Air Canada Express Flight 8646

The last flight of the day

Air Canada Express Flight 8646 was a scheduled service from Montreal-Trudeau to New York LaGuardia, operated by Jazz Aviation. For air traffic purposes it was Jazz Flight 646. The aircraft, registration C-GNJZ, was a CRJ900, formally designated CL-600-2 D 24. On board were 72 passengers, two flight attendants and two pilots.

The cockpit crew were beginning a four-day trip. They had already flown from Montreal to Quebec City and back, making the LaGuardia arrival their third and final scheduled flight that day. The captain had accumulated 3,560 total flight hours, including 1,600 at Jazz. The first officer had 718 total hours, including 435 at the airline. They had flown together on one previous occasion.

LaGuardia was operating at night in visual meteorological conditions. Light rain had been reported, and the runway was considered wet or contaminated for landing purposes. The latest available airport information reported good braking action throughout runway 4. These conditions describe the setting; the preliminary report does not yet identify weather or runway condition as a cause of the collision.

An emergency elsewhere on the airport

At 23:31:42, an emergency was declared near terminal B. Seven vehicles began responding: four aircraft rescue and firefighting trucks, a tool truck, an airstair truck and a Port Authority police vehicle. Their intended route took them along taxiway D, across runway 4 and toward the terminal. They initially gathered near the intersection of taxiways BB and D.

Rescue 35, an Oshkosh Striker 1500 fire truck, was part of this group. Its radio callsign was Truck 1. The tool truck, Truck 7, was intended to lead the response. At 23:35:47 its crew attempted to contact the tower, but a simultaneous transmission obscured the call. It tried again at 23:36:21, then coordinated with Truck 1 to make contact.

The vehicles were responding to an emergency, but crossing a runway still required coordination with aircraft movements. The preliminary sequence shows the rescue response and the airline arrival developing at the same time. It is the interaction of these two operations, rather than either one in isolation, that explains how a truck came to be in the path of a landing aircraft.

Clearance while the CRJ approached

At 23:35:07, the local controller cleared Jazz 646 to land on runway 4. The CRJ was on a five-mile final at an altitude of about 1,900 feet. As it descended, the rescue vehicles assembled and sought clearance to cross the same runway. At 23:36:44, the controller asked which vehicle needed to cross. Truck 1 moved to the front of the waiting group.

At 23:36:51, the red runway entrance lights at taxiway D illuminated. The approaching aircraft was about 1.2 nautical miles from that intersection and at an altitude of 287 feet. Truck 1 was stationary on taxiway D, approximately 460 feet from the runway edge. The lights provided a separate visual warning about runway traffic approaching the intersection.

Truck 1 identified itself and the accompanying vehicles, then requested permission to cross. The controller instructed them to cross at 23:37:04. At that moment, the CRJ was about a quarter mile from the runway on final approach, approximately 130 feet above ground. It was around 4,400 feet from the taxiway D intersection where the vehicles intended to pass.

Stop instructions and touchdown

Truck 1 read back the crossing clearance and began moving at 23:37:07. Four seconds later, the CRJ crossed the runway threshold. The truck was traveling at about 10 knots, or 11.5 miles per hour, and still approximately 410 feet from the runway edge. The two movements were closing on the same intersection from different directions.

At 23:37:12, after transmitting a taxi instruction to another airplane, the controller told Truck 1 to stop. The truck's speed continued to increase. Five seconds later, the CRJ's main landing gear touched down about 1,450 feet before taxiway D at a groundspeed of 128 knots. The first officer transferred control to the captain. The truck had crossed the hold-short line at about 24 miles per hour.

At 23:37:20, the controller again instructed Truck 1 to stop. The aircraft's braking began and its thrust reversers deployed. The truck was just over 100 feet from the runway and moving at about 29 miles per hour. The public report sets out this sequence in detail, but has not yet completed its assessment of why the stop instructions did not bring the truck to a halt.

The impact and evacuation

At 23:37:21, the runway entrance lights went out. About a second later, the aircraft's nose gear touched down. The CRJ was then approximately 400 feet from taxiway D and traveling at 106 knots, while Truck 1 was entering the runway at approximately 30 miles per hour. The collision followed roughly two seconds later.

