Probable Cause #43: Into the Water

Years ago I used to fly into the Toronto City Centre Airport (CYTZ), although in those days it was called the Toronto Island Airport, because it was built on an island in Lake Ontario.Back then, the only instrument approach that was available was an NDB approach over Lake Ontario towards the city and all of its buildings. It was scary when you knew that the CN Tower, which hails 2,056 feet into the Toronto skyline, was less than a mile northeast of the airport. But the beacon was on the airport property, so if you paid careful attention to your track, it worked out fine.

In July 2003, the pilot of a Beech B58TC Baron departed Lansing, Ill., for an instrument flight to Toronto City Centre where he was to take part in a business meeting. At 5:12 a.m. (all times are EDT), when the pilot called for his briefing, the forecast for City Centre had yet to be issued. However, the forecast for the Lester B. Pearson International Airport (CYYZ), located on the northwest side of the city, indicated that the lowest ceiling that was expected that morning was 2,000 feet overcast with a visibility of two miles in thundershowers and mist. The actual weather was consistent with the forecast at the time the pilot received his briefing. With no other indications that suggested otherwise, the Baron pilot probably thought he was in for an easy flight.About 30 minutes later, the CYTZ forecast was issued and it, too, suggested a similar forecast as that provided for CYYZ. However, beginning with the 6 a.m. AWOS for CYTZ, the ceilings were reported occasionally as low as zero feet and the visibility deteriorated to one mile between 6 and 7 a.m. At 7:24 a.m., an amended forecast was issued that indicated a temporary reduction in visibility to one mile in light rain and mist with scattered clouds at 200 feet until 10:00 a.m. The Baron pilot received this forecast when he checked the weather while enroute at 8:25 a.m.At 9:46 a.m., a new forecast was issued that indicated ceilings of zero feet and a visibility of 1/8-mile, improving to two miles visibility in mist and an 800-foot broken ceiling between 10 a.m. and noon. By this time, the pilot was about to enter a hold to wait his turn for the approach and it is believed that he never received this forecast.Tower controllers at City Centre noticed some improvement beginning at about 9:30 a.m. when the RVR was 2,800 feet and the top of the CN Tower was visible. At that time the aircraft ahead of the Baron began its approach to the airport. The crew later told investigators that they had visual contact with the surface of the lake when they were holding approximately 15 miles southeast of the airport at 3,000 feet. At 3.0 DME the crew noted that a lighthouse protruded through the fog layer, indicating that the fog depth was about 50 feet. At 2.2 DME the crew could see buildings and trees through the fog and some of the city's features.The approach in use was the Localizer/DME-B Approach, which only has circling minimums of 760 feet and two miles. The crew stated that as they began their turn to final approach, the precision approach path indicator (PAPI) was briefly visible, the mid-portion of the runway was briefly visible, but both ends of the runway were obscured by fog. As a result, the crew elected to execute a missed approach.At that time the tower controllers stated that conditions on the surface appeared to be "almost VFR," with an RVR of 6,000 feet. The AWOS at 9:44 a.m. reported the ceiling was unlimited with a few clouds at 8,800 feet, but that the visibility was still 1/4 mile.

The Baron's flight from the Chicago area to Toronto was uneventful. The pilot had filed for and flew at 13,000 feet until reaching London, Ontario, where he began a descent to 7,000 feet. He was later cleared to 4,000 feet and told to hold at the TILEL initial fix, located 15 miles east-southeast of the airport on the localizer. At 9:48 a.m. the Baron entered the hold at TILEL and the pilot was given an expect further clearance (EFC) time of 10 a.m. because the other aircraft ahead of him was still on the approach.At 9:50 a.m. the first flight crew declared their missed approach and they indicated to the controller that only part of the runway was in sight. The Baron pilot acknowledged hearing the report and the controller issued the 9:44 a.m. AWOS report. The pilot told controllers that he would like to fly the approach and he was cleared to do so with the stipulation that he cross TILEL at 3,000 feet.As the Baron passed TILEL inbound, the pilot was instructed to contact the City Centre Tower controller. The frequency change was accomplished and the Tower controller instructed the pilot to report the runway in sight or the missed approach. He was given the option of landing on Runway 8, 26 or 33.The pilot acknowledged the landing clearance and was not heard from again. After passing VOKUB, the final approach fix (FAF), the aircraft continued inbound on the localizer and descended below the minimum descent altitude (MDA) of 760 feet until it struck the water at approximately 3.6 DME, 1.6 miles short of the missed approach point.Several hours later, the Metropolitan Toronto Police Marine Unit located floating debris in the water. Using a side-scanning sonar, the aircraft was located the following day, but it would take another two weeks until it was recovered from Lake Ontario.

