Posts

Map of pipeline incidents across the US

Pipeline Incidents Continue to Impact Residents

Pipelines play a major role in the oil and gas extraction industry, allowing for the transport of hydrocarbons from well sites to a variety of infrastructure, including processing plants, petrochemical facilities, power generation plants, and ultimately consumers. There are more than 2.7 million miles of natural gas and hazardous liquid pipelines in the United States, or more than 11 times the distance from Earth to the moon.

With all of this infrastructure in place, pipelines are inevitably routed close to homes, schools, and other culturally or ecologically important locations. But how safe are pipelines, really? While they are typically buried underground and out of sight, many residents are concerned about the constant passage of volatile materials through these pipes in close proximity to these areas, with persistent but often unstated possibility that something might go wrong some day.

Safety talking points

In an attempt to assuage these fears, industry representatives and regulators tend to throw around variants of the word “safe” quite a bit:

Pipelines are the safest and most reliable means of transporting the nation’s energy products.
— Keith Coyle, Marcellus Shale Coalition

Although pipelines exist in all fifty states, most of us are unaware that this vast network even exists. This is due to the strong safety record of pipelines and the fact that most of them are located underground. Installing pipelines underground protects them from damage and helps protect our communities as well.
— Pipeline and Hazardous Materials Safety Administration (PHMSA)

Pipelines are an extremely safe way to transport energy across the country.
Pipeline 101

Knowing how important pipelines are to everyday living is a big reason why we as pipeline operators strive to keep them safe. Pipelines themselves are one of the safest ways to transport energy with a barrel of crude oil or petroleum product reaching its destination safely by pipeline 99.999% of the time.
American Petroleum Institute

But are pipelines really safe?

Given these talking points, the general public can be excused for being under the impression that pipelines are no big deal. However, PHMSA keeps records on pipeline incidents in the US, and the cumulative impact of these events is staggering. These incidents are broken into three separate reports:

  1. Gas Distribution (lines that take gas to residents and other consumers),
  2. Gas Transmission & Gathering (collectively bringing gas from well sites to processing facilities and distant markets), and
  3. Hazardous Liquids (including crude oil, refined petroleum products, and natural gas liquids).

Below in Table 1 is a summary of pipeline incident data from 2010 through mid-November of this year. Of note: Some details from recent events are still pending, and are therefore not yet reflected in these reports.

Table 1: Summary of pipeline incidents from 1/1/2010 through 11/14/2018

Report Incidents Injuries Fatalities Evacuees Fires Explosions Damages ($)
Gas Distribution 934 473 92 18,467 576 226 381,705,567
Gas Transmission & Gathering 1,069 99 24 8,614 121 51 1,107,988,837
Hazardous Liquids 3,509 24 10 2,471 111 14 2,606,014,109
Totals 5,512 596 126 29,552 808 291 4,095,708,513

Based on this data, on average each day in the US 1.7 pipeline incidents are reported (a number in line with our previous analyses), requiring 9 people to be evacuated, and causing almost $1.3 million in property damage. A pipeline catches fire every 4 days and results in an explosion every 11 days. These incidents result in an injury every 5 days, on average, and a fatality every 26 days.

Data shortcomings

While the PHMSA datasets are extremely thorough, they do have some limitations. Unfortunately, in some cases, these limitations tend to minimize our understanding of the true impacts. A notable recent example is a series of explosions and fires on September 13, 2018 in the towns of Lawrence, Andover, and North Andover, in the Merrimack Valley region of Massachusetts. Cumulatively, these incidents resulted in the death of a young man and the injuries to 25 other people. There were 60-80 structure fires, according to early reports, as gas distribution lines became over-pressurized.

The preliminary PHMSA report lists all of these Massachusetts fires as a single event, so it is counted as one fire and one explosion in Table 1. As of the November 14 download of the data, property damage has not been calculated, and is listed as $0. The number of evacuees in the report also stands at zero. This serves as a reminder that analysis of the oil and gas industry can only be as good as the available data, and relying on operators to accurately self-report the full extent of the impacts is a somewhat dubious practice.

View map fullscreen | How FracTracker maps work

This map shows pipeline incidents in the US from 1/1/2010 through 11/14/2018. Source: PHMSA. One record without coordinates was discarded, and 10 records had missing decimal points or negative (-) signs added to the longitude values. A few obvious errors remain, such as a 2012 incident near Winnipeg that should be in Texas, but we are not in a position to guess at the correct latitude and longitude values for each of the 5,512 incidents.

Another recent incident occurred in Center Township, a small community in Beaver County, Pennsylvania near Aliquippa on September 10, 2018. According to the PHMSA Gas Transmission & Gathering report, this incident on the brand new Revolution gathering line caused over $7 million in damage, destroying a house and multiple vehicles, and required 49 people to evacuate. The incident was indicated as a fire, but not an explosion. However, reporting by local media station WPXI quoted this description from a neighbor:

A major explosion, I thought it was a plane crash honestly. My wife and I jumped out of bed and it was just like a light. It looked like daylight. It was a ball of flame like I’ve never seen before.

From the standpoint of the data, this error is not particularly egregious. On the other hand, it does serve to falsely represent the overall safety of the system, at least if we consider explosions to be more hazardous than fires.

Big picture findings

Comparing the three reports against one another, we can see that the majority of incidents (64%) and damages (also 64%) are caused by hazardous liquids pipelines, even though the liquids account for less than 8% of the total mileage of the network. In all of the other categories, however, gas distribution lines account for more than half of the cumulative damage, including injuries (79%), deaths (73%), evacuees (62%), fires (71%), and explosions (78%). This is perhaps due to the vast network (more than 2.2 million miles) of gas distribution mains and service lines, as well as their nature of taking these hazardous products directly into populated areas. Comparatively, transmission and hazardous liquids lines ostensibly attempt to avoid those locations.

Is the age of the pipeline a factor in incidents?

Among the available attributes in the incident datasets is a field indicating the year the pipeline was installed. While this data point is not always completed, there is enough of a sample size to look for trends in the data. We determined the age of the pipe by subtracting the year the pipe was installed from the year of the incident, eliminating nonsensical values that were created when the pipeline age was not provided. In the following section, we will look at two tables for each of the three reports. The first table shows the cause of the failure compared to the average age, and the second breaks down results by the content that the pipe was carrying. We’ll also include a histogram of the pipe age, so we can get a sense of how representative the average age actually is within the sample.

A. Gas distribution

Each table shows some fluctuation in the average age of pipeline incidents depending on other variables, although the variation in the product contained in the pipe (Table 3) are minor, and may be due to relatively small sample sizes in some of the categories. When examining the nature of the failure in relation to the age of the pipe (Table 2), it does make sense that incidents involving corrosion would be more likely to afflict older pipelines, (although again, the number of incidents in this category is relatively small). On average, distribution pipeline incidents occur on pipes that are 33 years old.

