Gone Fishing: The 2005 Cheakamus River (Canada) Derailment

Cheakamus River is a tributary river to Squamish River, originating at Cheakamus Hydroelectric Dam and eventually feeding the Squamish River 70km/43mi downstream. The river runs alongside British Columbia Highway 99 and the British Columbia Railway line (BCR) and is known for whitewater rafting, kayaking and fishing, especially rainbow trout and salmon. The site of the derailment lies at a railway bridge over the river, 72km/45mi north of Vancouver and 218km/135.5mi west-southwest of Kamloops (both distances in linear distance) in the far southwest of Canada.

The location of the Cheakamus River Derailment in the southwestern corner of Canada.

The bridge over the Cheakamus River is part of the British Columbia Railway line, a 2320km/1440mi single-track largely unelectrified mainline connecting Prince George with North Vancouver. Opening in sections between 1915 and 1956 the line originally mostly supplied logging- and mining-operations on northern trips, taking their products to the harbor on the way back south. Ever since passenger services ceased in 2002 the line is almost exclusively used by freight trains, with occasional excursion-trains for tourists in between.

The site of the accident seen from above in 2019, the train came from the south (bottom of the image).

CN A47151–05 was a northbound mixed freight train travelling from Squamish to Prince George via Lillooet. The train consisted of five locomotives at the head of the train, an empty box car, 3 tank cars, 6 empty flatbed cars, 97 empty box and flatbed cars, 2 more locomotives and 27 further empty box cars. Only the three tank-cars on the train were loaded, carrying Sodium Hydroxide, better known as caustic soda, one of the most widespread chemical products being used in various applications from food preparation to paint stripping and paper production. The leading locomotive and the rear of the two helper locomotives were General Electric (GE) Dash 8–40CM units, a six-axle diesel freight locomotive introduced in 1990 with a power-output of 3281kW/4400hp.

BCR Dash 8–40CM number 4607, the leading locomotive from the accident, photographed in 2009.

The second, third and fifth locomotives at the head of the train as well as the leading of the two helpers were GE Dash 9–44CW units, the former’s successor with an introduction-date of 1993. The Dash 9–44CW is also a six-axle diesel freight locomotive, also producing 3281kW/4400hp and measuring 22.45m/73ft in length at a weight of 192.8 metric tons each.

BCR number 4645, the second locomotive in the train involved, photographed in 2008.

BCR number 766, running in fourth position, was an EMD SD40, a six-axle diesel freight locomotive introduced in 1966. Developed for both freight trains and shunting work these locomotives weight 160 metric tons at 20m/65.6ft in length and produce 2240KW/3000hp.

BCR number 766, the fourth locomotive on the train, photographed in 1990.

At the time of the accident the second, fourth and fifth locomotives were operational and producing a combined 11800hp, while the leading locomotive was running but not supplying traction-power and the two helpers were similarly running but not in gear/”loading”. The train measured approximately 2.8km/1.77mi in length at a total weight of 5002 metric tons.

On the 5th of August 2005 the mixed freight train was climbing the incline along Cheakamus River, recording 25.5mph at 6:44am. The train crew that had driven the train up to Squamish had received warnings about a problem with the DP-system (Distributed Power, the tactic of spreading locomotives out throughout the train), but since they noted no irregularities they had not told the new crew of those warnings. During the first stretch of the trip the driver had become suspicious of the helper locomotives not providing pulling-power, so he separated the remote-control for them from the throttle-control for the leading locomotives and selected throttle-level 8 for the helpers while setting the five locomotives at the head of the train to idle. This immediately caused the speed to plunge, confirming the driver’s suspicion about insufficient power, if any, coming from the helpers. Droppping to 13.7kph/8.5mph the driver activated the leading locomotive, adding six driven axles and 3281kW/4400hp to the traction-power.

Approaching the bridge over the Cheakamus River the driver manipulated the throttle-settings a few times, reaching a speed of as high as 55kph/34mph before reducing the speed to 36.5kph/22.7mph to adhere to an upcoming speed-limit as the rail line began navigating several turns at an incline of up to 2.14%. At 6:59am the train’s data-logger registered wheelslip on several axles, indicated by a sudden spike in recorded speeds not in line with the rest of the train. The head of the train proceeded to begin crossing the Cheakamus River Bridge before, at 7:14am, the train declared an emergency and automatically applied the brakes, coming to a stop just north of the bridge.Nine cars had separated from the rest of the train, derailing to the inside of the left hand turn at the bridge. Car 4, a tank car designated PROX 64041, had fallen off the bridge onto the northern bank of the river, with several more center beam cars also derailing and falling into the river or onto the brush land ahead of the bridge.

