kvlak limn

Transportation & Safety / PHMSA Pipeline Safety

What the Pipeline Spilled

Ledger of record: PHMSA incident and accident flagged files / one row per filed report / 2010-2025

Every time a pipeline in America fails hard enough to matter, somebody has to write it down: the barrels lost, the dollars burned, the people hurt. Sixteen years of those filings sit in one federal ledger. This page reads it straight - who spilled, what the pipe carried, how big the spills really run, and the distance between an incident that is significant and one that is serious. The two words are doing very different jobs, and the gap between them is where the truth about pipeline risk lives.

RATE ~300 significant incidents / yr
VOL 0.8M barrels lost, span total
LIFE 180 deaths on record
COST $10.6B property damage
Full PHMSA Pipeline Safety - Incident & Accident Flagged Files / 2010-2025
ST-01 / GEOGRAPHY

Where the Line Fails

Significant incidents by state, 2010-2025

Pipelines follow oil and gas, and so do their failures. Each state is shaded by the count of significant incidents filed there since 1986. Texas, sitting on the densest network of pipe in the country, out-fails every other state by nearly three to one - not because Texas pipe is worse, but because there is simply more of it to fail. The barrels tell the same geography harder: of the roughly 781 thousand net barrels lost across the span, the crude-oil corridor from the Gulf up through Kansas holds most of the stain.