When we look at the histogram (Figure 1) for the overall distribution of the age of the pipeline, we see that those in the first bin, representing routes under 10 years of age, are actually the most frequent. In fact, the overall trend, excepting those in the 40 t0 50 year old bin, is that the older the pipeline, the fewer the number of incidents. This may reflect the massive scale of pipeline construction in recent decades, or perhaps pipeline safety protocol has regressed over time.

Pipeline incidents charting

Figure 1. Age of pipeline histogram for gas distribution line incidents between 1/1/2010 and 11/14/2018. Incidents where the age of the pipe is unknown are excluded.

B. Gas Transmission & Gathering

Transmission & Gathering line incidents occur on pipelines routes that are, on average, five years older than their distribution counterparts. Corrosion, natural force damage, and material failures on pipes and welds occur on pipelines with an average age above the overall mean, while excavation and “other outside force” incidents tend to occur on newer pipes (Table 4). The latter category would include things like being struck by vehicles, damaged in wildfires, or vandalism. The contents of the pipe does not seem to have any significant correlation with the age of the pipe when we take sample size into consideration (Table 5).

The histogram (Figure 2) for the age of pipes on transmission & gathering line incidents below shows a more normal distribution, with the noticeable exception of the first bin (0 to 10 years old) ranking second in frequency to the fifth bin (40 to 50 years old).

It is worth mentioning that, “PHMSA estimates that only about 5% of gas gathering pipelines are currently subject to PHMSA pipeline safety regulations.” My correspondence with the agency verified that the remainder is not factored into their pipeline mileage or incident reports in any fashion. Therefore, we should not consider the PHMSA data to completely represent the extent of the gathering line network or incidents that occur on those routes.

Pipeline incidents chart

Figure 2. Age of pipeline histogram for transmission & gathering line incidents between 1/1/2010 and 11/14/2018. Incidents where the age of the pipe is unknown are excluded.

C. Hazardous Liquids

The average incident on hazardous liquid lines occurs on pipelines that are 27 years old, which is 6 years younger than for distribution incidents, and 11 years younger than their transmission & gathering counterparts. This appears to be heavily skewed by the equipment failure and incorrect operation categories, both of which occur on pipes averaging 15 years old, and both with substantial numbers of incidents. On the other hand, excavation damage, corrosion, and material/weld failures tend to occur on pipes that are at least 40 years old (Table 6).

In terms of content, pipelines carrying carbon dioxide happen on pipes that average just 11 years old, although there are not enough of these incidents to account for the overall departure from the other two datasets (Table 7).

The overall shape of the histogram (Figure 3) is similar to that of transmission & gathering line incidents, except that the first bin (0 to 10 years old) is by far the most frequent, with more than 3 and a half times as many incidents as the next closest bin (4o to 50 years old). Operators of new hazardous liquid routes are failing at an alarming rate. In descending order, these incidents are blamed on equipment failure (61%), incorrect operation (21%), and corrosion (7%), followed by smaller amounts in other categories. The data indicate that pipelines installed in previous decades were not subject to this degree of failure.

Pipeline incidents charting

Figure 3. Age of pipeline histogram for hazardous liquid line incidents between 1/1/2010 and 11/14/2018. Incidents where the age of the pipe is unknown are excluded.

Conclusions

When evaluating quotes, like those listed above, that portray pipelines as a safe way of transporting hydrocarbons, it’s worth taking a closer look at what they are saying.

Are pipelines the safest way of transporting our nation’s energy products? This presupposes that our energy must be met with liquid or gaseous fossil fuels. Certainly, crude shipments by rail and other modes of transport are also concerning, but movements of solar panels and wind turbines are far less risky.

Does the industry have the “strong safety record” that PHMSA proclaims? Here, we have to grapple with the fact that the word “safety” is inherently subjective, and the agency’s own data could certainly argue that the industry is falling short of reasonable safety benchmarks.

And what about the claim that barrels of oil or petroleum products reach their destination “99.999% of the time? First, it’s worth noting that this claim excludes gas pipelines, which account for 92% of the pipelines, even before considering that PHMSA only has records on about 5% of gas gathering lines in their pipeline mileage calculations. But more to the point, while a 99.999% success rate sounds fantastic, in this context, it isn’t good enough, as this means that one barrel in every 100,000 will spill.

For example, the Dakota Access Pipeline has a daily capacity of 470,000 barrels per day (bpd). In an average year, we can expect 1,715 barrels (72,030 gallons) to fail to reach its destination, and indeed, there are numerous spills reported in the course of routine operation on the route. The 590,000 bpd Keystone pipeline leaked 9,700 barrels (407,400 gallons) late last year in South Dakota, or what we might expect from four and a half years of normal operation, given the o.001% failure rate. In all, PHMSA’s hazardous liquid report lists 712,763 barrels (29.9 million gallons) were unintentionally released, while an additional 328,074 barrels (13.8 million gallons) were intentionally released in this time period. Of this, 284,887 barrels (12 million gallons) were recovered, meaning 755,950 barrels (31.7 million gallons) were not.

Beyond that, we must wonder whether the recent spate of pipeline incidents in new routes is a trend that can be corrected. Between the three reports, 1,283 out of the 3,853 (32%) incidents occurred in pipelines that were 10 years old or younger (where the year the pipeline’s age is known). A large number of these incidents are unforced errors, due to poor quality equipment or operator error.

One wonders why regulators are allowing such shoddy workmanship to repeatedly occur on their watch.


By Matt Kelso, Manager of Data and Technology, FracTracker Alliance

Pennsylvania Pipelines map by FracTracker Alliance

Pennsylvania Pipelines and Pollution Events

When people think about oil and gas extraction in Pennsylvania, they think about the tens of thousands of oil and gas wells in the state. It makes sense, because that’s where the process starts. However, while oil and other liquids can be shipped in tanker trucks, all of the producing gas wells in the state – whether they are small conventional wells or the giants of the Marcellus and Utica – must be connected by a network of pipelines.

Moving hydrocarbons from the well to processing facilities to power plants and residential customers all occurs within this giant midstream system, and the cumulative impact that pipelines have on the state is formidable. Let’s take a closer look at where the oil and gas pipelines are located in PA, their safety records, and major data gaps. Additionally, we’ve made available a detailed, interactive map of Pennsylvania pipelines and other important features such as water crossings.

Pipeline routes are everywhere in Pennsylvania

According to the Pipeline and Hazardous Materials Safety Administration (PHMSA), there were 92,407 miles of pipelines carrying natural gas and liquid petroleum products in Pennsylvania in 2017. That distance is equivalent to 151 round trips between Philadelphia and Pittsburgh on the Pennsylvania Turnpike, or more than three trips around the globe at the equator. This figure includes 78,022 miles of distribution lines (which takes gas from public utilities to consumers), 10,168 miles of transmission lines (which move gas between various processing facilities), 3,111 miles of petroleum liquid routes, and 1,105 miles of natural gas gathering lines (which take the gas from wells to midstream processing facilities).