The private pilot was originally licensed in 1960 and possessed multiengine and instrument ratings. He had approximately 700 hours of flight time in the Baron that he had owned for 11 years. He had undergone a biennial flight review and instrument competency check two days before the accident occurred. During that flight he flew two ILS approaches and one GPS approach in VFR weather conditions. The instructor told investigators that the pilot hand-flew the aircraft throughout the approaches and his handling of the aircraft was normal.Records indicate that the pilot flew 15 hours in the 90 days preceding the accident. He did not log any instrument time or any actual or practice approaches. Looking back to the beginning of 2002, the pilot flew three non-precision approaches in instrument conditions to his home field in Lansing. He also flew two training flights about 10 months before the accident occurred. Two ILS approaches were flown near his home airport and two non-precision approaches to Lansing were made.There is an ILS/DME Runway 8 approach into City Centre Airport, but it is not available if CYYZ is landing on Runways 23 and 24, as they were that morning.The investigation revealed no discrepancies with the aircraft or onboard equipment. The altimeter was found to be correctly set and the pilot flew all altitudes accurately, with the exception of descending below the MDA. The flight path of the aircraft was steady throughout the flight and it appeared that the pilot correctly flew the localizer inbound. The aircraft struck the water in a wings-level, nose-level attitude, suggesting that the pilot was in control up to the point of impact. The landing gear was extended and the flaps were set to what would be considered the approach setting. There was no indication of pilot incapacitation.

Investigators with the Transportation Safety Board of Canada turned their attention to "controlled flight into terrain" (CFIT). They identified several factors in this accident that are seen in many CFIT accidents. They are:

The weather was considered as well because it was below what had been forecast when the flight began and significantly below the revised forecast for the two-hour period when the aircraft arrived in the Toronto area. A new forecast was issued just before the pilot began his approach, but there is no evidence that the pilot ever received it.But would it have made any difference if he had? When the crew of the aircraft ahead of the Baron passed along that they had sporadic ground contact, the Baron pilot might have assumed that conditions were improving and that he would have a better chance of locating the runway than the first crew did. At that point the forecast itself probably would not have mattered, especially considering that the weather had not done what it was forecast to do anyway.Also, the RVR was indicating that the visibility on the runway was more than a mile, although the AWOS was still reporting a visibility of 1/4 mile. When the other crew reported seeing parts of the runway, it might have bolstered the pilot's impression that conditions were improving enough that by the time he got to the runway he'd be able to see it well enough to complete the landing. After all, the AWOS was reporting no ceiling.Based on the report made by the other crew, the fog itself was not very thick. It's entirely possible that the Baron was in and out of the clouds during the approach, and that the pilot was shifting his focus between scanning the instruments and looking outside the aircraft for evidence of the airport and the city.It's also possible the pilot was rushed as he began his approach. When he received his approach clearance, the aircraft was positioned close to TILEL at an altitude 1,000 feet above his cleared altitude. Radar indicates that he began an immediate descent with his airspeed increasing, so the aircraft's gear and flaps would probably have still been retracted.The aircraft passed TILEL 500 feet above the cleared altitude and soon afterward it leveled off and the speed decreased before the descent was resumed. The pilot probably extended the flaps and landing gear during this period and should have completed his before-landing checklist at this point.The aircraft then continued its descent, with a relatively high speed and rate of descent, eventually decreasing to what would be expected, about 120 kts and 700 fpm. They remained at those rates for about one minute. Had the pilot maintained this airspeed and rate of descent, he would have crossed the FAF at about 1,300 feet. It also would have placed the airplane at MDA at 3.0 DME, one mile from the missed approach point.But that's not what happened. Two miles before the aircraft reached the FAF, the rate of descent and the airspeed began to increase, and the aircraft crossed VOKUB at 1,000 feet, which is the minimum crossing altitude. But the 1,200-fpm rate of descent continued beyond VOKUB, which was greater than that needed to cross the 2.0-DME fix at the MDA of 760 feet. This rate continued until the aircraft struck the water.So, why did the aircraft fly into the water? We don't know if the pilot was monitoring his instruments during the final part of the descent or if he was looking out the window for the airport. It is likely that he saw the downtown buildings and possibly had some visual contact with the ground or water.Perhaps the water and the fog combined to create an optical illusion, presenting a false horizon that led the pilot to believe he was straight and level when he was instead descending towards the lake. The visibility above the very low-lying fog layer was probably good. It's possible that the aircraft's rate of descent was such that by the time the pilot realized he was in the fog bank, it was too late to look back at his instruments and arrest the descent before hitting the water.This accident points out the need for pilots to follow procedure when flying an instrument approach. One has to resist the temptation to look outside for the airport or runway too early when visibility is marginal or obscured. Optical illusions and the risk of spatial disorientation are real threats that have claimed numerous lives.