Alabama: 67 significant incidents, 9,897 bbl lost, 4 deaths Alaska: 14 significant incidents, 1,134 bbl lost, 0 deaths Arizona: 18 significant incidents, 0 bbl lost, 2 deaths Colorado: 68 significant incidents, 5,694 bbl lost, 2 deaths Florida: 49 significant incidents, 778 bbl lost, 4 deaths Georgia: 60 significant incidents, 3,549 bbl lost, 4 deaths Indiana: 68 significant incidents, 1,958 bbl lost, 1 deaths Kansas: 149 significant incidents, 24,133 bbl lost, 3 deaths Maine: no data Massachusetts: 28 significant incidents, 0 bbl lost, 3 deaths Minnesota: 78 significant incidents, 5,380 bbl lost, 2 deaths New Jersey: 91 significant incidents, 1,262 bbl lost, 1 deaths North Carolina: 61 significant incidents, 12,192 bbl lost, 2 deaths North Dakota: 66 significant incidents, 26,808 bbl lost, 2 deaths Oklahoma: 277 significant incidents, 30,912 bbl lost, 2 deaths Pennsylvania: 124 significant incidents, 12,612 bbl lost, 21 deaths South Dakota: 12 significant incidents, 50 bbl lost, 1 deaths Texas: 1,323 significant incidents, 353,678 bbl lost, 29 deaths Wyoming: 65 significant incidents, 7,269 bbl lost, 0 deaths Connecticut: 6 significant incidents, 304 bbl lost, 1 deaths Missouri: 79 significant incidents, 5,882 bbl lost, 2 deaths West Virginia: 34 significant incidents, 44,997 bbl lost, 1 deaths Illinois: 150 significant incidents, 36,155 bbl lost, 11 deaths New Mexico: 77 significant incidents, 13,158 bbl lost, 1 deaths Arkansas: 57 significant incidents, 8,449 bbl lost, 1 deaths California: 290 significant incidents, 5,841 bbl lost, 16 deaths Delaware: 1 significant incidents, 0 bbl lost, 0 deaths District of Columbia: 1 significant incidents, 0 bbl lost, 0 deaths Hawaii: 6 significant incidents, 480 bbl lost, 0 deaths Iowa: 60 significant incidents, 7,398 bbl lost, 0 deaths Kentucky: 52 significant incidents, 922 bbl lost, 2 deaths Maryland: 26 significant incidents, 181 bbl lost, 11 deaths Michigan: 146 significant incidents, 6,860 bbl lost, 8 deaths Mississippi: 77 significant incidents, 51,865 bbl lost, 1 deaths Montana: 32 significant incidents, 2,828 bbl lost, 1 deaths New Hampshire: 2 significant incidents, 15 bbl lost, 0 deaths New York: 91 significant incidents, 3,610 bbl lost, 11 deaths Ohio: 158 significant incidents, 1,694 bbl lost, 7 deaths Oregon: 7 significant incidents, 52 bbl lost, 0 deaths Tennessee: 48 significant incidents, 1,898 bbl lost, 3 deaths Utah: 20 significant incidents, 812 bbl lost, 2 deaths Virginia: 39 significant incidents, 102 bbl lost, 0 deaths Washington: 33 significant incidents, 320 bbl lost, 1 deaths Wisconsin: 34 significant incidents, 2,411 bbl lost, 1 deaths Nebraska: 42 significant incidents, 4,455 bbl lost, 5 deaths South Carolina: 24 significant incidents, 4,538 bbl lost, 0 deaths Idaho: 24 significant incidents, 370 bbl lost, 0 deaths Nevada: 12 significant incidents, 53 bbl lost, 0 deaths Vermont: 1 significant incidents, 0 bbl lost, 0 deaths Louisiana: 345 significant incidents, 78,435 bbl lost, 11 deaths Rhode Island: 3 significant incidents, 0 bbl lost, 0 deaths
Shade = significant incidents per state, 2010-2025, five quantile classes on the validated spill ramp. Alaska and Hawaii are inset by the Albers USA projection; territories fall outside its frame and appear only in the table.
Most incidents
  1. 01 Texas 1,323
  2. 02 Louisiana 345
  3. 03 California 290
  4. 04 Oklahoma 277
  5. 05 Ohio 158
Every state, in a table
State Incidents Barrels lost Damage $M Deaths
Texas TX 1,323 353,678 1,454.1 29
Louisiana LA 345 78,435 451.3 11
California CA 290 5,841 1,273.6 16
Oklahoma OK 277 30,912 295.4 2
Ohio OH 158 1,694 186 7
Illinois IL 150 36,155 272.6 11
Kansas KS 149 24,133 696.8 3
Michigan MI 146 6,860 1,324.4 8
Pennsylvania PA 124 12,612 254.7 21
New Jersey NJ 91 1,262 71.2 1
New York NY 91 3,610 52.6 11
Missouri MO 79 5,882 88.5 2
Minnesota MN 78 5,380 94.5 2
New Mexico NM 77 13,158 45.2 1
Mississippi MS 77 51,865 50.6 1
Indiana IN 68 1,958 102.1 1
Colorado CO 68 5,694 97.7 2
Alabama AL 67 9,897 132 4
North Dakota ND 66 26,808 166.6 2
Wyoming WY 65 7,269 38.6 0
North Carolina NC 61 12,192 215.8 2
Iowa IA 60 7,398 36.8 0
Georgia GA 60 3,549 42.5 4
Arkansas AR 57 8,449 167.2 1
Kentucky KY 52 922 99 2
Florida FL 49 778 36.5 4
Tennessee TN 48 1,898 75.6 3
Nebraska NE 42 4,455 34.5 5
Virginia VA 39 102 56.4 0
Wisconsin WI 34 2,411 91 1
West Virginia WV 34 44,997 58.6 1
Washington WA 33 320 95.5 1
Montana MT 32 2,828 205 1
Massachusetts MA 28 0 1,883.7 3
Maryland MD 26 181 22 11
Idaho ID 24 370 14.6 0
South Carolina SC 24 4,538 16.9 0
Utah UT 20 812 114.7 2
Arizona AZ 18 0 27.4 2
Alaska AK 14 1,134 89.8 0
South Dakota SD 12 50 68.1 1
Nevada NV 12 53 3.9 0
Oregon OR 7 52 20.9 0
Connecticut CT 6 304 5.1 1
Hawaii HI 6 480 5.2 0
Rhode Island RI 3 0 1.8 0
New Hampshire NH 2 15 3.2 0
Delaware DE 1 0 0.1 0
Vermont VT 1 0 0.3 0
District of Columbia DC 1 0 0.6 0
OP-02 / LEDGER