Of note – The last category’s estimate is almost certainly a drastic underestimation. As of June 7th, there were 3,781 unconventional well pads in Pennsylvania, according the Pennsylvania Department of Environmental Protection (DEP), and all of the pads need to be connected to gathering lines. A 2014 report by the Nature Conservancy estimates that 19 acres of land are cleared for each well pad, which would work out to 3.1 miles of gathering lines for a typical 50-foot right-of-way. Multiplied out, 3,781 wells pads would require a total of 11,721 miles of gathering lines – well over PHMSA’s estimate of a 1,105 miles (See Table 1 for estimate comparisons).

Table 1. Varying estimates of gathering lines in Pennsylvania.*

Source

Unconventional Well Pads

Average Gathering Line Length (Miles) Statewide Total Estimated Miles
Nature Conservancy 3,781 3.1 11,721
Bradford County 3,781 3.5 13,234
PHMSA  3,781  0.3 1,105

*Estimates based on Nature Conservancy and Bradford County data are based on calculating the average length of segments, then multiplying by the number of well pads in the state to find the statewide total. The PHMSA estimate was calculated in reverse, by dividing the purported total of gathering lines by the number of well pads to find the average mileage.

Early map of gathering lines in Bradford County, PA by FracTracker (Pennsylvania Pipelines)

Figure 1: Location of gathering lines (2014) and oil and gas wells (2018) in Bradford County, Pennsylvania. Note the pockets of newer wells that are not connected to the older gathering line network.

In 2014, the FracTracker Alliance digitized a published map of gathering lines in Bradford County, allowing us to analyze the data spatially (Figure 2). These efforts yield similar results, with gathering lines averaging 3.5 miles in length. Not counting segments of transmission lines included in the data, such as Stagecoach, Sunoco, and Kinder Morgan’s Tennessee Gas Pipeline, there were 1,003 miles of gas gathering lines just in Bradford County in 2014.

Almost all of this data is based only on unconventional oil and gas activity, and therefore ignores the more than 96,000 conventional oil and gas (O&G) wells active in the state. We do not have a reasonable estimate on the average length of gathering line segments are for this network. It is reasonable to assume that they tend to be shorter, as conventional wells are often closer together than unconventional well pads, but they must still network across vast portions of the state.

Table 2. Estimated length of gathering lines for conventional wells in Pennsylvania by variable average lengths

Average Length (Miles) Conventional Wells Total Miles
0.5 96,143 48,072
1.0 96,143 96,143
1.5 96,143 144,215
2.0 96,143 192,286
2.5 96,143 240,358
3.0 96,143 288,429

If the average gathering line for conventional wells in Pennsylvania is at least 1 mile in length, then the total mileage of gathering lines would exceed all other types of gas and petroleum pipelines in the state. Conversely, for the PHMSA figure of 1,105 miles to be accurate, the average gathering line for all conventional wells and unconventional well pads in Pennsylvania would be 0.011 miles, or only about 58 feet long.

Pipelines are dangerous

As pipelines impact residents in many ways, there are numerous reason why communities should try to understand their impacts – including basic planning, property rights, sediment runoff into streams, to name a few. Perhaps the most significant reason, however, is the potential for harmful incidents to occur, which are more common than anyone would like to think (See Table 3). Some of these incidents are quite serious, too.

Table 3. Nationwide pipeline incidents statistics from PHMSA from January 1, 2010 through July 13, 2018

Report Events Fatalities Injuries Explosions Evacuees Total Damages
Gas Distribution 909 92 432 220 16,949 $348,511,528
Gas Transmission / Gathering 1,031 23 94 49 8,557 $1,085,396,867
Hazardous Liquids 3,368 10 24 14 2,467 $2,531,839,207
Grand Total 5,308 125 550 283 27,973 $3,965,747,602

As of the July 13, 2018 download date, the PHMSA report covers 3,116 days.

Incidents Per Day

This means that nationally per day there are 1.7 pipeline incidents, almost 9 people evacuated, and $1,272,704 in damages, including the loss of released hydrocarbons.

On average, there is a fatality every 25 days, an injury every six days, and an explosion every 11 days. The location of those explosions obviously has a lot to do with the casualty count and aggregate property damage.

How do Pennsylvania pipelines hold up? As one might expect from a state with so many pipelines, Pennsylvania’s share of these incidents are significant (See Table 4).

Table 4. Pennsylvania pipeline incidents statistics from PHMSA from January 1, 2010 through July 13, 2018

Report Events Fatalities Injuries Explosions Evacuees Total Damages
Gas Distribution 29 8 19 12 778 $6,769,061
Gas Transmission / Gathering 30 0 2 2 292 $51,048,027
Hazardous Liquids 49 0 0 1 48 $9,115,036
Grand Total 108 8 21 15 1,118 $66,932,124

Within Pennsylvania, an incident is reported to PHMSA every 29 days, an injury or fatality can be expected every 107 days, and the daily average of property damage is $21,480.

The issue with under-reported gathering lines notwithstanding, PHMSA lists Pennsylvania with 92,407 miles of combined gas and hazardous liquid pipelines, which is roughly 3.3% of the nationwide total, and there is no reason to believe that PHMSA’s issue with accounting for gathering lines is unique to the Keystone State.

Just 2% of the total number of incidents are in Pennsylvania. In terms of impacts, however, the state has seen more than its fair share – with 6.4% of fatalities, 3.8% of injuries, 5.3% of explosions, and 3.9% of evacuations. Property damage in Pennsylvania accounts for just 1.7% of the national total, making it the only category examined above for which its share of impacts is less than expected, based on total pipeline miles.

Pipeline location data not widely available

Pipeline data is published from a variety of public agencies, although almost none of it is really accessible or accurate.

For example the Department of Homeland Security (DHS) publishes a number of energy-related datasets. While they do not publish gas pipelines, they do have a 2012 dataset of natural gas liquid routes, which is a significant portion of the hazardous liquid inventory. From an analytical point of view, however, this dataset is essentially worthless. Many of these pipelines are so generalized that they don’t make a single bend for multiple counties, and the actual location of the routes can be miles from where the data are represented. Communities cannot use this as a tool to better understand how pipelines interact with places that are important to them, like schools, hospitals, and residential neighborhoods. The dataset is also incomplete – the original Mariner East natural gas pipeline, which has been around for decades, isn’t even included in the dataset.

Screenshot from PHMSA's public pipeline viewer

Figure 2: This text appears to viewers of PHMSA’s public pipeline viewer.

Another data source is PHMSA’s National Pipeline Mapping System Public Viewer. While this source is rich in content, it has several intentional limitations that thwart the ability of the public to accurately analyze the pipeline network and understand potential impacts:

  1. Data can only be accessed one county at a time, which is impractical for long interstate transmission routes,
  2. Data can not be be downloaded, and
  3. The on-screen representation of the routes disappears when users zoom in too far.

Within Pennsylvania, the Department of Environmental Protection (DEP) maintains the Pennsylvania Pipeline Portal, which contains a lot of information about various recent pipeline projects. However, with the sole exception of the Mariner East II project, the agency does not provide any geospatial data for the routes. The reason for this is explained on the Mariner East II page:

These shapefiles are the GIS data layers associated with the permits that have been submitted for the proposed pipeline project. These shapefiles are not required as part of a permit application and are not commonly submitted but were provided to the Department by Sunoco Pipeline, L.P.

