This article was first published today as an op-ed in The New York Times.
Every day more than one million barrels of oil flow to refineries in the United States from western Canada's oil sands region. Producers hope to quadruple that amount in the next decade, arguing that oil from a friendly neighbor will deliver an extra degree of national security.
But this oil is no ordinary crude oil, and it carries with it risks that we're only beginning to understand. Its core ingredient — bitumen — is not pumped from wells but is strip-mined or boiled loose underground.
Industry insiders long considered bitumen to be a "garbage" crude. But now that the light, sweet oil we covet has become more scarce and its price has skyrocketed, bitumen has become worth the trouble to recover. At room temperature, bitumen has the consistency of peanut butter, thick enough to hold in your hands. To get it through pipelines, liquid chemicals must be added to thin it into what's known as dilbit, short for diluted bitumen.
Last month, the National Transportation Safety Board issued a report that was harshly critical of the federal government's regulation and oversight of pipeline safety following a spill of more than one million gallons of dilbit into the Kalamazoo River in Michigan in 2010. The accident underscored not only how different dilbit is from conventional oil, but how unprepared we are for the impending flood of imports.
After the dilbit gushed into the river, it began separating into its constituent parts. The heavy bitumen sank to the river bottom, leaving a mess that is still being cleaned up. Meanwhile, the chemical additives evaporated, creating a foul smell that lingered for days. People reported headaches, dizziness and nausea. No one could say with certainty what they should do. Federal officials at the scene didn't know until weeks later that the pipeline was carrying dilbit, because federal law doesn't require pipeline operators to reveal that information.
The 2010 spill could have been worse if it had reached Lake Michigan, as authorities originally feared it might. Lake Michigan supplies drinking water to more than 12 million people. Fortunately, the damage was restricted to a tributary creek and about 36 miles of the Kalamazoo, used primarily for recreation, not drinking water.
This close call hasn't deterred the energy industry from announcing plans to build or repurpose more than 10,000 miles of pipelines to carry dilbit to the United States and global markets. That includes the controversial Canada-to-Texas Keystone XL pipeline, which would pass through the Ogallala aquifer, the nation's largest drinking water aquifer. It supplies drinking water for eight states and about 30 percent of the groundwater used for irrigation.
The nation's pipeline network was designed to handle conventional crude oil and is governed by laws and regulations that were written long before the unique risks and hazards associated with dilbit began to emerge. In fact, dilbit is exempt from an excise tax that pays for oil spill cleanups, because the 1980 law that created the tax did not consider bitumen from the "tar sands" to be crude oil.
After the spill, Congress passed new pipeline safety legislation, but it will take years for its modest provisions to have any impact. It does not require pipeline companies to reveal whether their lines are shipping dilbit. And while it does require a study of how dilbit affects pipeline corrosion, the scientists conducting that study met for the first time only last month, and their work is not likely to be completed before new pipelines are built or old ones are repurposed.
The N.T.S.B.'s investigation of the Michigan spill identified "a complete breakdown of safety" at Enbridge, the pipeline's operator. But it also revealed that pipeline rules are weakly enforced. One telling fact: Enbridge discovered defects in the area where the pipeline eventually ruptured as early as 2005, and reported them to regulators. Yet the company was able to delay making repairs without breaking any rules.
The N.T.S.B. also found that Enbridge's leak detection system did not work as advertised. The company had said that its sensors could spot a leak and shut down in less than 10 minutes. TransCanada, the company that is building the Keystone XL, makes similar claims. Yet it took operators in Enbridge's Canadian control room 17 hours to realize their pipeline had torn open. Sensors triggered 16 alarms but operators continued to pump dilbit into the line, believing the problem was an air bubble, until someone in Michigan saw oil on the ground and called Enbridge's emergency line.
The leak-detection problem is industry-wide. Oil spill data maintained by federal regulators show that over the last 10 years, advanced leak detection systems identified only one out of every 20 reported pipeline leaks. Members of the public detected and reported leaks at four times that rate.
Now that the industry is aiming to fill almost a quarter of America's domestic oil needs from western Canadian sources, we need a transparent and informed discussion about dilbit's risks and benefits, up-to-date laws and regulations, and improved leak detection.
The N.T.S.B. chairwoman likened Enbridge employees to Keystone Kops in their handling of the Michigan dilbit disaster. It is a label that could come to apply to the rest of us if we don't guard against future catastrophe.