Historical spill counts have risen sharply since 2021

The number of historical spill records filed each year was relatively flat from 2016 through 2020, then accelerated dramatically. From 2020 to 2025 filings increased more than fifteenfold, from 125 to 1,948 unique spills per year. This reflects a wave of decommissioning activity across Colorado’s aging Denver-Julesburg Basin well fleet, as older wells reach the end of their productive lives and operators are required to plug, cap, and remediate them before closure.

Weld County accounts for 87 percent of all historical spills

The geographic distribution of historical spills is highly concentrated. Weld County, the heart of Colorado’s Denver-Julesburg Basin oil and gas activity, is where 5,303 of the 6,084 total historical spills reported occurred. The next largest county, Adams, has just 187. This reflects both the density of oil and gas development in the Denver-Julesburg Basin and the scale of ongoing abandonment and reclamation activity there.

Three operators account for 76 percent of historical spill records

Noble Energy, Kerr McGee, and PDC Energy together account for nearly three-quarters of all historical spill records on file. All three are major Denver-Julesburg Basin operators that own large numbers of legacy wells. Noble Energy was acquired by Chevron in 2020, and Kerr McGee’s assets were absorbed into Crestone Peak Resources and then Civitas Resources, meaning the responsibility for many of the historical spills in this report now rests with these companies.

Tank batteries and wells are the most common failure points

Tank batteries (33 percent) and wells (30 percent) together account for nearly two-thirds of historical spill sites, with flowline systems contributing another 19 percent. Almost all historical spills, 96 percent, occurred on private land, with federal and state lands representing a small fraction. This has implications for landowner notification and long-term remediation obligations.

Why are historical spills being discovered now?

41 percent — Plugging, abandonment, and cut-and-cap operations. When a well is being permanently closed, either through a standard plugging process or under a state-ordered procedure to seal off the wellhead, crews often discover contaminated soil underneath the wellhead, separator, or underground vault. The contamination is usually old, with no clear record of how much spilled, and cleanup is planned as a separate, later project. This is the single largest discovery trigger in the dataset.

22 percent — Tank battery deconstruction. Workers deconstructing a tank battery find contaminated soil beneath produced water vaults, dump lines, or tanks. The contamination is almost always described as pre-existing with no active leak at the time of discovery.

13 percent — Flowline decommissioning. When companies remove old pipelines from service, they often find oil-contaminated soil at pipe bends or spots where sensors detect hydrocarbons in the ground. Reports frequently note that an exact release point cannot be determined due to the historical nature of the contamination.

7 percent — Facility decommissioning. Closure of production facilities, like separators, compressor stations, and other surface equipment, triggers soil assessment that uncovers historical contamination not previously documented.

6 percent — Construction and excavation activities. Routine construction work, including new pipelines, surface upgrades, and road grading, unexpectedly turns up contaminated soil. These are typically incidental discoveries rather than targeted investigations.

2 percent — Site reclamation and pit closure. Closing an old production pit or conducting a planned remediation project triggers excavation that exposes legacy contamination not previously on record.

<1 percent — Regulatory inspection. A state field inspector spots pooled fluids or stained soil during a routine site visit and initiates a report. These represent a very small share of discoveries, suggesting inspection capacity is not the primary driver of detection.

Note: Percentages are of classified records only; 7 percent of records could not be classified by discovery reason and are excluded from the chart above.

Why the sharp rise in filings from 2021 onward?

The more-than-fifteenfold increase in historical spill filings between 2020 and 2025 is not primarily a sign that more spills are happening. It reflects several converging forces that are accelerating how quickly legacy contamination gets documented. ECMC itself reaches the same conclusion: the agency’s 2025 Director’s Report on the Evaluation of Cumulative Impacts attributes the steady year-over-year climb in reported spills since 2020 largely to 2021 rule changes that lowered spill reporting thresholds and required operators to report historical releases discovered during facility decommissioning.

The SB 19-181 “Mission Change” rulemaking made historical spills reportable

Colorado’s landmark 2019 oil and gas reform law directed the ECMC to prioritize public health and environmental protection, leading to the “Mission Change” rulemaking, a comprehensive rewrite of state oil and gas rules effective January 15, 2021. The rewritten Rule 912 added discovery-based reporting triggers on top of the old volume-based thresholds: operators must now report the discovery of 10 cubic yards or more of impacted material from a current or historic spill, as well as any impacted groundwater or surface water, without waiting for confirmation sampling. Companion changes to Rules 911 and 913 expanded closure documentation requirements—previously focused on pits and partially buried vessels—to all related facilities when a well is plugged and abandoned, including a Form 27 site investigation and remediation workplan at facility closure. As ECMC’s 2025 cumulative impacts report acknowledges, these threshold and closure-documentation changes are a principal driver of the rise in reported spills—legacy contamination that previously went undocumented must now enter the record.