The files were accepted by the Department to aid in the review of the application material given the large scale of the project. The shapefiles ease the review by displaying some information contained in the hardcopy of the plans and application in a different format.

The Department of Conservation and Natural Resources (DCNR) does make oil and gas infrastructure data available, including pipelines, where it occurs on state forest land.

Pennsylvania Pipelines Map

Considering the risks posed by pipelines, their proliferation in Pennsylvania, and this critical juncture in their development with an implicit opportunity to document impacts, FracTracker believes it is important now to develop an accurate interactive statewide map of these projects, fortify it with essential data layers, and facilitate citizen reporting of the problems that are occurring.

Other than the Mariner East II route and the state forest data available from DCNR, all of the pipeline routes on our Pennsylvania Pipeline Map, below, have been painstakingly digitized – either from paper maps, PDFs, or other digital media – to make geospatial data that can analyzed by interacting with other datasets. These layers are only as good as their sources, and may not be exact in some cases, but they are orders of magnitude better than data produced by public agencies such as DHS.

Figure 3: FracTracker’s Pennsylvania Pipeline Map. View fulll screen to explore map further, view water crossings, and other details not visible at the statewide map view.

Data Layers on Pennsylvania Pipelines Map

  • Incidents

    PHMSA incidents (7-13-2018). Pipeline incidents that were reported to the Pipeline and Hazardous Material Safety Administration. These reports contain significant information about the incidents, including location coordinates, and are shown on the map with white circles.

    Note that a few of the location coordinates appear to be erroneous, as two reports appear outside of the state boundary.

  • Spills

    Mariner East II – Inadvertent Returns (6-1-2018). This data layer shows inadvertent returns – or spills – related to the construction of the Mariner East II pipeline. This is a combination of two reports, including one where the spills that impacted waterways, and those categorized as upland spills. These are represented on the map by orange dots that vary in size depending on the amount of fluid that spilled. Some of the locations were provided as latitude / longitude coordinates, while others are estimates based on the description. In a few cases, the latitude value was adjusted to intersect the pipeline route. In each case, the adjusted location was in the correct county and municipality.

  • Water Crossings

    Known Stream & Wetland Crossings (2018). This shows the locations where the known pipeline routes intersect with streams and other wetlands on the National Wetland Inventory. These are organized by our four pipeline layers that follow, including FracTracker Vetted Pipelines (1,397 crossings), DCNR Pipelines (184 crossings), PHMSA Gas Pipelines (6,767 crossings), and Bradford County Gathering Lines (867 crossings). These crossings are shown as diamonds that match the colors of the four listed pipeline layers.

  • Vetted Pipelines

    FracTracker Vetted Pipelines (2018). This pipeline layer is an aggregation of pipeline routes that have been digitized in recent years. Much of this digitization was performed by the FracTracker Alliance, and it is an available layer on our mobile app. These are largely newer projects, and contain some routes, such as the Falcon Ethane Pipeline System, that have not been built yet. In some cases, multiple versions of the pipeline routes are printed, and we may not have the final version of the route in all circumstances. FracTracker Vetted Pipelines are represented with a red line.

  • DCNR Pipelines

    DCNR Pipelines (2018). This includes pipeline routes on state forest lands, and is shown as green lines on the map.

  • PHMSA Pipelines

    PHMSA Gas Pipelines (2018). This includes data digitized from the PHMSA Public Pipeline Viewer. This source contains gas and liquid pipelines, but only gas pipelines are included in this analysis. These routes are shown in a bright purplish pink color.

  • Bradford Lines

    Bradford County Gathering Lines (2014). This layer was digitized by the FracTracker Alliance after Bradford County published a printed map of gathering lines within the county in 2014. It is the only county in Pennsylvania that we have gathering line data for, and it is shown on the map as a yellow line.

  • Nearby Waterways

    Streams & Wetlands with 1/2 Mile of Pipelines (2018). This clipped layer of the National Wetlands Inventory is provided for visual reference of the wetlands near known pipeline routes. Due to the large amount of data, this layer is only visible when users zoom in to a scale of 1:500,000, or about the size of a large county.


By Matt Kelso, Manager of Data and Technology

This article is the first in a two-part series on Pennsylvania pipelines. Stay tuned!

Shell Pipeline - Not Quite the Good Neighbor

Shell Pipeline: Not Quite the “Good Neighbor”

In August 2016, Shell Pipeline announced plans to develop the Falcon Ethane Pipeline System, a 97-mile pipeline network that will carry more than 107,000 barrels of ethane per day through Pennsylvania, West Virginia, and Ohio, to feed Shell Appalachia’s petrochemical facility currently under construction in Beaver County, PA.

FracTracker has covered the proposed Falcon pipeline extensively in recent months. Our Falcon Public EIA Project explored the project in great detail, revealing the many steps involved in risk assessments and a range of potential impacts to public and environmental health.

This work has helped communities better understand the implications of the Falcon, such as in highlighting how the pipeline threatens drinking water supplies and encroaches on densely populated neighborhoods. Growing public concern has since convinced the DEP to extend public comments on the Falcon until April 15th, as well as to host three public meetings scheduled for early April.

Shell’s response to these events has invariably focused on their intent to build and operate a pipeline that exceeds safety standards, as well as their commitments to being a good neighbor. In this article, we investigate these claims by looking at federal data on safety incidents related to Shell Pipeline.

Contrary to claims, records show that Shell’s safety record is one of the worst in the nation.

The “Good Neighbor” Narrative

Maintaining a reputation as a “good neighbor” is paramount to pipeline companies. Negotiating with landowners, working with regulators, and getting support from implicated communities can hinge on the perception that the pipeline will be built and operated in a responsible manner. This is evident in cases where Shell Pipeline has sold the Falcon in press releases as an example of the company’s commitment to safety in public comments.

Figure 1. Shell flyer

A recent flyer distributed to communities in the path of the Falcon, seen in Figure 1, also emphasizes safety, such as in claims that “Shell Pipeline has a proven track record of operating safely and responsibility and remains committed to engaging with local communities regarding impacts that may arise from its operations.”

Shell reinforced their “good neighbor” policy on several occasions at a recent Shell-sponsored information meeting held in Beaver County, stating that, everywhere they do business, Shell was committed to the reliable delivery of their product. According to project managers speaking at the event, this is achieved through “planning and training with first responders, preventative maintenance for the right-of-way and valves, and through inspections—all in the name of maintaining pipeline integrity.”

Shell Pipeline also recently created an informational website dedicated to the Falcon pipeline to provide details on the project and emphasize its minimal impact. Although, curiously, Shell’s answer to the question “Is the pipeline safe?” is blank.

