What happened
On January 26, 2004, a Do 328-100 operating a scheduled flight from Hamburg arrived at Saarbrücken during heavy snowfall and low visibility. Due to weather conditions preventing a non-precision approach on runway 09, the crew elected to perform an ILS approach on runway 27, accepting a 10 kt tailwind.
During the approach, the crew was briefly held in a pattern while runway conditions were being assessed. The tower reported that the runway was being cleared and provided recent braking action values. Upon landing, the aircraft touched down approximately 850 meters after the threshold. The pilot utilized reverse thrust and ground spoilers, successfully decelerating the aircraft from 110 kt to under 30 kt within 11 seconds. However, the aircraft continued to roll without further deceleration for approximately 25 seconds. The pilot applied the wheel brakes only as the aircraft approached taxiway B. The aircraft eventually exited the paved surface, coming to a stop on the grass approximately 10 meters beyond the end of the runway. The incident resulted in no injuries, though the aircraft sustained light damage and a runway light was destroyed.
The investigation
The BFU examined flight data recorder (FDR) and cockpit voice recorder (CVARG) data, radar records from Deutsche Flugsicherung, and runway friction measurements taken by a Skiddometer. The investigation focused on the discrepancy between the reported braking action and the actual friction levels, the impact of the 10 kt tailwind, and the timing of the pilot's braking application.
Findings
- The aircraft was operating within its weight limits, but the 10 kt tailwind exceeded the operational limits specified in the manual for a contaminated runway.
- The primary cause was the late application of wheel brakes, which, when combined with the tailwind, the descending slope of the runway, and degraded braking conditions, led to the overrun.
- The braking action values communicated to the crew were inaccurate; while the reported values suggested "good" conditions, the actual friction was likely "medium to poor" due to ongoing snowfall and the presence of slush.
- Friction measurements taken by the Skiddometer prior to landing were unreliable because the instrument was measuring over displacing contamination (wet snow/slush), which does not reflect actual braking performance.
- Poor visibility due to heavy snow likely caused the pilot to misjudge the distance to the runway end.