Who Spilled the Most

Top operators by net barrels lost

Ranked by net barrels lost - what was released and never recovered. The liquid haulers own this ledger, and they should: they move the most liquid through the most pipe. Watch the bottom of the list instead. The gas utilities sit there with zero barrels and the worst death tolls on the page, because gas does not spill - it finds a basement and waits for a spark. Bars share one scale; length is the comparison.

  1. 01
    ENTERPRISE PRODUCTS OPERATING LLC HVL/NGL
    175,068 bbl 236 incidents $169M damage 6 deaths
  2. 02
    ENERGY TRANSFER COMPANY HVL/NGL
    62,238 bbl 85 incidents $47M damage 1 death
  3. 03
    COLONIAL PIPELINE CO Refined
    25,956 bbl 81 incidents $337M damage 2 deaths
  4. 04
    ONEOK NGL PIPELINE, LLC Refined
    18,762 bbl 69 incidents $6M damage
  5. 05
    PHILLIPS 66 PIPELINE LLC HVL/NGL
    17,955 bbl 60 incidents $41M damage 1 death
  6. 06
    ENTERPRISE CRUDE PIPELINE LLC Crude
    14,137 bbl 116 incidents $96M damage
  7. 07
    SUNOCO PIPELINE L.P. Crude
    13,299 bbl 88 incidents $107M damage
  8. 08
    MARATHON PIPE LINE LLC Crude
    8,280 bbl 75 incidents $141M damage
  9. 09
    KINDER MORGAN LIQUID TERMINALS, LLC Refined
    7,463 bbl 54 incidents $36M damage
  10. 10
    MAGELLAN PIPELINE COMPANY, LP Refined
    6,850 bbl 80 incidents $60M damage
  11. 11
    PLAINS PIPELINE, L.P. Crude
    4,158 bbl 86 incidents $222M damage
  12. 12
    SHELL PIPELINE CO., L.P. Crude
    2,006 bbl 53 incidents $91M damage
  13. 13
    BUCKEYE PARTNERS, LP Refined
    1,747 bbl 76 incidents $89M damage
  14. 14
    PACIFIC GAS & ELECTRIC CO Gas dist.
    no spill volume - gas system
    0 bbl 87 incidents $840M damage 12 deaths
  15. 15
    SOUTHERN CALIFORNIA GAS CO Gas dist.
    no spill volume - gas system
    0 bbl 64 incidents $96M damage 4 deaths
  16. 16
    CONSUMERS ENERGY CO Gas dist.
    no spill volume - gas system
    0 bbl 60 incidents $80M damage 5 deaths

Barrels are net loss (released minus recovered). Operator totals span mergers and name changes.

OP-03 / CROSSCHECK

Barrels Are Not Bodies

Net barrels lost vs deaths, per operator

Cross the two ledgers and they refuse to agree. The liquid haulers that top the volume ranking sit along the floor of this chart: tens to hundreds of thousands of barrels, one or two deaths each. The gas utilities sit pinned to the left wall - zero barrels, ever, and 21 of the 31 deaths among these sixteen operators. A barrel count measures the mess. It does not measure the harm.