U.S. Pipeline Incident Data

Every few years FracTracker revisits data on pipeline safety incidents that is maintained by the Pipeline and Hazardous Materials Safety Administration (PHMSA). In our last national analysis we found that there have been 4,215 pipeline incidents resulting in 100 reported fatalities, 470 injuries, and property damage exceeding $3.4 billion.

These numbers were based on U.S. data from 2010-2016 for natural gas transmission and gathering pipelines, natural gas distribution pipelines, and hazardous liquids pipelines. It is also worth noting that incident data are heavily dependent on voluntary reporting. They also do not account for incidents that were only investigated at the state level.

Shell Pipeline has only a few assets related to transmission, gathering, and distribution lines. Almost all of their pipeline miles transport highly-volatile liquids such as crude oil, refined petroleum products, and hazardous liquids such as ethane. Therefore, to get a more accurate picture of how Shell Pipeline’s safety record stacks up to comparable operators, our analysis focuses exclusively on PHMSA’s hazardous liquids pipeline data. We also expanded our analysis to look at incidents dating back to 2002.

Shell’s Incident Record

In total, PHMSA data show that Shell was responsible for 194 pipeline incidents since 2002. These incidents spilled 59,290 barrels of petrochemical products totaling some $183-million in damages. The below map locates where most of these incidents occurred. Unfortunately, 34 incidents have no location data and so are not visible on the map. The map also shows the location of Shell’s many refineries, transport terminals, and off-shore drilling platforms.

Open the map fullscreen to see more details and tools for exploring the data.


View Map Fullscreen | How FracTracker Maps Work

Incidents Relative to Other Operators

PHMSA’s hazardous liquid pipeline data account for more than 350 known pipeline operators. Some operators are fairly small, only maintaining a few miles of pipeline. Others are hard to track subsidiaries of larger companies. However, the big players stand out from the pack — some 20 operators account for more than 60% of all pipeline miles in the U.S., and Shell Pipeline is one of these 20.

Comparing Shell Pipeline to other major operators carrying HVLs, we found that Shell ranks 2nd in the nation in the most incidents-per-mile of maintained pipeline, seen in table 1 below. These numbers are based on the total incidents since 2002 divided by the number of miles maintained by each operator as of 2016 miles. Table 2 breaks Shell’s incidents down by year and number of miles maintained for each of those years.

Table 1: U.S. Pipeline operators ranked by incidents-per-mile

Operator HVL Incidents HVL Pipeline Miles Incidents Per Mile (2016)
Kinder Morgan 387 3,370 0.115
Shell Pipeline 194 3,490 0.056
Chevron 124 2,380 0.051
Sunoco Pipeline 352 6,459 0.049
ExxonMobile 240 5,090 0.048
Colonial Pipeline 244 5,600 0.044
Enbride 258 6,490 0.04
Buckeye Pipeline 231 7,542 0.031
Magellan Pipeline 376 12,928 0.03
Marathan Pipeline 162 5,755 0.029

Table 2: Shell incidents and maintained pipeline miles by year

Year Incidents Pipeline Miles Total Damage Notes
2002 15 no PHMSA data $2,173,704
2003 20 no PHMSA data $3,233,530
2004 25 5,189 $40,344,002 Hurricane Ivan
2005 22 4,830 $62,528,595 Hurricane Katrina & Rita
2006 10 4,967 $11,561,936
2007 5 4,889 $2,217,354
2008 12 5,076 $1,543,288
2009 15 5,063 $11,349,052
2010 9 4,888 $3,401,975
2011 6 4,904 $2,754,750
2012 12 4,503 $17,268,235
2013 4 3,838 $10,058,625
2014 11 3,774 $3,852,006
2015 12 3,630 $4,061,340
2016 6 3,490 $6,875,000
2017 9 no PHMSA data $242,800
2018 1 no PHMSA data $47,000 As of 3/1/18

Cause & Location of Failure

What were the causes of Shell’s pipeline incidents? At Shell’s public informational session, it was said that “in the industry, we know that the biggest issue with pipeline accidents is third party problems – when someone, not us, hits the pipeline.” However, PHMSA data reveal that most of Shell’s incidents issues should have been under the company’s control. For instance, 66% (128) of incidents were due to equipment failure, corrosion, welding failure, structural issues, or incorrect operations (Table 3).

Table 3. Shell Pipeline incidents by cause of failure

Cause Incidents
Equipment Failure 51
Corrosion 37
Natural Forces 35
Incorrect Operation 25
Other 20
Material and/or Weld Failure 15
Excavation Damage 11
Total 194

However, not all of these incidents occurred at one of Shell’s petrochemical facilities. As Table 4 below illustrates, at least 57 incidents occurred somewhere along the pipeline’s right-of-way through public areas or migrated off Shell’s property to impact public spaces. These numbers may be higher as 47 incidents have no mention of the property where incidents occurred.

Table 4. Shell Pipeline incidents by location of failure

Location Incidents
Contained on Operator Property 88
Pipeline Right-of-Way 54
Unknwon 47
Originated on Operator Property, Migrated off Property 3
Contained on Operator-Controlled Right-of-Way 2
Total 194

On several occasions, Shell has claimed that the Falcon will be safely “unseen and out of mind” beneath at least 4ft of ground cover. However, even when this standard is exceeded, PHMSA data revealed that at least a third of Shell’s incidents occurred beneath 4ft or more of soil.

Many of the aboveground incidents occurred at sites like pumping stations and shut-off valves. For instance, a 2016 ethylene spill in Louisiana was caused by lightning striking a pumping station, leading to pump failure and an eventual fire. In numerous incidents, valves failed due to water seeping into systems from frozen pipes, or large rain events overflowing facility sump pumps. Table 5 below breaks these incidents down by the kind of commodity involved in each case.

Table 5. Shell Pipeline incidents by commodity spill volumes

Commodity Barrels
Crude Oil 51,743
Highly Volatile Liquids 6,066
Gas/Diesel/Fuel 1,156
Petroleum Products 325
Total 59,290

Impacts & Costs

None of Shell’s incidents resulted in fatalities, injuries, or major explosions. However, there is evidence of significant environmental and community impacts. Of 150 incidents that included such data, 76 resulted in soil contamination and 38 resulted in water contamination issues. Furthermore, 78 incidents occurred in high consequence areas (HCAs)—locations along the pipeline that were identified during construction as having sensitive environmental habitats, drinking water resources, or densely populated areas.

Table 6 below shows the costs of the 194 incidents. These numbers are somewhat deceiving as the “Public (other)” category includes such things as inspections, environmental cleanup, and disposal of contaminated soil. Thus, the costs incurred by private citizens and public services totaled more than $80-million.

Table 6. Costs of damage from Shell Pipeline incidents

Private Property Emergency Response Environmental Cleanup Public (other) Damage to Operator Total Cost
$266,575 $62,134,861 $11,024,900 $7,308,000 $102,778,856 $183,513,192

A number of significant incidents are worth mention. For instance, in 2013, a Shell pipeline rupture led to as much as 30,000 gallons of crude oil spilling into a waterway near Houston, Texas, that connects to the Gulf of Mexico. Shell’s initial position was that no rupture or spill had occurred, but this was later found not to be the case after investigations by the U.S. Coast Guard. The image at the top of this page depicts Shell’s cleanup efforts in the waterway.

