So much pipeline safety information!

Did you know we recently updated our briefing papers and added papers on three additional topics? These papers are a great source of ‘Pipeline Safety 101’ information – they are short and each focused on a single topic. They do not need to be read in order, though some basic understanding is helpful before reading about the more technical issues. Check them out if you haven’t already! Here is the full list:


Ignition of Natural Gas Transmission Pipelines

Question of the week:

A fairly large, 24 inch I believe, natural gas transmission pipeline recently failed here in Pennsylvania, and I was surprised that it did not ignite. I thought when natural gas pipelines ruptured they normally catch fire. Can you tell me why this one didn’t?


I think you are referring to the recent William’s Transco failure in Lycoming County, Pennsylvania that is described in a newspaper account here. We can’t really tell you why that one did not ignite, because few specifics are known about that failure at this time, but it is not unusual for gas transmission pipeline to rupture or leak without igniting. It all really depends whether the gas coming out finds an ignition source, which normally in an open area such as where this rupture occurred would be from sparks from the pipeline and rocks flying around due to the pressure of the escaping gas, or even static caused by the rapidly escaping gas.


We took a quick look at all the significant natural gas transmission pipeline incidents in the past 5 years and came up with this graph that shows for the various types of pipeline incidents whether they ignited or not. As you can see more often than not pipelines do not ignite when there are incidents, even when the lines completely rupture. We suspect that part of the reason people think they ignite more often is that when they do the incidents are quite spectacular and tend to make the news, whereas when they don’t ignite people hear much less about them.



It is not unusual for these types of gas transmission pipelines to be operating at 800-1000 psi or more. Compare that to your car tires that operate at 30-35 psi and you get a sense of how much pressure is in these pipelines. Even just the pressure of the gas escaping can cause some impressive damage as the picture below shows. This picture shows a similar pipeline rupture in Washington State where there was no ignition. The crater is just from the force of the gas escaping. Notice the piece of pipe in the upper right hand corner of the picture. That is how far the force of the escaping gas threw that piece of heavy pipe.




Hope that helps answer your question.

Santa Barbara Oil Spill

Question of the dayI see a pipeline just spilled oil into the ocean near Santa Barbara, California. What can you tell me about the company that operates that pipeline and their safety record?


Thanks for the question. The cause and the size of the spill are still being determined, and the clean up will go on for weeks we are sure.  Here is a quick analysis of the company that operates that pipelines

Quick Analysis Plains Pipeline L.P pipeline that failed in California

With the spill yesterday into the Pacific Ocean from a pipeline operated by a subsidiary of Plains All American Pipeline L.P. we have received a lot of calls about why the pipeline failed and how Plains safety record compares to other companies operating similar pipelines. To date we have not seen any information about why the pipeline may have failed, so speculating on that would serve little purpose. The Plains All American Pipeline L.P. system that failed is referred to in the Pipeline and Hazardous Materials Safety Administration’s (PHMSA) database as Plains Pipeline L.P, and their operator identification number is 300. Plains Pipeline L.P. is only part of the entire Plains All American Pipeline system which according to their website includes 17,800 miles of pipeline.

According to PHMSA Plains Pipeline L.P operates 6437 miles of hazardous liquid pipelines in 16 states, with 480 miles of it in California.[1] In the past ten years they have reported 175 pipeline incidents,[2] which caused nearly $24 million of property damage. Of those 175 incidents only 11 were in California. There have been 20 enforcement actions initiated against this company resulting in $284,500 in fines.[3] Of those enforcement actions none of them were for issues specific to California.

Here is some information about the incidents this system has had in the past ten years. Plains

The Pipeline Safety Trust today took a look at the incident data from PHMSA for the past 5 years (2009 – 2013) and compared Plains All American’s incident rates to the national average. Here is what we learned.





The number of Incidents reported to PHMSA for all hazardous liquid pipelines is increasing, but incidents for crude oil pipelines are increasing at a faster rate. The number of incidents on crude oil pipelines operated by Plains Pipeline L.P. follows this trend, and is increasing faster then the national average.





Since the mileage of pipelines has changed over the past few years we also normalized this analysis by looking at the number of incidents per mile of pipeline. We found that the rate nationally for crude oil pipelines is twice that of other types of hazardous liquid pipelines, and that the rate of incidents/mile of pipe for crude oil pipelines operated by Plains Pipeline L.P. was about 14% higher than the national average for crude oil pipelines.







