraft it is more important to alleviate undercarriage side-stresses than it is to avoid pylon scrapes (as is the main consideration for the 747). So in these aircraft, such as the MD11, a slight wing down after kick-straight in the flare will minimize lateral movement on touchdown. The overriding consideration is that it should be a smooth (and not firm) non-symmetric arrival on that upwind gear. In this accident, because of the unmonitored autothrottle closure at 50 feet, the aircraft picked up a high sink rate and touched down hard on the right wing gear. There has been considerable informed debate about the MD11 undercarriage and wing strength since before the Newark accident. (see this link: http://www.airlinesafety.com/faq/faq9.htm  ). It would seem that, for a safer operation, autothrottle should
be cancelled for wind strengths above a certain speed (i.e. where gusts may be encountered). This would include most significant crosswind situations. This would ensure that handling pilots kept a ready, responsible and responsive hand on the throttles. In this HK case, with both hands on the yoke, the MD11 captain probably did not note the autothrottle action and the resulting speed decay and sink rate caught him out. If he'd had his right hand on the power levers it may well have been a different outcome. He could have overridden the autothrottle and used incremental "power as required". Applying the "out of sight, out of mind" philosophy to the situation, because his right hand was NOT on the throttles there was no tactile awareness of that being the instinctive solution to the close-in sink-rate and speed loss. Been there, done that, seen it hundreds of times. Usually it simply results in a firm landing and discussion as to why. In the MD11 it's just another trap for young players. 

Hopefully this will come out in the investigation and some action will be taken. It's a problem for, and applies to, more than just the MD11. Sometimes you just cannot leave these handling technique matters to individuals. You have to make strong recommendations in the Operating Handbooks together with a clear exposition of "why". 

It falls into that same category of runway handling accidents on take-off or landing for which there is never an available fix. i.e. you can simply not expect that a pilot experiencing post-rotate problems will NB that his flaps or slats are retracted (he'll just never know). You cannot expect (Little Rock MD-82) that a pilot struggling to maintain directional control and aquaplaning will NB that his spoilers haven't actuated. Elimination of these built-in "traps" would go a long way towards reducing runway accidents, warning systems notwithstanding. 

Apart from pilot handling considerations it is really not acceptable that such errors and subsequent hard landings should lead to catastrophic failures and aircraft on fire inverted with trapped passengers. Together with the prior pronouncements on MD11 handling and autopilot aspects it is really adding up to an honest airworthiness review requirement [of the MD11 as a pax carrier]. 