0 5 10 15 050k100k150k200k net barrels lost, span total deaths ENTERPRISE PRODUCTS OPERATING LLC: 175,068 bbl lost, 6 deaths, 236 incidents Enterprise Products ENTERPRISE CRUDE PIPELINE LLC: 14,137 bbl lost, 0 deaths, 116 incidents SUNOCO PIPELINE L.P.: 13,299 bbl lost, 0 deaths, 88 incidents PACIFIC GAS & ELECTRIC CO: 0 bbl lost, 12 deaths, 87 incidents Pacific Gas & Electric PLAINS PIPELINE, L.P.: 4,158 bbl lost, 0 deaths, 86 incidents ENERGY TRANSFER COMPANY: 62,238 bbl lost, 1 death, 85 incidents Energy Transfer COLONIAL PIPELINE CO: 25,956 bbl lost, 2 deaths, 81 incidents Colonial MAGELLAN PIPELINE COMPANY, LP: 6,850 bbl lost, 0 deaths, 80 incidents BUCKEYE PARTNERS, LP: 1,747 bbl lost, 0 deaths, 76 incidents MARATHON PIPE LINE LLC: 8,280 bbl lost, 0 deaths, 75 incidents ONEOK NGL PIPELINE, LLC: 18,762 bbl lost, 0 deaths, 69 incidents SOUTHERN CALIFORNIA GAS CO: 0 bbl lost, 4 deaths, 64 incidents SoCalGas PHILLIPS 66 PIPELINE LLC: 17,955 bbl lost, 1 death, 60 incidents CONSUMERS ENERGY CO: 0 bbl lost, 5 deaths, 60 incidents Consumers Energy KINDER MORGAN LIQUID TERMINALS, LLC: 7,463 bbl lost, 0 deaths, 54 incidents SHELL PIPELINE CO., L.P.: 2,006 bbl lost, 0 deaths, 53 incidents
Liquid systems (crude, refined, HVL) Gas systems
Sixteen operators, span totals: net barrels lost against deaths. Gas systems report no barrels by definition, so their dots sit on the zero line - the volume ledger cannot see them at all.
Every operator, in a table
OperatorSystemBarrels lostDeathsInjuries
ENTERPRISE PRODUCTS OPERATING LLC Liquid 175,068 6 13
ENTERPRISE CRUDE PIPELINE LLC Liquid 14,137 0 2
SUNOCO PIPELINE L.P. Liquid 13,299 0 0
PACIFIC GAS & ELECTRIC CO Gas 0 12 77
PLAINS PIPELINE, L.P. Liquid 4,158 0 0
ENERGY TRANSFER COMPANY Liquid 62,238 1 3
COLONIAL PIPELINE CO Liquid 25,956 2 4
MAGELLAN PIPELINE COMPANY, LP Liquid 6,850 0 0
BUCKEYE PARTNERS, LP Liquid 1,747 0 0
MARATHON PIPE LINE LLC Liquid 8,280 0 0
ONEOK NGL PIPELINE, LLC Liquid 18,762 0 2
SOUTHERN CALIFORNIA GAS CO Gas 0 4 16
PHILLIPS 66 PIPELINE LLC Liquid 17,955 1 1
CONSUMERS ENERGY CO Gas 0 5 9
KINDER MORGAN LIQUID TERMINALS, LLC Liquid 7,463 0 0
SHELL PIPELINE CO., L.P. Liquid 2,006 0 0
SP-04 / THE SPLIT

Significant Is Not Serious

Recent 10-year annual average

The regulator counts in nested tiers, and the words are doing precise work. Significant is the wide net: a fire, a big enough spill, fifty thousand 1984 dollars of damage. Serious is the narrow one: a death or a hospitalization. Out of roughly seven hundred significant incidents a year, about one in fourteen hurts a person badly enough to count. Quote the big number as if it measured bodies and you inflate the toll fourteen-fold; quote only the deaths and you hide how often the line actually lets go. The page holds both numbers apart on purpose.

All reported incidents 563/yr

Every incident in PHMSA's flagged files that year - the significant tier is a subset. Excludes the categories PHMSA leaves out of its trend files (gravity HL, reporting-regulated-only HL gathering, Type R gas gathering).

Significant 285/yr
51% of prior

Death, injury needing hospitalization, fire/explosion, >=5 bbl HVL or >=50 bbl liquid, or >=$50k (1984 dollars).

