Narrative:

On short final to a low night IMC ILS approach; before the 'approaching minimums' call; the pilot flying inadvertently selected toga in addition to deselecting autothrottle and autopilot. As pilot monitoring; I could see a glow from the ALS but I don't know what the pilot flying saw; he did not verbalize his intention to start manual control of the aircraft. My first indication was the flight director (FD) bars commanding a climb and a 30 knot increase in target airspeed. Looking lower over the nose; I could see the ALS and TDZ. With the FD providing GA guidance; the pilot flying had difficulty flying raw ILS data and transitioning to visual flight reference with the PAPI. With a wind shift from left quartering tailwind to right crosswind; the pilot flying executed some track and glideslope oscillations and we got a sink rate warning from the GPWS below 100 feet RA. I called for nose up and power while backing up the pilot flying on the controls. A nominal touchdown and roll out was made. In the short time frame of the event from FD mode change; through what is our status (VMC); to departure from glideslope; I considered trying to re-establish proper FD ILS guidance; but thought outside visual reference and less heads down FMS management would be better. Short answer; better training. On a subsequent cavu night; cleared for the visual approach; cleared for landing using ILS guidance; I selected toga at 400 feet AGL to see if FD ILS guidance could be re-established quickly. With the pilot flying going high on the glideslope in short order; localizer but not glideslope guidance was reacquired. From this limited analysis; it appears a quick de-selection of the FD's per fom would have been our best course to eliminate confusing guidance. For decades; standard simulator training has entailed very predictable transitions from IMC to VMC landing; or no landing environment sighted/fouled runway GA's and missed approaches all at the charted approach minimums. With landing visibility; maintaining the aircraft's current glidepath and course with minimal adjustment gives a high successful landing rate; there's not much time for the pilot to deviate. Perhaps adding a variable approach scenario with higher ceiling and visibility (the more common real world situation) to practice the IMC to VMC transition with erroneous FD guidance added to improve pilots' competence and response to an abnormal. Globally; the requirements to use the highest levels of automation and maintain electronic glideslope when available has degraded large airline commercial pilots' basic airmanship abilities from VFR maneuvering to raw data instrument navigation. An old rule of thumb was aviate; navigate; communicate. To that I would put communicate as the fourth priority. On the line; I see making inputs to the FMC as the highest priority in 99% of the first officers. Prime example: new ATC clearance; 'abc turn right heading 330; direct xyz when able.' with the aircraft in heading select; autopilot engaged; instead of quickly complying with ATC instructions with a one-step twist of the heading select knob on the MCP; the first officer goes heads down to the CDU and executes six to seven keystrokes; verify with pilot monitoring; execute; and select LNAV while the aircraft has traveled four to six miles straight ahead; not turning as cleared by ATC. I just think it is very indicative of an over reliance on automation and a deterioration of thinking and flying ability.

Google
 

Original NASA ASRS Text

Title: B737-800 Captain reports an inadvertent TOGA selection by the flying First Officer; approaching minimums; followed quickly by deselecting autopilot and autothrust. The approach is continued to landing in marginal VMC with some glideslope excursions and the flight director indicating go-around.

Narrative: On short final to a low night IMC ILS approach; before the 'approaching minimums' call; the Pilot Flying inadvertently selected TOGA in addition to deselecting autothrottle and autopilot. As Pilot Monitoring; I could see a glow from the ALS but I don't know what the Pilot Flying saw; he did not verbalize his intention to start manual control of the aircraft. My first indication was the flight director (FD) bars commanding a climb and a 30 knot increase in target airspeed. Looking lower over the nose; I could see the ALS and TDZ. With the FD providing GA guidance; the Pilot Flying had difficulty flying raw ILS data and transitioning to visual flight reference with the PAPI. With a wind shift from left quartering tailwind to right crosswind; the Pilot Flying executed some track and glideslope oscillations and we got a SINK RATE warning from the GPWS below 100 feet RA. I called for nose up and power while backing up the Pilot Flying on the controls. A nominal touchdown and roll out was made. In the short time frame of the event from FD mode change; through what is our status (VMC); to departure from glideslope; I considered trying to re-establish proper FD ILS guidance; but thought outside visual reference and less heads down FMS management would be better. Short answer; better training. On a subsequent CAVU night; cleared for the visual approach; cleared for landing using ILS guidance; I selected TOGA at 400 feet AGL to see if FD ILS guidance could be re-established quickly. With the Pilot Flying going high on the glideslope in short order; LOC but not glideslope guidance was reacquired. From this limited analysis; it appears a quick de-selection of the FD's per FOM would have been our best course to eliminate confusing guidance. For decades; standard simulator training has entailed very predictable transitions from IMC to VMC landing; or no landing environment sighted/fouled runway GA's and missed approaches all at the charted approach minimums. With landing visibility; maintaining the aircraft's current glidepath and course with minimal adjustment gives a high successful landing rate; there's not much time for the pilot to deviate. Perhaps adding a variable approach scenario with higher ceiling and visibility (the more common real world situation) to practice the IMC to VMC transition with erroneous FD guidance added to improve pilots' competence and response to an abnormal. Globally; the requirements to use the highest levels of automation and maintain electronic glideslope when available has degraded large airline commercial pilots' basic airmanship abilities from VFR maneuvering to raw data instrument navigation. An old rule of thumb was Aviate; Navigate; Communicate. To that I would put communicate as the fourth priority. On the line; I see making inputs to the FMC as the highest priority in 99% of the first officers. Prime example: new ATC clearance; 'ABC turn right heading 330; direct XYZ when able.' With the aircraft in Heading Select; autopilot engaged; instead of quickly complying with ATC instructions with a one-step twist of the Heading Select knob on the MCP; the First Officer goes heads down to the CDU and executes six to seven keystrokes; verify with Pilot Monitoring; execute; and select LNAV while the aircraft has traveled four to six miles straight ahead; not turning as cleared by ATC. I just think it is very indicative of an over reliance on automation and a deterioration of thinking and flying ability.

Data retrieved from NASA's ASRS site as of July 2013 and automatically converted to unabbreviated mixed upper/lowercase text. This report is for informational purposes with no guarantee of accuracy. See NASA's ASRS site for official report.