http://www.awgnet.com/safety/news/nzfedx11.htm 

NTSB Weighs Findings Of FedEx MD-11 Probe  James T. McKenna/Washington  The U.S. National Transportation Safety Board is to meet next week to decide whether pilot error or an aircraft design or operations flaw led to the July 1997 crash in which a Federal Express MD-11 freighter ended up on its back and burning.  The five presidential appointees on the board are scheduled to meet here Sept. 8 to review the results of NTSB investigators' two-year probe of the July 31, 1997, crash at Newark (N.J.) International Airport and weigh their recommendations on its probable cause.  Their deliberations could be delayed, however, depending on the initial findings of the probe into the Aug. 22 crash of an MD-11 passenger aircraft that also flipped onto its back after landing in Hong Kong. The NTSB's lead investigator on the FedEx crash, Robert Benzon, was dispatched with a team of safety board specialists to assist in the Hong Kong probe.  In scrutinizing the FedEx crash, Benzon's team of investigators from the safety board, Boeing, the airline and its pilots union, General Electric and other parties have focused on the performance of the MD-11's flight crew as well as the design, maintenance and operation of the aircraft.  Under NTSB supervision, Boeing has performed metallurgical and other analyses of the accident aircraft's right wing and right main landing gear, both of which failed and broke free of the aircraft during the attempted landing.  Investigators have reviewed reports of previous hard landings involving the accident aircraft and other MD-11s as well as DC-10s, the type from which the MD-11 is derived. They also have reviewed the training and work history of the accident aircraft's flight crew and the supervision by FedEx and the FAA of MD-11 flight crew training.  Officials familiar with the draft report of the probe told Aviation Week & Space Technology investigators are expected to recommend to the five-member board that the probable cause of crash be attributed to the captain's failure to control his descent rate and aircraft attitude during the final approach, flare and touchdown.  Board members now are reviewing the draft reports, its proposed findings of cause and safety recommendations.  FedEx Flight 14 was bound for Newark from Anchorage with 145,000 lb. of cargo from Asia, including what the company called 400 lb. of "routine hazardous materials.'' As it touched down on the centerline about 1,475 ft. from the threshold Runway 22 Right, records of the investigation show, the aircraft bounced off the pavement, touched down again, then flipped off the runway at about 1:35 a.m. on July 31. At landing, the wind on the runway was reported to be from about 240 deg. at 5 kt. Buildings on and around the approach end of Runway 22R at Newark can cause gusts of variable direction and intensity, according to pilots who regularly fly into the airport (AW&ST Aug. 11, 1997, p. 38).  At the first touchdown, the aircraft was doing 149 kt. and descending at 500 fpm. The flight data recorder (FDR) indicates the first touchdown had a vertical acceleration of +1.67g and a lateral acceleration to the right of 0.2g. After the bounce, according to the FDR and eyewitnesses, the aircraft slowly rolled right and yawed right while in the air.  The aircraft bounced about 7 ft. off the runway, staying airborne about 3 sec. It pitched down and landed again with a vertical acceleration of +1.7g and a lateral acceleration of about 0.4g to the right.  The second touchdown occurred 2,275 ft. from the threshold, with the aircraft still on the centerline but rolling right. At that point, the nose gear, the right main landing gear and the nacelle of the engine under the right wing contacted the runway. The runway was severely gouged 40 ft. to the right of the centerline for about 1,000 ft. in an area that would have been under the nacelle of the No. 3 engine. The gouging first appeared about 2,513 ft. from the threshold.  Data from the aircraft manufacturer indicate the aircraft would have to have been pitched down 2.5 deg. and rolled 12 deg. right--and the struts of the nose and right gear fully compressed--for the No. 3 engine nacelle to strike the runway. FDR data shows the aircraft was rolled right only 9 deg. when the nacelle struck. Some investigators believe the right main gear already had failed at that point.  The engine subsequently separated from the wing. The spar structure of the right wing was fractured near the wing root, and the wing broke free of the aircraft. The beam to which the main gear is attached on the MD-11 was cracked roughly in half during the accident.  The MD-11F came to a stop on its back 4,800 ft. down and to the right of the runway, about 200 yards from the airport's terminal complex. The bulk of the right wing came to rest about 1,000 ft. short of that position. Sections of landing gear and one of the aircraft's General Electric CF6-80C2D1F wing-mounted engines were scattered with other debris around the wreckage.  NTSB investigators were spurred to seek clues in the nonvolatile memory of computer units on the aircraft after they learned the FDR had stopped recording about 5 sec. before the broken aircraft came to a halt. Wires to the FDR apparently were severed shortly after the aircraft touched down the second time.  Investigators also experimented with extracting data from the nonvolatile memory of the full-authority digital engine controllers from the aircraft. Although investigators believe Flight 14's engines were functioning normally, developing an ability to extract and analyze such data could prove useful in future investigations.  Both of Flight 14's pilots escaped with minor injuries through one of the cockpit windows, as did two Federal Express employees and a worker from another carrier who were riding jump seats on the flight. The resulting fire destroyed the aircraft.  The aircraft that crashed, N611FE, was the first of more than 160 MD-11s in service to have been destroyed in an accident. In the two years since then, three other MD-11s have crashed, including the Mandarin Airlines MD-11 that crashed last week in Hong Kong.