The impact with the rocky ground had ripped the tank car open, spilling about 80% of the 86.4 metric tons of Sodium Hydroxide it had been carrying. Within minutes the material coated most of the surrounding area up and down the stream on a length of 45m/148ft. Having already notified dispatch of an unknown emergency causing an autonomous stop the driver now added information about the spillage of hazardous cargo, with CN sending specialists to the site by noon. Within a short time the chemical spread downstream, with dead fish soon showing up on the banks of the river, often showing skin-burns and gill tissue damage due to the high-PH corrosive chemical they came in contact with. In the following days over half a million fishes were declared dead because of the spill before the chemical dissolved enough to lose its harmfulness, largely ridding the river of fish. There are, at least, no records of humans suffering injuries during the derailment or spillage.

A pile of dead fish collected on a stretch of the river’s bank near the site.

While specialists took care of the tank car and the cleanup downstream of the site investigators examined the train and rail line, trying to figure out why the freight cars derailed in the first place. With the tracks showing no sign of a preexisting defect or sabotage they focused soley on the train, soon getting suspicious about #4621, the helper-locomotive which received the remote-control input from the driver for the two helpers. Notably, the locomotive was facing the wrong way, with the cab to the rear of the train while the leading locomotives had obviously been travelling cab-first. This in itself is not a fatal error, as the remote-control system can be set up to feed opposite-direction information to the helpers so that a “forwards”-input by the driver is registered as “reverse” by the helper, making the locomotives move in the same direction. However, investigators discovered that the system hadn’t been set up properly during assembly of the train. This meant that, as soon as the assembled train was put in gear, the leading and the helper locomotives tried to pull the train in opposite directions.

The helpers’ onboard control system recorded data indicating that this error had been noticed, automatically cutting power to the axles as a fail safe-mode. Normally the train would then be stopped and the driver would have to reset the helper locomotives, during which he would properly set the directional mode. However, due to the route from North Vancouver to Squamish being relatively level the leading locomotives had enough power for the passive helpers to go unnoticed.

A diagram from the report showing the elevation of the rail-line from North-V

Looking closer at the configuration of the train the investigators found that both the leading locomotive and the “master”-unit of the two helpers were older Dash 8-units, who, in contrast to the Dash 9-units making up most of the train’s locomotives, did not tell drivers the status or power-output of remote-controlled locomotives. Had the helpers and leading 2 locomotives each switched places, putting a pair of Dash 9-units in charge, the drivers would have discovered the error already on the trip into Squamish and been able to fix their mistake at the latest when handing the train off to the new crew before the train started up the mountain route towards Lillooet.

The way the train approached Cheakamus River Bridge it acted like a regular train without helper-locomotives, as those had effectively been turned into 2 very heavy train cars way down the train. However, being unaware of this, the crew still ran the train as if it had distributed power. This, along with the excessive power on the front end and the sharp, inclined turns, eventually caused the freight cars between the locomotives to “stringline” the curve and derail. Essentially, the weight of the train, especially the inoperable helpers, contrasted with the leading locomotives pulling from the front, caused the train to “straighten out” instead of following the curved track. The basic concept of this phenomenon is the same as if you were to take a loosely lying piece of yarn and pull on both ends. Effectively, too much power was pulling on too much train from the front while gravity wanted the train to go the other way. A shorter train would have been less likely to stringline, as would a heavier train (more weight pressing the cars onto the tracks) or the same train on a less curvy route.

A conventional train without distributed power actually has limitations in how much horsepower is allowed to be applied, reducing the risk of the train being dragged off the rails in such situations as the one that led to the accident. Similarly, there are rules how many empty cars a train can contain and where the loaded cars are to be placed. This results in the simple fact that, the moment the train crew noticed the helpers not properly working, the train should have stopped ascending the rail line as it no longer fulfilled the requirements for a train without distributed power.

After the accident CN improved training of train crews in the operation of Distributed Power and added sections to their handbook about what to do if helpers on a train with distributed power fail to provide propulsion. They also reduced the maximum number of cars in conventional trains travelling northbound on the line to 80. The train involved in the accident had 144, more than allowed without helpers even before the derailment. CN also had to pay for the cleanup-effort, while a precise number was never published the damage and spill cost the company at least 6.66 Million Euros/7 Million USD. The accident wasn’t the only derailment of a CN-train leading to a toxic spill, causing widespread criticism claiming that the railway wasn’t taking the dangers involved in hazardous cargo serious.

Only six years after the accident did fish-populations in the river reach sizes worth mentioning again. The Dash 8–40 has largely disappeared from CN‘s fleet, with the company putting the last units up for sale in 2020. With more modern units with a more sophisticated control system a repeat of the accident on the same cause is near-impossible, further being avoided by tighter rules and restrictions on hazardous cargo.


Note: As previously explained I’m currently unable to publish these posts on Reddit as I used to, if you feel like a post deserves to be shared there feel free to do so in my place.



Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store