http://www.irlgov.ie/tec/aaiu/2001Re.../2001-0010.htm FDR readouts should count as facts IMHO:
At 3.5 miles from touchdown and 1 minute 17 seconds to run, the autopilot was disengaged. The barometric height was 1966 ft, RADALT was 1110 ft, CAS was 162 kt and the heading read 219°magnetic. Manual control was initiated but the autothrottle remained engaged and in speed mode. Ceilings were reported as 1800 ft AGL with good visibility, and the approach was continued visually. From this point to touchdown, the aircraft was below the glideslope, ranging from 0.32ddm (3.6 dots) to twice that amount 2.5 seconds before touchdown.
At 1000 ft AGL, the speed selected was at 155 kt or (V ref+12), and remained at that value until touchdown. The actual approach was flown at an average speed of about 163 kt, but excursions of 150 kt to 173 kt were recorded due to gusts and turbulence. Strong headwinds were indicated on the initial approach and decent, but by the time the aircraft was at 700 ft radio altitude, the headwind component had largely vanished. The gusts, turbulence, and variable crosswinds contributed to continuous oscillations in all parameters, particularly pitch attitude.
At 0.6 miles and 14.4 seconds to go to touchdown the “glideslope” warning sounded. The barometric height was now 1016 ft and the RADALT was 186 ft. The CAS registered 160kt with an aircraft heading of 233° magnetic. The “glideslope” warning was again repeated as the aircraft was still below the glideslope.