Another incident found that a Shell crude oil pipeline ruptured twice in less than a year in the San Joaquin Valley, CA. Investigations found that the ruptures were due to “fatigue cracks” that led to 60,000 gallons of oil spilling into grasslands, resulting in more than $6 million in environmental damage and emergency response costs. Concerns raised by the State Fire Marshal’s Pipeline Safety Division following the second spill in 2016 forced Shell to replace a 12-mile stretch of the problematic pipeline, as seen in the image above.

Conclusion

These findings suggest that while Shell is obligated to stress safety to sell the Falcon pipeline to the public, people should take Shell’s “good neighbor” narrative with a degree of skepticism. The numbers presented by PHMSA’s pipeline incident data significantly undermine Shell’s claim of having a proven track record as a safe and responsible operator. In fact, Shell ranks near the top of all US operators for incidents per HVL pipeline mile maintained, as well as damage totals.

There are inherent gaps in our analysis based on data inadequacies worth noting. Incidents dealt with at the state level may not make their way into PHMSA’s data, nor would problems that are not voluntary reported by pipeline operators. Issues similar to what the state of Pennsylvania has experienced with Sunoco Pipeline’s Mariner East 2, where horizontal drilling mishaps have contaminated dozens of streams and private drinking water wells, would likely not be reflected in PHMSA’s data unless those incidents resulted in federal interventions.

Based on the available data, however, most of Shell’s pipelines support one of the company’s many refining and storage facilities, primarily located in California and the Gulf states of Texas and Louisiana. Unsurprisingly, these areas are also where we see dense clusters of pipeline incidents attributed to Shell. In addition, many of Shell’s incidents appear to be the result of inadequate maintenance and improper operations, and less so due to factors beyond their control.

As Shell’s footprint in the Appalachian region expands, their safety history suggests we could see the same proliferation of pipeline incidents in this area over time, as well.

NOTE: This article was amended on 4/9/18 to include table 2.

Header image credit: AFP Photo / Joe Raedle

By Kirk Jalbert, FracTracker Alliance

Pipeline Regulations & Impact Assessments, a Primer

Part of the Falcon Public EIA Project

Pipelines are categorized by what they carry — natural gas, oil, or natural gas liquids (NGLs) — and where they go — interstate or intrastate. The regulatory system is complicated. This primer is a quick guide to the agencies that may be involved in Falcon’s permit reviews.

Regulating Pipelines

The siting of natural gas pipelines crossing state or country boundaries is regulated by the Federal Energy Regulatory Commission (FERC). Meanwhile, determination of the location of natural gas routes that do not cross such boundaries are not jurisdictional to FERC, instead determined by the owner pipeline company. Hazardous liquids and NGL pipelines are not regulated for siting by FERC regardless of their location and destination. However, FERC does have authority over determining rates and terms of service in these cases. The U.S. Army Corps of Engineers gets involved when pipelines cross navigable waters such as large rivers and state Environmental Protection Agencies.

Pipeline design, operation, and safety regulations are established by the Pipeline and Hazardous Materials Safety Administration (PHMSA), but these regulations may vary state-by-state as long as minimal federal standards are met by the pipeline project. Notably, PHMSA’s oversight of safety issues does not determine where a pipeline is constructed as this is regulated by the different agencies mentioned above – nor are PHMSA’s safety considerations reviewed simultaneously in siting determinations done by other agencies.

An excerpt from the U.S. Army Corps’ EIS of the Atlantic Sunrise pipeline

These federal agencies are required by the National Environmental Policy Act (NEPA) to prepare an Environmental Impact Statement (EIS) investigating how the pipeline pertains to things like the Clean Water Act, the Endangered Species Act, the National Historic Preservation Act, as well as state and local laws. The image above, for instance, is a caption from the Army Corp’s assessment of the Atlantic Sunrise, a natural gas pipeline.

An EIS is based on surveying and background research conducted by the company proposing the project, then submitted to agencies as an Environmental Impact Assessment (EIA). An EIS can exceed hundreds of pages and can go through many drafts as companies are asked to refine their EIA in order to qualify for approval.

An excerpt from the PA DEP’s review of water crossings for the Mariner East 2 pipeline

Pipeline proposals are also evaluated by state and local agencies. In Pennsylvania, for instance, the PA DEP is responsible for assessing how to minimize pipeline impacts. The DEP’s mission is to protect Pennsylvania’s air, land and water from pollution and to provide for the health and safety of its citizens through a cleaner environment. The PA Fish and Boat Commission oversees the avoidance or relocation of protected species. Local township zoning codes can also apply, such as to where facilities are sited near zoned residential areas or drinking reservoirs, but these can be overruled by decisions made at the federal level, especially when eminent domain is granted to the project.

Regulating the Falcon

For the Falcon pipeline, an interstate pipeline that will transport ethane (an NGL), FERC will likely have authority over determining rates and terms of service, but not siting. Construction permitting will be left state agencies and PHMSA will retain its federal authority with the Pennsylvania Public Utilities Commission (PUC) acting as PHMSA’s state agent to ensure the project complies with federal safety standards and to investigate violations. The Army Corps will almost certainly be involved given that the Falcon will cross the Ohio River. As far as we know, the Falcon will not have eminent domain status because it supplies a private facility and, thus, does not qualify as a public utility project.

Questioning Impact Assessments

The contents of EIAs vary, but are generally organized along the lines of the thematic categories that we have created for assessing the Falcon data, as seen above. However, there is also much that EISs fail to adequately address. The Army Corp’s assessment of the Atlantic Sunrise is a good example. The final EIS resulting from the operators EIA includes considerations for socioeconomic impacts, such effects on employment and environmental justice, as seen in the excerpt below. But potential negative impact in these areas are not necessarily linked to laws requiring special accommodations. For instance, federal regulations mandate achieving environmental justice by “identifying and addressing, as appropriate, disproportionately high and adverse human health or environmental effects” of projects subject to NEPA’s EIS requirement. However, there are no laws that outline thresholds of unacceptable impact that would disallow a project to proceed.

An excerpt from the Atlantic Sunrise EIS addressing environmental justice concerns

Furthermore, the narratives of EIAs are almost always written by the companies proposing the project, using sources of data that better support their claims of minimal or positive impact. This is again seen in the Atlantic Sunrise EIS, where several studies are cited on how pipelines have no affect on property values or mortgages, with no mention of other studies that contradict such findings. Other factors that may be important when considering pipeline projects, such as concerns for sustainability, climate change, or a community’s social well-being, are noticeably absent.

Complicating matters, some pipeline operators have been successful in skirting comprehensive EIAs. This was seen in the case of the Mariner East 2 pipeline. Despite being the largest pipeline project in Pennsylvania’s history, a NEPA review was never conducted for ME2.

