Summary of Risk Arguments

CPE has been interested in the hazard introduced by the bridges at the Second Narrows of Burrard Inlet, and we feel that the hazard has not been properly analysed for risk.  We have looked at three studies undertaken by Kinder Morgan Trans Mountain, and offer our assessment of these studies.  We believe that the risk studies should have considered the probability of bridge collision as a major item, but they have not.

Three studies concerning an assessment of navigational issues have been done for Kinder Morgan Trans Mountain between 2012 and 2014. These concerned the Westridge Terminal, the Strait of Georgia and the Haro Strait, all conducted by LANTEC Marine Inc.  

The three reports were:

  1. Summary Report of Manoeuvring Assessment, Haro Strait Tanker Escort Tug Force Analysis

Desktop Simulation Manoeuvring Study, LANTEC Marine Inc., 23 January 2012

  1. Summary Report of Manoeuvring Assessment, Westridge Terminals Vancouver Expansion

Design Options 11 and 12, LANTEC Marine Inc., 4 October, 2013

  1. Summary Report of Manoeuvring Assessment, Strait of Georgia Proposed Tug Escort Desktop

Simulation Manoeuvring Study,  LANTEC Marine Inc.,  22 August 2014

In addition, a fourth report was produced, concerning tug escort procedures while assisting a tanker in distress near the Second Narrows. This followed the initiation, in 2006,  of a review by the British Columbia Coast Pilots (BCCP) and the Vancouver Fraser Port Authority (VFPA) to scrutinise their procedures for commercial traffic transiting the Second Narrows:

A Case Study In Improving Tug Escort Procedures With The Aid Of Simulation, by Gregory Brooks (Towing

Solutions Inc., USA) and Garland Hardy (LANTEC Marine Inc., Canada)

All these studies rely on the application of a computer simulation or a computer-based Bridge Simulator.  In the latter, captains and pilots can be virtually exposed to normal or emergency situations, including mechanical failures, different weather conditions and coordination between tankers and the accompanying tugboats.

The simulator used is a very sophisticated type of software developed by Kongsberg in Norway, under the name of K-Sim® Polaris – Ships bridge simulator, and is used for maritime training of ships’ bridge crews.  At the request of Trans Mountain, it was applied to the simulation of possible incidents in either the Georgia Strait or Haro Strait, and for the design of the docking terminal. It can forecast for example, the expected outcome after a ship’s rudder failure and the subsequent assistance from tugboats. The studies also included communication logistics between a distressed tanker and tugboats. In the fourth study, a computer simulation program was used to study the outcome of the interaction between a distressed vessel, operating in the Second Narrows, and the assisting tugboats under different environmental conditions.

While these types of software can be said to be essential tools for the consideration of maritime incidents or training of the ships’ bridge personnel, they have been applied to specific conditions or to the study a specific, isolated situation.  What is needed, as part of a risk analysis, is a study of the safety of the operation that would span the distance starting at the Kinder Morgan terminal at the head of the Burrard Inlet, through to the Pacific Ocean. The simulations can be adapted for this purpose, however, by considering and studying a large sample of situations generated at random.

A risk analysis would consider, for example, a rudder incident at a random location, under possible adverse weather conditions, and under tugboat assistance with chance human errors of communication between the crews.  Taken all together, it would then be possible to estimate whether the operation could be completed without incident.

Our conclusion is that, while the approaches used in the four reports are the correct ones and rely on very sophisticated tools, the application is limited and cannot be used to reach conclusions about the safety of the proposal.

It must also be said that the simulations can only consider random emergencies for the crew to confront, but do not say anything about whether or not a spill will be triggered nor the magnitude of its consequences if this does occur.

The simulator in the fourth study could have been used to estimate the probability of collisions with the bridge piers at the Second Narrows in  the Burrard Inlet.  For example, emergency situations like a rudder failure in a laden tanker going west,  between the terminal and the bridges, under different weather conditions and tugboat assistance, to estimate the chances that the tanker would not transit successfully under the bridges. This, in fact, would have been a complete computer simulation of the operation’s performance for a risk analysis. This should have been done to estimate the probability of incidents occurring (like groundings) that could lead to spills, but it was not, probably because of time and budget constraints.

We believe that the risk studies should have considered the probability of collision with the bridge as a major item, not in just the simulation of specific emergencies.

In Canada the Highway Bridge Code S6 provides a method of estimating this probability, but it is approximate in that it relies on data which reflects collision history for bridges in other parts of the world. In any case, given the possible catastrophic consequences of bridge damage or collapse, either a tool like the Polaris simulator or the steps specified in the Code S6 should have been used, but nothing of this sort has been done.

