On January 12, API held an hourlong press conference titled “State of American Energy”. Along with this press conference, they released a series of tweets, one of which showed some charts on pipeline safety that looked impressive, but were incredibly misleading. As the Pipeline Safety Trust’s new Data and GIS Analyst, I’d like to provide some context around the data and illustrate the ways in which it is misleading.
TL;DR While IPE (impacting people &the environment) incidents are down over the past 11 years, picking a five-year horizon to look at their progress takes advantage of a high-incident period ending in 2016, an unusually awful time for pipeline safety. Furthermore, despite the assertion by API’s chart, progress has been anything but linear for pipeline safety and is still far from reaching 0 incidents.
First, Incidents Impacting People & the Environment (IPE) are down over the past 11 years, when these incidents were first defined and subsequently tracked (Plot 1). However, significant incidents remain at a similar level over the same period (2010: 123; 2021:134). Organizations like API also tend to present incidents per barrel-mile, which offers us a rate of incidence relative to pipeline mileage and throughput, but this only obscures the actual number of incidents occurring nationwide. While pipelines continue to expand in length and use, rates of incidents per barrel/mile will continue to look more favorable for the industry. However, there has been minimal progress since 2010 on reducing the actual number of significant incidents observed in the US.
Incidents Impacting People & the Environment (IPE) are defined as resulting in a fatality, injury, explosion, fire, evacuation, property damage, or incidents that have been determined to have an impact on wildlife or water. The incident is labeled IPE if the incident is not contained fully on operator property, the unintentional release volume is at least five gallons in a high consequence area (HCA), at least five barrels outside of an HCA, contaminates surface water or contaminates the soil. Significant Incidents are those which result in any of the following: (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; and (4) Liquid releases resulting in an unintentional fire or explosion (PHMSA).
API’s Tweet also presents a graph that shows three lines falling to the bottom of the x-axis. It’s not necessarily a problem to set an x-axis at a non-zero value when the values offer a relatively narrow range well above 0. However, showing the 5-year trends falling to 0 obscures the long road ahead toward safe pipelines. Plot 2 presents the same data API alludes to in their tweet but extends to 11 years and shows a broader range of values.
Using corrosion failure as an example of improvement over five years is incredibly misguided when considering the 10-year change. Compared to five years, there has been a smaller decrease in IPE incidents over 11 years attributable to corrosion failure (15%). A terrible 2013-2016 period distorts the shorter-term conclusion presented by API. Plot 3 shows the percent change in incidents attributable to all causes and suggests most causes don’t show any obvious 11-year trend. Causes referenced in API’s tweet appear highlighted, while others are displayed in gray. API’s tweet presents the declines as relatively linear, while the data shows a more complex picture.
The percent change from 2010 does show a generally positive picture for IPE causes. As IPE’s are down, most of the causes have improved (except for “Other Natural Force Damage”, which had a 0% change). However, presenting this in a more responsible manner suggests the year-to-year changes are not a simple linear downtrend for any cause. It is worth noting some of these causes are beyond operator control, such as natural force damage, but others are failures of operators and industry.
Finally, API chose to share data about hazardous liquids, and PHMSA’s hazardous liquids data does not include gas pipelines. Plot 5 shows how gas transmission and distribution lines have improved slightly, but vary considerably from year-to-year, while gathering lines and Underground Natural Gas Storage have shown no significant trend at all. The latter two result in relatively few incidents per year, however issues with the lack of regulation and collection of data on gathering lines obfuscate the data considerably (PHMSA FAQ on Gas Gathering).
From the data presented it appears that in-Line inspections have been reduced in the years since 2010 and therefore not a lot of reduction has occurred in this time. The companies must know which lines are old and corroding. Either they need to use ILI or increased Cathodic protection to reduce failure numbers. It would be very interesting to a plot of the numbers of ILI corrosion runs by year since 2010.