Oil Pipelines by the Real Numbers

Reporters ask us a lot about numbers, and we see both accurate and misleading figures being thrown around in the press and even on unnamed official websites, so we’re expanding here on our February 10 blog post that touched on numbers. That post mentioned both numbers from the Pipeline and Hazardous Materials Safety Administration (PHMSA) and industry numbers published recently by API/AOPL (American Petroleum Institute / Association of Oil Pipe Lines) in their Pipeline Safety Excellence (PSE) initiative “Pipelines by the Numbers.” A fundamental concern the Trust has with these industry numbers is the lack of transparency about where their numbers come from.

The federal government through DOT- Pipeline and Hazardous Materials Safety Administration (PHMSA) collects information from pipeline operators annually and on incidents that occur (annual data and incident data, respectively). This data is fairly comprehensive and publicly available. PSE uses data from their own secret Pipeline Performance Tracking System (PPTS) with no public access to this data. The “Pipelines by the Numbers” report does not tell the reader what filters are used to pull the numbers, or why they differ so from the PHMSA incident data.

The data used by the Trust is typically based on PHMSA 20-year pipeline significant incident trends. In this example, we filtered that for onshore hazardous liquid (HL) pipelines (accessed on Feb 25, 2014), as we’re comparing our numbers to those of the liquid pipeline industry (API/AOPL). The Trust relies on the ‘significant incident’ dataset (rather than ‘all incidents’ or ‘serious incidents’) because we think it provides the most honest and transparent reflection of the incidents that show shortcomings in pipeline safety regulations and in operator safety cultures. Serious incidents only capture information when a death or serious injury occurs, and many catastrophic incidents are left out because the environmental destruction or personal property damage incurred is not enough to warrant the ‘serious’ categorization. The significant incidents dataset includes all incidents with $50,000 or more in total costs, measured in 1984 dollars, and including the value of the lost product; it also includes all serious incidents, any hazardous liquid release of 50 barrels or more, any HVL release of 5 barrels or more, and any liquid release resulting in a fire or explosion. We do not generally use the ‘all incidents’ dataset that captures the smaller accidents that occur, because reporting criteria for what a reportable incident is has changed quite a bit over time and can result in seemingly odd fluctuations when looking at all incidents.

We’re going to use this space to compare what we see in the PHMSA numbers, to what the hazardous liquid pipeline industry has published through their API/AOPL PSE initiative “Pipelines by the Numbers.” For shorthand purposes, when we say the word “industry” below in this post, we are referring to the hazardous liquid pipeline industry and the numbers put forth by API/AOPL.

1    FACT: More than 5 million gallons of hazardous liquids spilled in 2013 (including crude oil, refined products like gasoline, highly volatile liquids and others).

Industry states 99.999% of all hazardous liquid products are delivered safely each year, equating to all but .001% of the 14.9 billion barrels delivered in 2013. That means 5,009,524 gallons (we use PHMSA numbers here, but using industry numbers it would be 6,258,000 gallons) were spilled in 2013 in hazardous liquid pipeline incidents. (This is the one time we do use the ‘all incidents’ dataset to portray the volume that does, in fact, NOT get delivered safely each year.) Even if we filter the data for only the significant incidents, that figure remains close to 5 million gallons (4,978,706 gallons) spilled that year. By way of comparison, that’s equivalent to five spills the size of Marshall Michigan just for the volume spilled in 2013.

2    FACT: Greater than 67% of incidents since 2002 were caused by things within the operator’s control.

PHMSA collects incident data from operators, including a designation of what caused the incident, and that information is posted on PHMSA’s website. Among the causes are several that are entirely within the operator’s control: Corrosion, Incorrect Operation, Material/Weld/Equipment Failure, and Excavation Damage by the operator or its contractor. Together, these incident cause categories account for an average of 75% of all incidents since 2002 being within the operator’s control (1,622 total significant incidents from 2002-2014; 1,214 of those caused by things within the operators control). And things are not improving over time: every other year since 2002, more than 2/3 of significant incidents have been caused by things within the operators’ control.

3    FACT: 315 million residents are at risk of pipeline failures.

Industry reports that 315 million US consumers and workers benefit from pipelines daily – roughly the US Census population. Both are hyperboles adding nothing to what should be an important discussion about pipeline safety.

4    FACT: Over 50% increase in significant onshore HL pipeline incidents between 2001-2013.

Industry states a 50% DROP in incidents from 1999-2013. Frankly, we have no idea where they got this number, unless they only considered incidents that occurred off company property on Sundays (we’re kidding, kind of). 1999 had the highest number of significant pipeline incidents between 1995-2012 (142), so it was a convenient year to begin if looking to find a drop, but even starting at that relatively high point, there has been a 13% increase in significant hazardous liquid pipeline incidents.