Another study for Trans Mountain considers only the vessel traffic (including tankers and other vessels) under the bridges. It was produced by Ausenco Engineering Canada, under the name Trans Mountain Pipeline ULC, TMEP Simulation Study Analysis of Second Narrows Transits, November 2013. This is essential information to estimate the probability of vessel collisions with the bridges, as required by the method specified in the Code S6. But this study only reports on the traffic data and is not an evaluation of the collision probability.

None of these studies, although they provide pieces of the puzzle, constitute  a risk analysis.

A proper risk analysis would have required the definition of situations considered for failures of the system (for example, an oil spill of a given volume, a collision with the bridges). Oil spills have been considered by Kinder Morgan, but not collision with the bridges.

Since, normally, the factors affecting the failure of the system exhibit uncertainty (for example, weather conditions, tugboat assistance procedures, speeds, human errors, etc.), it is not possible to achieve a system with zero probability of failure.

Instead, for each situation, one needs to estimate the probability that failure would occur over the operating life of the system.

What is acceptable risk? Today, buildings are designed to withstand a certain level of force generated from earthquakes.  There is a two percent chance that a force will exceed what the building can handle.  Kinder Morgan’s analysts have come up with figures that tell us that the risk of any spill occurring is 19%.  For a major spill, (eight million, two hundred and fifty thousand litres) the risk is 10% over the life of the project (50 years).  That is 3000 times larger than the MV Marathassa spill of April 8, 2015 when this cargo ship spilled several thousand litres of bunker fuel into Vancouver’s English Bay.

The risk analysis submitted by Kinder Morgan estimates the probability of an oil spill of a given volume, and does this by reporting to NEB the return periods of spills of different volumes. For example, a spill of 8,250 m3 (which would qualify as a “major” spill) is reported to have a return period of 473 years. This is confusing, as it could be read as estimating a period of 473 years between such spills. The risk analysis submitted by Kinder Morgan wrongly uses this interpretation, which conveniently diminishes the appreciation of the risk. In fact, a return period is the average time elapsed between consecutive spills, and 473 years statistically means that there is a 10% chance of a 8,250 m3 spill at some time during the expected 50 years of operation. This number is considered too high by CPE, but the methodology used by Kinder Morgan could have been verified by using the Simulator applications as described above. Since bridge collision was not included, the probability of failure would be possibly greater than 10% in 50 years.

Still, calculating probabilities of failure does not in itself constitute a risk analysis, because risk must involve an estimation of the consequences of the failure.

Unfortunately, another drawback of the risk analyses submitted to the NEB by Kinder Morgan is the absence of quantification of consequences. What would be the estimated damage produced by a spill of, for example, 8,250 m3? What would be estimated cost of a difficult Dilbit clean up? What would be the cost of damage to the fisheries? Who would pay for the costs? How can the risk be reduced by implementing different mitigation procedures or moving the operations somewhere else? While the studies done concentrate on return periods or probabilities of groundings, when it comes to consequences the public has to rely only on well-meant wishes: that there will be a world class spill response mechanism, a world class communication system between vessels and controllers, and, in the end, that there will be a second-to-none cadre of captains and pilots highly experienced on the navigational demands of the BC coast. What would be the consequences to the regional and national economy of bridge failures due to collisions? Because very little has been done quantitatively about consequences, we therefore know very little about risks, and approval of the project should take this deficiency very much into account.

While the Kinder Morgan submission implies that there will be a 10% chance of a spill greater than 8,250m3 in a 50 year-period,  this also means that there is a 90% chance that such spill will not occur in a 50 year interval. The argument that tankers have operated safely in the same route for 50 years is faulty. First, we are now discussing different, bigger tankers and a larger number of them. Second, the event that nothing has been observed for 50 years is a quite probable event, in fact, with a 90% chance of occurring. With these odds, if nothing has been observed, then it has been a matter of luck. Still, this does not deny that there is a 10% chance that an incident could occur and should be planned for.  For a comparison, a “Big One” earthquake is also estimated to have a return period of about 500 years, or also roughly a 10% chance of occurring during a 50 year window. We have not had a Big One during the last 50 years, but nobody uses this argument to conclude that we should feel safe from such an event. In fact, to the contrary, Vancouver is considered to be in a high seismic risk area, and millions are being currently spent to mitigate the consequences of such an event.