The truck turned left just before impact, and the airplane's rudder moved about six degrees left. The last recorded aircraft groundspeed was 90 knots, equivalent to about 104 miles per hour. The collision occurred at the intersection of runway 4 and taxiway D. The CRJ stopped upright near taxiway B, initially with its nose down because of extensive forward-fuselage damage. The fire truck came to rest on its left side nearby.

Both pilots died. The 72 passengers and two flight attendants survived. Passengers evacuated through the four overwing exits, and the responding vehicles diverted from the terminal emergency to assist. The aft flight attendant described an orderly evacuation despite darkness and not knowing what had happened. He remained with four passengers until rescue personnel entered and helped them leave.

During evacuation, the airplane tilted until its tail contacted the ground. The NTSB recorded no aircraft fire. Six people sustained serious injuries: one aircraft crew member, three passengers and both truck occupants. Thirty-three passengers sustained minor injuries. The report's statement that 39 people went to hospitals includes the truck crew, so it should not be presented as 39 injured airliner passengers.

What the wreckage showed

The worst aircraft damage was ahead of the first passenger row. The cockpit and nose were crushed and partly separated. The pilot seats, unused flight-deck observer seat and forward flight attendant seat were found separated from the aircraft. In contrast, all passenger seats remained attached, with only limited deformation in the first two left-side seats and minor or no damage elsewhere.

Investigators found no anomalies in the pitch, roll or yaw control surfaces. The slats and flaps were fully extended. Although the spoilers were found retracted in the wreckage, the recorder showed that ground spoilers had activated about six seconds before the end of its data. The thrust reversers were found fully deployed. These differences illustrate why post-accident positions must be checked against recordings.

The truck's right-side structure and equipment compartments were badly damaged. Its water tank fragmented and its contents were released; the foam tank was breached. The operator cab remained largely intact. The preliminary examination documents how the impact affected the two vehicles, without yet completing the analysis of crash forces or occupant protection.

Why the surface system did not warn

LaGuardia had Airport Surface Detection Equipment, Model X, known as ASDE-X. It combines radar and other inputs to show aircraft and vehicle movement and provide conflict alerts to controllers. FAA analysis found that it did not issue a visual or audible warning of this collision.

None of the seven responding vehicles had transponders. Their radar returns intermittently merged and separated as the vehicles waited close together. The system could not establish reliable individual tracks. Near the time of the accident, it displayed two radar targets for the group rather than seven distinct vehicles. It could not reliably associate Truck 1's movement with the approaching airplane to predict the conflict.

This does not mean that the entire airport surveillance system had simply stopped functioning. The report explains a tracking limitation involving the closely grouped, unequipped vehicles. It also notes that preventive maintenance had been performed the previous day. The presence of surveillance equipment alone therefore did not guarantee that every moving object would be identified well enough to trigger a warning.

Lights, radio and the continuing investigation

The runway entrance lights did operate. Surveillance video showed them red while the rescue vehicles waited and as Truck 1 advanced. The system is designed to extinguish the lights at an intersection about two to three seconds before a landing airplane reaches it. Their going out just before impact is therefore different from evidence that the runway had become safe to cross.

Truck 1 also had a situational awareness tablet. It could show positions and alert the crew when approaching a runway, but it was not designed to predict an impending collision. The preliminary report separates these functions rather than treating every airport display as a collision avoidance system.

Two controllers were on duty. The ground controller, also acting as controller-in-charge, was dealing with the other aircraft's rejected takeoffs and the subsequent terminal emergency. The local controller took over radio instructions on both ground and local frequencies. These duties and their timing remain part of the human-performance investigation.

As of 5 October 2026, the NTSB investigation remains ongoing and the available report is preliminary. Recorders from the airplane and truck, interviews, surveillance video and physical evidence support the sequence described here. They do not yet support a final finding about the relative responsibility of controllers, vehicle crews, airport procedures or equipment. The account will need revision when the NTSB completes that analysis.