Serious 26/yr
9% of prior

A fatality or an injury requiring in-patient hospitalization.

Fatal 1/yr
4% of prior

At least one death. A subset of serious.

TR-05 / TREND

The Line Holds, the Deaths Stay Low

2010-2025

Significant incidents held roughly flat across the span, from about 265 a year to 276, with a mid-decade bulge in between. The 2010 rule revision, pinned at the left, widened what operators must report near the start of this window. The serious count, the line hugging the floor, barely moved at all. The system is not filing many more incidents than it did in 2010. It is not killing more people either.

0 100 200 300 400 2010201520202025 2010 rule 276 significant 27 serious
Significant incidents Serious (a death or a hospitalization)
Significant and serious incidents filed per year, 2010-2025. The vertical pin marks the 2010 reporting-rule revision, which widened what counts. The serious tier is a subset of the significant tier.
Year-by-year, in a table
YearSignificantSeriousDeathsBarrels lost
2010 265 34 19 49,191
2011 286 31 11 57,170
2012 256 28 10 29,114
2013 305 24 8 85,376
2014 302 27 19 21,914
2015 333 26 9 80,951
2016 310 38 16 46,080
2017 303 23 7 44,918
2018 293 36 6 70,448
2019 317 25 11 26,104
2020 284 27 15 105,426
2021 269 25 9 41,206
2022 262 13 1 30,351
2023 272 23 16 21,163
2024 264 22 12 48,772
2025 276 27 11 23,233
TR-06 / GAUGES

Same Years, Four Needles

Each panel on its own scale

Put the four measures side by side and none of them tells the same story. Significant incidents hold in a band near 300 a year. Serious incidents drift inside a narrower band still. Deaths spike and fall with single catastrophes - 2010, 2014, 2023 - and never settle into a direction. Barrels swing on the price and flow of crude. Anyone quoting one of these lines as if it were the other three is telling you a story the ledger does not.

Four readings of the same yearly ledger, 2010-2025. Panels share the x axis but not the y scale; each panel's peak year sets its top gridline and is printed in its title.
ER-07 / ERAS

The Ledger Tilts Toward Gas

Significant incidents by decade and system

Split the ledger into two halves and sort each by what the failing system carried. The gas share climbs across the window - from about 48% of incidents in the earlier band to 51% in the recent one, tipping the majority from liquid to gas. Part of that is real (an aging distribution network under growing cities), part is definitional (gathering lines only entered broad reporting in 2022). Either way, the modern pipeline incident is now as likely to be a gas leak as an oil spill.

0 1,000 2,000 2010-2017 48% gas 2,360 2018-2025 51% gas 2,237
Gas systems (distribution, transmission, gathering, LNG) Liquid systems (crude, refined, HVL)
Two even halves of the modern ledger (2010-2017 and 2018-2025). Bar length is total significant incidents on one shared scale; the printed percentage is the gas share of that era.
The decades, in a table
DecadeGas incidentsLiquid incidentsTotalDeaths
2010-2017 1,126 1,234 2,360 99
2018-2025 1,149 1,088 2,237 81
CM-08 / COMMODITY

What the Pipe Carried

Incidents / barrels / deaths

Score each cargo three ways and the columns refuse to agree. Gas transmission leads on incidents, yet registers not a single barrel - gas leaks and explodes, it does not pool. The volume ledger belongs to highly volatile liquids / ngl, which spills oceans and kills almost no one. Each bar is scaled to its own column's maximum; read down a column to rank the cargos, across a row to see one cargo's whole ledger.

