As oil pipeline operators await new federal rules on leak detection, one spill after another teaches a familiar lesson: Usually, it's the general public, or the pipeline company employees near the scene, who discover that a pipe is leaking—not the detection technology that the industry often brags about when it tries to reassure a nervous public.
It's a reality that our own Lisa Song's groundbreaking reporting uncovered almost a year and a half ago in one of the stories that was awarded the Pulitzer Prize for national reporting, a prize she shared with two colleagues at InsideClimate News.
"Few Oil Pipeline Spills Detected by Much-Touted Technology," the headline said in Sept. 2012.
Song, one of a team who investigated a catastrophic oil spill in Michigan and exposed the critical gaps in federal pipeline safety regulations, had analyzed thousands of government records and concluded that "leak detection systems do not provide as much protection as the public has been led to believe." Indeed, 95 percent of the time leaks were detected first by people, not by "state-of-the-art" remote sensing technology, she discovered.
The same point is being made again in the wake of a major September spill in North Dakota that was not discovered for 11 days, until a farmer stumbled across it.
The Wall Street Journal took a look at the same data that formed the basis for Song's report, and not surprisingly, the newspaper reached the same conclusion.
"High-Tech Monitors Often Miss Oil Pipeline Leaks," its Jan. 21 headline read. Alison Sider, the Journal's Texas energy and environment reporter, designed her own analysis of the same federal database and concluded: "The overwhelming majority of these pipeline spills, ruptures and leaks were discovered by somebody near the accident site."
Others have found the same thing. The Pipeline and Hazardous Materials Safety Administration, or PHMSA, hired a contractor to look at the situation. The findings were reported in late 2012 by The New York Times, a few months after Song's report.
"Study Finds Flaws in Pipeline Leak Detection Systems," the Times headline reported this time. That reporter, Dan Frosch, wrote: "Members of the general public are more likely to identify oil and gas spills than the pipeline companies' own leak detection systems."
It's all enough to bring back memories from the first season of Saturday Night Live, when the deadpan comic Chevy Chase poked fun at his own network, starting one Weekend Update sketch after another with the same headline: "Generalissimo Francisco Franco is still dead."
More seriously, it shows that for all the passage of time, the problem of how best to detect leaks endures. And it's at a time when the industry plans to build thousands of miles of new pipelines over the next five years—including the controversial Keystone XL, which would carry Canadian diluted bitumen (dilbit) across a critically important drinking water aquifer. TransCanada, the pipeline's builder, maintains that the line's remote sensors could quickly detect oil spills.
Just days before the 2010 dilbit spill in Michigan, that pipeline's operator, Enbridge, told PHMSA it could remotely detect a rupture in eight minutes. In reality, it took 17 hours for the company to confirm the spill, the Pulitzer-winning InsideClimate News investigation found.
After that Enbridge spill, Congress passed an overhaul of safety laws, but it did not set a very urgent schedule and important changes to federal rules are still under consideration by PHMSA, an underfunded and understaffed division of the U.S. Department of Transportation. The agency leaves much responsibility to the states, and much leeway to pipeline operators.
There should be fresh news on this front before too long. PHMSA had promised members of Congress to move forward on new leak regulations during 2013. It has received its reports, consulted with industry and held workshops and seminars. The deadline for an actual regulation, under the pipeline safety law that Obama signed in 2011, was to be "as soon as practical after Jan. 3, 2014," according to agency documents.
PHMSA spokesman Damon Hill wouldn't say when a new rule might appear, but he said the agency has not been spinning its wheels.
"These activities, in addition to routine information collections, are helping PHMSA to better understand, measure, and assess the performance of various pipeline leak detection types and issues and facilitate ways to reduce safety risks," he wrote in an email. "The information and research we are gathering will inform our decisions on any future regulatory and oversight activities related to leak detection and pipeline integrity issues."
Whatever PHMSA decides, what's clear from the media's repeated examinations of the data is that the technology now in wide use is far from perfect.
Although the journalists' methods were somewhat different, their results were consistent.
Song took a comprehensive approach, covering 10 years. She also ran separate analyses to compare larger oil spills with smaller ones.
Sider looked at the four most recent years, including 2013. She also analyzed a smaller subset of spills—those that began at a spot along the pipeline right-of-way, not inside pumping stations, and those where the leak could have been spotted either by control rooms or by a widespread technique called computational pipeline monitoring.
Both Sider and Song consulted with pipeline specialists as they did their research. And the reporters' bottom line was essentially identical—to trust in current leak-detection technology is to invite an unpleasant surprise.
And yet, despite the history of reports on the risks, people seem startled when they learn that leak detection systems find so few leaks.
When Gov. Jack Dalrymple of North Dakota found out that a huge leak of oil from a pipeline had soaked acres of farmland without being noticed until a farmer stumbled onto the mess, he said he was "shocked" that technology hadn't sounded the first alarm.