An aging well fleet entering mass decommissioning

The Denver-Julesburg Basin saw a major drilling boom in the 2010s. Many of those wells, and the older legacy wells predating the boom, are now reaching the end of their productive life. When an operator plugs and abandons a well or deconstructs a tank battery, ECMC rules, updated in the Mission Change rulemaking, require environmental assessment of the site. As the volume of plugging and abandonment activity has climbed sharply since 2021, so has the volume of historical releases being uncovered and documented. ECMC’s 2025 cumulative impacts report confirms the pattern statewide: of roughly 2,300 spills and releases reported to ECMC in 2025, about 85 percent were historical spills, mostly discovered during facility decommissioning and well plugging—only 337 were recent or active spills.

Cut-and-cap program scaling up

ECMC’s updated “cut and cap” rules require operators to follow a specific set of procedures when plugging decommissioned wells. As this program has ramped up, especially across Weld County’s large inventory of older wells, each site assessment completed under the updated rules is an opportunity for industry to find and report soil contamination that has been present but unreported for years.

Operator consolidation surfacing inherited liabilities

Several large acquisitions in the Denver-Julesburg Basin, notably Chevron’s purchases of Noble Energy in 2020 and PDC Energy in 2023, Occidental’s absorption of Kerr-McGee via its 2019 Anadarko acquisition, and the 2021 three-way merger of Bonanza Creek, Extraction, and Crestone Peak that created Civitas Resources (now part of SM Energy), brought new scrutiny to inherited well portfolios. Post-acquisition due diligence and accelerated decommissioning programs have pushed large backlogs of legacy contamination into the ECMC reporting system.

What the trend tells us

The rise in historical spill filings is largely good news from a regulatory standpoint. It means the stronger rules related to decommissioning and remediating wells are working—a view ECMC shares, noting in its 2025 cumulative impacts report that finding and remediating large numbers of historical spills during site closures is ultimately positive for protecting groundwater and the surface environment. But it also underscores how much historic contamination was not caught in real time, and how much legacy liability remains embedded in Colorado’s landscape. Finally, the fact that less than one percent of historical spills are caught by state inspectors reveals a need for increased on-the-ground oversight of both active and inactive well sites by state regulators.

 Data source: ECMC (Energy and Carbon Management Commission, formerly COGCC), downloaded from the ECMC online spill database.

Record selection: Filtered to rows where Spill Type = “Historical” and Date of Discovery falls between January 1, 2016 and December 31, 2025. Records with a Date of Discovery before 2016 (9 records) or in 2026 (674 records bearing future-dated entries present in the source file) are excluded. Each unique Tracking # is counted as one spill event; where multiple rows share a Tracking #, the first occurrence by row order is used for categorical fields. Note: the 2016 data begins in May; January through April have no records, likely reflecting the period before ECMC’s historical spill reporting requirements took effect. The 2016 count therefore represents a partial year.

Spill count: 6,084 unique historical spill records after deduplication and date filtering.

Volume: Most historical spill records do not include a reported volume; the release typically occurred years or decades before discovery, making volume estimation impossible. Volume figures are not reported in this analysis.

Discovery reason classification: The Root Cause and Spill Description fields were analyzed using keyword text classification to categorize each spill by the activity that led to its discovery. Categories and counts: plugging, abandonment, and cut-and-cap operations (41 percent); tank battery deconstruction (22 percent); flowline decommissioning (13 percent); facility decommissioning (7 percent); construction and excavation (6 percent); site reclamation and pit closure (2 percent); ECMC field inspection (<1 percent); unknown or not reported (7 percent). Percentages are of 6,084 unique tracking numbers.

Inorganic-only filings: Approximately 28 of the 6,084 records (less than 0.5 percent) appear to involve inorganic-only soil parameter exceedances — such as elevated pH, sodium adsorption ratio (SAR), or arsenic — rather than hydrocarbon releases, based on zero reported liquid volume combined with descriptions referencing inorganic constituents or background-level determinations. These records are retained in the dataset because ECMC does not provide a formal mechanism to distinguish them from confirmed hydrocarbon releases. Their inclusion has no material effect on any figure in this report.