* * *

Related Articles

By Kirk Jalbert, FracTracker Alliance

The Falcon: Class Locations, Buildings & Recreational Areas

Part of the Falcon Public EIA Project

In this segment of the Falcon Public EIA Project we begin to explore the different ways that pipelines are assessed for potential risk to populated areas. We outline a methods dictated by the Pipeline and Hazardous Materials Safety Administration (PHMSA) called Class Locations. This methods identifies occupied structures in proximity to a pipeline.

Quick Falcon Facts

  • 67% of the Falcon route will qualify as Class 1, 27% as Class 2, and 3% as Class 3.
  • More than 557 single family residences and 20 businesses within 660ft of the pipeline.
  • Three recreational parks and a planned luxury housing development also at risk.

Map of Falcon Class Locations

The following map will serve as our guide in breaking down the Falcon’s Class Locations. Expand the map full-screen to explore its contents in greater depth. Some layers only become visible as you zoom in. A number of additional layers are not shown by default, but can be turned on in the “layers” tab. Click the “details” tab in full-screen mode to read how the different layers were created.



View Map Fullscreen | How FracTracker Maps Work

Pipeline Class Locations

Pipeline “Class locations” determine certain aspects of how a pipeline is constructed. Essentially, a pipeline’s route is segmented into lengths that are each given different classifications as outlined in PHMSA guidelinesIn general terms, a segment’s Class is established by first calculating a buffer that extends 220 yards (660ft) on either side of the pipeline’s center in 1-mile continuous lengths. This buffer area is then analyzed for how many building structures are present. Classes are then assigned to each 1-mile segment using the follow criteria:

  • Class 1: a segment with 10 or fewer buildings intended for human occupancy
  • Class 2: a segment with more than 10, but less than 46 buildings intended for human occupancy
  • Class 3: a segment with 46 or more buildings intended for human occupancy, or where the pipeline lies within 100 yards of any building, or small well-defined outside area occupied by 20 or more people on at least 5 days a week for 10 weeks in any 12-month period (i.e. schools, businesses, recreation areas, churches)
  • Class 4: a segment where buildings with four or more stories aboveground are prevalent

The finer details of these calculations and their adjustments are complex, however. For instance, Class locations can be shortened to less than 1-mile lengths if building densities change dramatically in an certain area. The example image below shows one of the ways available to operators for doing this, called the “continuous sliding” method:

Calculating Class Locations
(source: PHMSA)

Class location designations may also be adjusted over time as densities change. For instance, if new homes were built in proximity to a previously constructed pipeline, the operator may be required to reduce their operating pressure, strengthen the pipeline, or conduct pressure tests to ensure the segment would technically meet the requirements of a higher Class. Alternatively, operators can apply for a special permit to avoid such changes.

What Class Locations Dictate

Pipeline segments with higher Classes must meet more rigorous safety standards, which are enforced either by PHMSA or by their state equivalent, such as the Pennsylvania Utility Commission. These include:

  • Soil depth: Class 1 locations must be installed with a minimum soil depth of 30 inches (18 inches in consolidated rock). Class 2, 3, and 4 locations require a minimum soil depth of 36 inches (24 inches in consolidated rock)
  • Shut-off valves: Class locations determine the maximum distance from shut-off valves to populated areas, as follows: Class 1 (10 miles), Class 2 (7.5 miles), Class 3 (4 miles), and Class 4 (2.5 miles).
  • Operating pressure: Classes also regulate the maximum allowable operating pressure (MAOP) of pipeline segments
  • Structural integrity: Classes determine where thicker walled materials must be used to withstand higher pressures, as well as different structural testing methods used in safety inspections

By replicating the 600 foot buffer from the Falcon’s centerline (used as the standard distance for determining Class Locations) we found that 67% of the Falcon route will qualify as Class 1, 27% as Class 2, and 3% as Class 3. These are represented on our interactive maps as green, yellow, and orange segments, respectively. An additional segment is marked as having an “unknown” Class on our maps (shaded in gray). This is the stretch crossing the Ohio River, where Shell’s Class location analysis has not been updated to reflect the route change that occurred in the summer of 2017.

Residential Structures

In total, there are 557 single family residences, 20 businesses, and a church within the 660ft buffer. Shell’s data also identify non-occupied structures along the route, such as sheds, garages, and other outbuildings. There are 535 such structures, but we did not have the time to replicate the locations of these sites. It is also important to note that the points on our interactive map represent only those identified by Shell, which we believe is an incomplete assessment of occupied structures based on our quick review of satellite maps.

Three residential structures lie directly within the 50-foot right-of-way. One of these homes, located in a Class 2 segment in Independence Township, is shown below. The Falcon will come as close as 20 feet to the edge of the structure and surround the home on three sides.

An occupied residence in the right-of-way

Neighborhoods in the following five communities account for the entirety of Falcon’s Class 3 locations. These would be considered the most “at risk” areas along the route in terms of proximity to the number of occupied structures. For instance, below is a satellite view of the Class 3 section of Raccoon Township.

  • Rumley Township, Harrison OH
  • Knox Township, Jefferson County OH
  • Raccoon Township, Beaver County PA
  • Independence Township, Beaver County PA
  • Mount Pleasant Township, Washington County PA

Raccoon Township residences & Municipal Park in a Class 3

Recreational Areas

In the above image we also see the location of Raccoon Township Municipal Park (in purple), home to a number of ballfields. Two similar recreation areas are located in the 660ft Class Location buffer: Mill Creek Ballpark, in Beaver County PA, and Clinton Community Park, in Allegheny County PA.

However, the Raccoon Township park is notable in that the Falcon cuts directly through its property boundary. Shell intends to bore under the park using HDD techniques, as stated in their permit applications, “to avoid disturbance to Beaver County baseball field/recreational park,” also stating that, “this HDD may be removed if the recreational group will allow laying the pipeline along the entrance roadway.”

New Housing Developments

One discovery worth attention is that the Falcon runs straight through an under-construction luxury housing development. Located in Allegheny County, PA, its developer, Maronda Homes, bills this growing community as having “picturesque landscapes, waterfront views and a peaceful collection of homes.” Shell mentions this development in their permit applications, stating:

Maronda Homes is in the planning and design stage of a very large housing development and SPLC [Shell Pipeline LC] worked closely with the developer and the Project was rerouted to avoid most of the housing sites.

It stands to reason that this neighborhood will eventually rank as one of the densest Class 3 areas along the Falcon route. Whether or not the pipeline is updated with higher safety standards as a result remains to be seen. The image below illustrates where the Falcon will go relative to lots marked for new homes. This property lots diagram was obtained from Shell’s GIS data layer and can be viewed on the FracTracker interactive map as well.

The Falcon intersects a luxury home development

1/31/18 Note: the Pittsburg Post-Gazette obtained newer lot line records for a portion of the Maronda Farms during their investigation into this story. These new records appear to have some alterations to the development, as seen below.