Gas transmission
1,133
-
33
Crude oil
1,000
113,265
3
Gas distribution
981
-
130
Refined products
739
179,944
2
Highly volatile liquids / NGL
583
488,208
11
Gas gathering
146
-
0
LNG facilities
15
-
1
CM-09 / SHARES

The Death Share Is Not the Spill Share

Share of incidents vs share of deaths

Take each commodity's slice of the incident ledger and follow it to its slice of the death ledger. Gas distribution - the small pipes under streets and into basements - files about 21% of significant incidents and takes 72% of the deaths. Every liquid line falls: crude oil is about a fifth of the incidents and a sliver of the deaths. The pipe that reaches your house is the one that kills.

Share of incidents Share of deaths Gas transmission 25% 18% Gas transmission Crude oil 22% 1.7% Crude oil Gas distribution 21% 72% Gas distribution Refined products 16% 1.1% Refined products HVL / NGL 13% 6.1% HVL / NGL Gas gathering 3.2% 0.0% Gas gathering LNG 0.3% 0.6% LNG
Gas systems Liquid systems
Each commodity's share of all significant incidents (left) and of all pipeline deaths (right), span totals. Labels nudged apart where values crowd; the thin connectors point back to the true position.
The shares, in a table
CommoditySystemIncidentsShareDeathsShare
Gas transmission Gas 1,133 25% 33 18%
Crude oil Liquid 1,000 22% 3 1.7%
Gas distribution Gas 981 21% 130 72%
Refined products Liquid 739 16% 2 1.1%
Highly volatile liquids / NGL Liquid 583 13% 11 6.1%
Gas gathering Gas 146 3.2% 0 0.0%
LNG facilities Gas 15 0.3% 1 0.6%
SZ-10 / SIZE

Most Spills Are Small. The Barrels Are Not.

Liquid systems, net barrels per incident

Sort every liquid-system incident by how much it actually released and the ledger splits in two. Three out of four spills stay under fifty barrels - a patch of ground, a cleanup crew, a filing. But the 7.6% of incidents that top five hundred barrels carry 92% of everything ever spilled. The typical incident is a puddle. The typical barrel is lost in a flood.

How often each size happens share of liquid incidents 61% under 5 bbl 1,417 spills 18% 5 to 50 bbl 427 spills 13% 50 to 500 bbl 301 spills 6.1% 500 to 5,000 bbl 141 spills 1.6% over 5,000 bbl 36 spills Where the barrels went share of all net barrels lost 31%61% 0.1% under 5 0.9% 5 to 50 6.7% 50 to 500 781,417 bbl total
Five ordered size classes of net barrels lost per incident, liquid systems only (gas reports no barrels). Top: each class's share of incidents. Bottom: the same classes, same shades, as a share of total volume.
The size classes, in a table
Size classSpillsShare of spillsBarrels lostShare of barrels
under 5 bbl 1,417 61% 454 0.1%
5 to 50 bbl 427 18% 7,204 0.9%
50 to 500 bbl 301 13% 52,667 6.7%
500 to 5,000 bbl 141 6.1% 245,077 31%
over 5,000 bbl 36 1.6% 476,015 61%
CS-11 / CAUSE

Why the Line Lets Go

Share of incidents vs share of barrels

The backhoe gets the headlines; the rust gets the pipe. Corrosion and material or weld failure - metal quietly giving up - account for about 33% of significant incidents and an even larger cut of the volume. Excavation damage runs the other way: 14.0% of incidents but only 22.1% of the barrels, because a struck line is found in minutes while a corroded seam can bleed for days. The dominant pipeline hazard is not an accident. It is age.

