Pipeline Safety Data: Reading the Record
How pipeline incident records are built, what the categories mean, and where to find the source behind each safety claim in the United States.
Pipeline incident records in the United States are compiled from operator reports and published by federal regulators, and they are organized by system type rather than by headline. Anyone who wants to check a claim about a spill, an explosion, or a rupture should start with the system category, then the cause code, then the reporting threshold that applied in that year. The numbers rarely speak for themselves, but the filing structure does tell a reader what was counted and what was left out.
What counts as a reportable incident?
A reportable incident is not the same as a news event. Federal rules set thresholds that trigger a written report: a release above a set volume, a fire or explosion, a fatality or injury requiring hospitalization, or damage above a dollar figure that is adjusted over time. Below those thresholds, an event may still be real and still be recorded somewhere, but it will not appear in the national incident file.
That gap matters when someone compares two decades of data. A rise in reported incidents can reflect more pipeline miles, more reporting categories, or a lower threshold, not necessarily worse operation. A fall can reflect fewer miles in service or a change in how a cause is coded.
The reporting pipeline runs from the operator to the federal agency that collects the forms, then into public datasets. Each step introduces a decision: what to count, how to label it, and whether to revise it later. Readers who want the underlying documents rather than a summary can work through the incident records collected at pipeline safety data, which organizes the filings by system and by cause.
Which pipeline systems are tracked separately?
Gas transmission, gas distribution, hazardous liquids, and liquefied natural gas each have their own reporting forms and their own physical characteristics. A transmission line moves gas at high pressure between compressor stations, often across state lines. A distribution main sits under a street and serves homes and businesses at lower pressure. Hazardous liquid lines move crude, refined products, and chemicals. LNG facilities handle gas in liquid form at very low temperature.
Because the systems differ, the incident counts differ. A distribution incident is often an excavation strike in a residential area. A transmission incident may involve a weld or a compressor station. A liquids incident may involve a pump station, a valve, or a tank. Comparing a distribution total to a transmission total without noting the mileage base produces a misleading ratio.
Mileage is the denominator that most casual comparisons forget. Annual mileage reports give the size of each system, so a count can be turned into a rate. A state with many distribution miles will have more distribution incidents than a state with few, even if both operate to the same standard.
What do the cause categories actually mean?
Cause codes come from the filed reports, and they are assigned after investigation. Common categories include corrosion, material and weld failure, excavation damage, equipment failure, incorrect operation, and natural force damage. Each category contains subcategories, and the subcategory is often where the useful detail sits.
Corrosion, for example, splits into external corrosion, internal corrosion, and stress corrosion cracking. Those are different mechanisms with different inspection methods. A single line labeled corrosion in a summary table can hide three distinct problems.
Excavation damage is another category that rewards patience. It covers third party strikes, operator excavation, and damage that occurred years earlier but only failed later. The date of the incident and the date of the damage are not always the same, and a reader who assumes they are will misread the trend.
Material and weld failures are coded from metallurgical findings. The wording in the file is often technical and sometimes ambiguous. A phrase like selective seam weld corrosion describes a specific defect in a specific era of pipe manufacturing. Treating it as generic corrosion loses the point.
How should a reader interpret a single incident file?
The file is a form, not a narrative. It contains fields for location, system, commodity, volume released, ignition, injuries, fatalities, and cause. Some fields are mandatory, some are optional, and some are completed with estimates. The quality of a field depends on who filled it out and how soon after the event.
A useful habit is to read the cause field last. Start with the system and the commodity, then the volume and whether ignition occurred, then the consequences. Only then look at the cause code, because the code is an interpretation placed on top of the physical facts.
Another habit is to check the revision history. Operators sometimes amend a report after further investigation. The amended version is the one that should be cited, but older summaries may still carry the original figure.
What is the regulatory frame around these records?
The rules sit in Title 49 of the Code of Federal Regulations, parts 190 through 199. Part 191 covers gas reporting, part 192 covers gas pipeline safety standards, part 194 covers liquids response plans, part 195 covers hazardous liquid pipeline safety, and part 199 covers drug and alcohol testing. The reporting requirements that generate the incident data come from these parts.
Integrity management is the framework that requires operators of certain lines to identify threats, inspect, and remediate. Inline inspection tools, often called smart pigs, travel inside the pipe and measure wall thickness, cracks, or geometry. The results feed into the integrity plan, and the plan feeds into what gets repaired.
The regulatory text is public and searchable. A reader who wants to know why a particular incident was reportable can find the threshold in the part that applies to that system. That is more reliable than a secondary summary, including this one.
Where do the limits of comparison show up?
Comparisons across years, states, or operators run into several limits. Reporting thresholds change. Cause codes change. Mileage grows or shrinks. A single large incident can dominate a yearly total and make a trend look worse than the underlying rate.
A defensible comparison states the system, the years, the threshold, and the mileage base. It also states what was excluded. Without those four elements, a chart can be accurate and still mislead.
Glossaries help here. Terms like release, rupture, leak, and failure have specific meanings in the filings. A rupture is a specific mechanical event, not a synonym for any spill. Using the terms loosely makes two datasets look comparable when they are not.
For a woodturner, the connection may seem distant. It is not. A workshop that heats with natural gas, or that sits near a distribution main, depends on the same record keeping. The same discipline that makes a gouge cut predictable, measuring the bevel, checking the grain, applies to reading a safety record: know the tool, know the material, and know what the number does not say.
A reader who wants to check a claim about a local line can start with the system type, find the mileage, and read the cause code with the subcategory attached. That path leads back to the filed form, which is the only place where the claim can be confirmed or corrected.
Precise safety and material claims are grounded in the editorial methodology; workshop guidance is identified as practical synthesis.