5    FACT: 119 people killed or injured by onshore HL pipeline incidents since 1999 and $2.4 billion in property damages from pipeline spills in the past 10 years.

These are numbers they’d rather you not think about, preferring instead some variation of “everything is awesome”.

6    FACT: $141 million spent by oil & gas industry to lobby Congress last year.

Again, industry points out how much money the industry spends on pipeline maintenance and safety initiatives; and yet the industry spends 14 times the amount on lobbying than HL operators spend on pipeline safety research and development in any given year. Is “getting to zero” as important as “influencing” politicians?

7    FACT: 1.5 regulators review spill response plans for 192,000 miles of pipelines    and                             0 unannounced spill response drills held by federal regulators.

Under the Oil Pollution Act of 1990, HL pipeline operators are required to produce an ‘oil spill response plan’ or ‘facility response plan’ that details their preparations for a spill and submit this plan to PHMSA every 5 years. Only 1.5 PHMSA staff members are assigned to review these plans that cover 192,000+ miles of pipelines, the lowest by far of any of the four agencies who review these types of plans. The low staffing level results in PHMSA failing to require unannounced drills, as is required.

8    FACT: Corrosion caused more incidents in each of the last 3 years than in any year since 1997.

Industry claims that corrosion-caused pipeline incidents are down 75% since 1999. What??! PHMSA tracks corrosion as a cause (out of 7 overall cause categories) of incidents. We simply don’t understand where the industry claim could have come from. See for yourself; we’ve included the PHMSA data below so you can see both the numbers and percentages of incidents caused by corrosion each year. It’s remained a fairly steady 25%, but of late the total number of corrosion-caused incidents has been higher, not lower, than in previous years.


  1995 1996 1997 1998 1999 2000 2001 2002 2003 2004
# onshore HL sig. incidents 154 171 153 131 142 128 107 129 122 125
# corrosion-caused subset 32 51 44 31 24 29 32 34 28 35
% corrosion-caused 20.8% 29.8% 28.8% 23.7% 16.9% 22.7% 29.9% 26.4% 23.0% 28.0%
  2005 2006 2007 2008 2009 2010 2011 2012 2013  
# onshore HL sig. incidents 120 104 107 119 107 119 137 126 160  
# corrosion-caused subset 27 32 26 31 26 27 38 36 38  
% corrosion-caused 22.5% 30.8% 24.3% 26.1% 24.3% 22.7% 27.7% 28.6% 23.8%  


Everything is Awesome!…. or is it?

This recent tweet by Brigham McCown (former head of PHMSA in the Bush Administration) about rail and pipeline safety gave us pause. It praised government agency oversight and questioned congressional initiatives to push for greater safety, begging the question: Are we missing something? Should we as a country really be patting ourselves on the back for successful pipeline safety improvements?

In a more recent tweet he stated that “#pipelines have decreased accidents by 50%” – the same thing that PHMSA’s Deputy Director for Governmental, International and Public Affairs told media in Montana when talking about the recent spill into the Yellowstone River. That would be great if it were true, but it’s just not.

If only we were more technologically advanced, we would create an animation of Mr. McCown and the heads of the large oil and gas industry groups singing “Everything is Awesome” and touting the ‘continuous improvement’ and ‘excellent job reducing incidents, spills, injuries’ that have apparently happened of late while the rest of us have been watching oil spill into the Yellowstone and fireballs threaten more pipeline neighbors.

We would love to agree with Mr. McCown, but cannot reconcile his sentiment with what we know about the reality of pipeline safety today and the recent record of pipeline failures. He specifically called into question the actions of Congressman DeFazio, later suggesting he should “stand down.” Before we start in on our own perspective, here is part of what Peter DeFazio had to say to the Department of Transportation’s (DOT) Inspector General (this excerpt comes after discussion of the Pipeline and Hazardous Materials Safety Administration’s (PHMSA) slow action and failure to issue new rules regarding safety of rail tank cars):

In multiple pipeline accident investigations over the last 15 years, the NTSB [National Transportation Safety Board] has identified the same persistent issues, most of which PHMSA has failed to address on its own accord. Each and every time, Congress has been forced to require PHMSA to take action, most recently in the Pipeline Safety, Regulatory Certainty, and Job Creation Act of 2011 (P.L. 112-90). Three years later, almost none of the important safety measures in the Act have been finalized.