Share of significant incidents Share of net barrels lost
EQUIPMENT FAILURE 23.7% 15.6% CORROSION FAILURE 21.3% 9.3% EXCAVATION DAMAGE 14.0% 22.1% MATERIAL FAILURE OF PIPE OR WELD 10.1% 24.0% INCORRECT OPERATION 9.9% 5.7% OTHER OUTSIDE FORCE DAMAGE 9.0% 11.8% NATURAL FORCE DAMAGE 6.0% 8.8% OTHER INCIDENT CAUSE 3.2% 0.0% PIPE, WELD, OR JOINT FAILURE 1.7% 0.0% OTHER ACCIDENT CAUSE 1.2% 2.7%
Each cause's share of all significant incidents (dark bar) and of all net barrels attributed to a cause (orange bar), span totals, one shared scale.
The causes, in a table
CauseIncidentsShareBarrels lostShareNote
EQUIPMENT FAILURE 1,088 23.7% 122,042 15.6%
CORROSION FAILURE 977 21.3% 72,945 9.3%
EXCAVATION DAMAGE 645 14.0% 172,716 22.1%
MATERIAL FAILURE OF PIPE OR WELD 462 10.1% 187,532 24.0%
INCORRECT OPERATION 453 9.9% 44,274 5.7%
OTHER OUTSIDE FORCE DAMAGE 413 9.0% 92,260 11.8%
NATURAL FORCE DAMAGE 276 6.0% 68,908 8.8%
OTHER INCIDENT CAUSE 149 3.2% 0 0.0%
PIPE, WELD, OR JOINT FAILURE 80 1.7% 0 0.0%
OTHER ACCIDENT CAUSE 54 1.2% 20,740 2.7%
MD-12 / METHOD

Methodology

Notes on the Data

The figures on this page follow PHMSA Pipeline Safety - Incident & Accident Flagged Files (2010-2025 reports). PHMSA's "flagged files" are the bulk product the agency itself uses to build its twenty-year trends: every incident and accident report operators are required to file, plus the derived flags - significant, serious, and the harmonized cause categories - that let the records be compared across decades of changing report forms.

Full Every figure is a live aggregate of PHMSA's flagged files, 2010-2025.

What's real

This build is a real ingest. The state map, the operator ledger and scatter, the yearly trend and gauges, the era split, and the commodity, share, size-class, and cause readings are all counted straight from PHMSA's modern-era flagged files (2010-present) across all four system types - hazardous liquid, gas distribution, gas transmission and gathering, and LNG - 4,595 significant incidents in all. The path is the data layer only: src/lib/source.ts reads the flagged-file workbooks, scripts/build-data.ts aggregates the rows into this exact shape - including the era and spill-size slices - and HANDOFF.md lists the download steps. Drop a fresher flagged-file ZIP into data/raw/ and rerun npm run data to refresh, with nothing else changed.

The words, precisely

Significant incidents meet at least one threshold: a fatality or in-patient hospitalization; a fire or explosion; 5 or more barrels of highly volatile liquid or 50 or more barrels of other liquid; or $50,000-plus in damage measured in 1984 dollars. Serious is the narrower tier - a death or a hospitalization - and is a subset of significant. "Barrels lost" is net loss: volume released minus volume recovered. Gas systems report no barrels, so their rows sit at zero volume by definition, not by omission - which is exactly why this page keeps crossing the volume ledger against the death ledger.

How the derived slices are cut

The era bands split the span into two even halves (2010 through 2025) and sort each by system type: distribution, transmission, gathering, and LNG count as gas; crude, refined products, and HVL count as liquid. The spill-size classes cover liquid systems only and bin each incident by its net barrels lost. Shares in the slope and cause charts are computed within their own slice's total, so each chart is internally consistent against the figures it draws from.

What you're not seeing

This is the modern-era window: the flagged files reach back to 1986, but the pre-2010 report forms use incompatible columns and are not ingested here (the swap-point to backfill them is documented in HANDOFF.md). The 2010 rule revision, pinned on the trend chart, widened what operators must report near the start of the window. Gas gathering lines were only broadly brought into reporting in 2022, so the earlier band understates them and the recent gas share runs a little hot. Operator names change with mergers and are reported as the operator of record at the time, so a single company can appear under several names across the span. And an incident is a reported release, not a measure of long-term environmental or health harm, which no single dataset captures.


Generated 2026-07-10 12:59 UTC

Source: PHMSA Pipeline Safety - Incident & Accident Flagged Files