Maronda Farms, updated lot lines

Issues with Setbacks

There are no setback restrictions for building new homes in proximity to a pipeline. Parcels will eventually be sectioned off and sold to home buyers, begging the question of whether or not people in this community will realize a hazardous liquid pipeline runs past their driveways and backyards. This is a dilemma that residents in a similar development in Firestone, Colorado, are now grappling with following a recent pipeline explosion that killed two people, seen below, due to inadequate building setbacks.

A pipeline explodes in a Colorado home development
(source: InsideEnergy, CO)

Interestingly, we researched these same Maronda Farms parcels in FracTracker’s Allegheny County Lease Mapping Project only to discover that Maronda Homes also auctioned off their mineral rights for future oil and gas drilling. New homeowners would become victims of split-estate, where drilling companies can explore for oil and gas without having to seek permission from property owners, amplifying their level of risk.

* * *

Related Articles

Updated Pipeline Incident Analysis

By Matt Kelso, Manager of Data & Technology

As massive new pipeline projects continue to generate news, the existing midstream infrastructure that’s hidden beneath our feet continues to be problematic on a daily basis. Since 2010, there have been 4,215 pipeline incidents resulting in 100 reported fatalities, 470 injuries, and property damage exceeding $3.4 billion.

Chart 1: Cumulative impacts pipeline incidents in the US. Data collected from PHMSA on November 4th, 2016. Operators are required to submit incident reports within 30 days.

Figure 1: Cumulative impacts pipeline incidents in the US. Data collected from PHMSA on November 4th, 2016. Operators are required to submit incident reports within 30 days.

In our previous analyses, pipeline incidents occurred at a rate of 1.6 per day nationwide, according to data from the Pipeline and Hazardous Materials Safety Administration (PHMSA). Rates exceeding 1.9 incidents per day in 2014 and 2015 have brought the average rate up to 1.7 incidents per day. Incidents have been a bit less frequent in 2016, coming in at a rate of 1.43 incidents per day, or 1.59 if we roll results back to October 4th in order to capture all incidents that are reported within the mandatory 30 day window.

Chart 1: Pipeline incidents per day for years between 2010 and 2016. Incidents after October 4, 2016 may not be included in these figures.

Figure 2: Pipeline incidents per day for years between 2010 and 2016. Incidents after October 4, 2016 may not be included in these figures.

These figures are the aggregation of three reports, namely natural gas transmission and gathering pipelines (828 incidents since 2010), natural gas distribution (736 incidents), and hazardous liquids (2,651 incidents). Not all of the hazardous liquids are petroleum related, but the vast majority are. 1,321 of the releases involved crude oil, and an additional 896 involved other liquid petroleum products, accounting for 84% of hazardous liquid incidents. The number could be higher, depending on the specific substances involved in the 399 highly volatile liquid (HVL) related incidents. The HVL category includes propane, butane, liquefied petroleum gases, ethylene, and propylene, as well as other volatile liquids that become gaseous at ambient conditions.

What is causing all of these pipeline incidents?

Figure 3: Cause of pipeline incidents for all reports received from January 1, 2010 through November 4, 2016.

Nonprofits, academics, and concerned citizens looking for accurate pipeline data will find that it is restricted, with the argument that releasing accurate pipeline data constitutes a threat to national security.  This makes little sense for several reasons. First, with over 2.4 million miles of pipelines, they are nearly omnipresent. Additionally, similar data access restrictions only apply to midstream infrastructure such as pipelines and compressor stations, whereas the locations for wells, refineries, and power plants are all publicly available, despite the presence of the same volatile hydrocarbons at these facilities. Additionally, pipelines are purposefully marked with surface placards to help prevent unintentionally impacting the infrastructure.

In fact, a quick look at the causes of pipeline incidents reveal that it it much more dangerous to not know where the pipelines are located. In the “Other Incident Cause” category (Figure 3) there are 152 incidents that were caused by unsuspecting motor vehicles. When this is combined with incidents resulting from excavation damage, we have 558 cases where “not knowing” about the pipeline’s location likely contributed to the failure. On the other hand, there are 14 incidents (only .003%) where the cause is identified as intentional. While even one case of tampering with pipeline infrastructure is unacceptable, PHMSA incident data indicate that obfuscated pipelines are 40 times more likely to cause a problem when compared to sabotage. Equipment failures and corrosion account for more than half of all incidents.

Where do these incidents occur?

PHMSA is not allowed to make accurate pipeline location data available for download, but such rules apparently do not apply to pipeline incidents. The following map shows the 4,215 pipeline releases since 2010, highlighting those that have resulted in injuries and fatalities.

Pipeline incidents in the US. Please zoom in to access specific incident data.  To see the legend and other tools, Please click here.

 

Pipeline Incidents by State for reports received 1/1/2010 through 11/4/2016.

Figure 4: Pipeline incidents by state for reports received 1/1/2010 through 11/4/2016.

While operators are required to submit the incident’s location as a part of their report to PHMSA, data entry errors are common in the dataset. The FracTracker Alliance has been able to identify and correct a few of the higher profile errors, such as the February 9, 2011 explosion in Allentown, PA, the report for which had mangled the latitude and longitude values so badly that the incident was rendered in Greenland. Other errors persist in the dataset, however. Since 2010, pipeline incidents have occurred in Washington, DC, Puerto Rico, and 49 states (the exception being Vermont). Ten states have at least 100 incidents apiece during the past six years (see Figure 4), and more than a quarter of all pipeline incidents in that time frame have occurred in Texas.

Which operators are responsible?

Figure 5: This table shows the ten operators with the most reported incidents, along with the length of their pipeline network.

Figure 5: This table shows the ten operators with the most reported incidents, along with the length of their pipeline network.

Altogether, there are 521 pipeline operators with reported releases, although many of these are affiliated with one another in some fashion. For example, the top two results in Figure 5 are almost certainly both subsidiaries of Enterprise Products Partners, L.P.

The real outlier in Figure 5, in terms of incidents per 100 miles, is Kinder Morgan Liquid Terminals; LLC. However, this is one of ten or more companies that share the Kinder Morgan name when reporting pipeline inventories. When taken in aggregate, companies with the Kinder Morgan name accounted for 142 incidents over a reported 7,939 miles, for a rate of 1.8 incidents per 100 miles. It should be noted that this, along with all of the statistics in Figure 5, are entirely based on matching the operator name between the incident and inventory reports.  Kinder Morgan’s webpage boasts of 84,000 miles of pipelines in the US — there are numerous possible explanations for the discrepancy in pipeline length, including additional Kinder Morgan subsidiaries, as well as whether gathering lines that aren’t considered to be mains are on both lists.

The operators responsible for the most deaths from pipeline incidents since 2010 include Pacific Gas & Electric (15), Washington Gas Light (9), and Consolidated Edison Co. of New York (8). Of course, the greatest variable in whether or not a pipeline explosion kills people or not is whether or not the incident happens in a populated location. In the course of this analysis, there were 230 explosions and 635 fires over 2,500 days, meaning that there is pipeline explosion somewhere in the United States every 11 days, on average, and a fire every fourth day. The fact that only 65 of the incidents resulted in fatalities indicates that we have been rather lucky with incidents in the midstream sector.