Finalizing these rules is imperative; our Nation’s vast 2.5 million-mile pipeline network is aging. According to PHMSA, more than 50 percent of these pipelines were constructed in the 1950s and 1960s. The potential for catastrophic accidents is not a matter of if, but when. DOT must be prepared to address any pipeline or hazardous material safety deficiencies now, not just when we mandate action.

For these reasons, I am concerned with the agency’s ability to address significant safety issues and am requesting an audit of PHMSA’s pipeline and hazardous materials safety programs. Specifically, I request an evaluation of the agency’s effectiveness in addressing: congressional mandates, and NTSB, Government Accountability Office, and Office of Inspector General recommendations in a timely manner; the process PHMSA utilizes for implementing such mandates and recommendations; the sufficiency of PHMSA’s efforts to coordinate with the modal administrations and address safety concerns raised by those administrations; and any impediments to agency action.

We think Mr. DeFazio is spot-on. Not only because of what needs to be done to prevent future incidents, but also because of what has already happened. In the first ten days of 2015 alone, there were 11 reported pipeline incidents, 4 of them significant, and all of those four were from causes within the operator’s control. It is too early to know the total damage done, but we do know that one of the incidents not yet included in the PHMSA database resulted in an estimated 30,000 gallons of crude oil spilling into the Yellowstone River in Montana and polluting the drinking water supply for the town of Glendive, requiring emergency changes to the water treatment plant to filter out benzene from the spill.

Since 1995, the number of significant pipeline incidents has risen by 3% – going from a rolling 3-yr average of 275 (1995-1997) to 283 (2012-2014). And the most prevalent cause of significant incidents is something within the operators’ control: material/weld/equipment failure. In fact, according to PHMSA data, the majority of significant incidents in the past 10 years have been caused by three things within the operator’s control: corrosion, incorrect operation and material/weld/equipment failure.

Operators will say that many more miles of pipelines have been installed in recent years, and they are still the safest form of oil and gas transportation. But even with the incident data normalized by miles of pipeline, we’ve still got lots of improving to do. 


You will see that throughout this piece, we refer to “significant incidents.” The types of incidents are divided by PHMSA into “serious”, “significant” and “all”. Significant incidents include:

  1. Fatality or injury requiring in-patient hospitalization
  2. $50,000 or more in total costs, measured in 1984 dollars
  3. Highly volatile liquid releases of 5 barrels or more or other liquid releases of 50 barrels or more
  4. Liquid releases resulting in an unintentional fire or explosion.

Serious incidents are a subset of significant incidents and only include those causing a fatality or in-patient hospitalization. And serious incidents have indeed decreased over the past 20 years, and we celebrate that. Mr. McCown and Ms. Klinger were perhaps thinking only of serious incidents when making their claims. But looking at serious incidents only does not tell the whole story about pipeline safety. In the case of the Yellowstone spills or the Marshall, Michigan spill, no one was killed or hospitalized, so they were significant incidents. The records of serious incidents will not tell you about catastrophes even of that magnitude. The difference between whether an incident is categorized as serious or significant is sometimes as random as whether someone whose house was destroyed by pipeline explosion happened to be home at the time. Had they been home, the incident would have been serious. They weren’t, so the incident was significant. The pipeline wasn’t safer just because no one was killed. They just weren’t home. If we only looked at serious incidents trends, we’d miss learning from many significant incidents and mislead the public into thinking fewer incidents were happening.

Another metric that the pipeline industry likes to tout is that 99.999% of the oil and refined products shipped by pipeline arrives safely at its destination. Wow! That is really awesome, you say, starting to hum along with Mr. McCown and the crew. But hold on. How much is that .001%, you wonder? Well, as API/AOPL tell us in their new “Pipelines by the Numbers“, 14.9 billion barrels of oil and refined products were shipped by pipeline in 2013. So that .001% works out to be 149 thousand barrels, or 6.258 million gallons of oil and refined products that did NOT make it safely to its destination.

6.258 million gallons spilled from pipelines. That’s the equivalent of:

One spill the size of the Enbridge spill at Marshall, Michigan EVERY OTHER MONTH for a year;


Two spills the size of the 1999 Bellingham, Washington incident EVERY MONTH for a year;


One spill the size of the Exxon Valdez every other year.

Our point (in case you missed it): It’s a lot of oil and refined product spilling and leaking from pipelines every year. “#Pipelines” have NOT reduced accidents by 50%. Things are NOT awesome. Pipelines may be the safest means we have, but they’re not nearly safe enough. And it’s Congress’ duty to call out the regulator when things aren’t good enough. That’s what Representative DeFazio did, and that’s why.