Bitumen tailings spills in Canada and the failure of monitoring and management

Alberta is the world’s largest exploiter of bitumen. Production is concentrated near the banks of the Athabasca River in northeastern Alberta upstream of the globally significant Peace-Athabasca Delta (Figure 1). It would be reasonable to assume that the regulator, currently the Alberta Energy Regulator (AER), would have a scientifically credible system of spill monitoring and management. This study tests that assumption.

Background and setting

The AER is responsible for responding to and managing all energy industry activities, including gathering environmental information related to spills and other incidents. Operators are required to report to the AER any unapproved liquid spill greater than 2 m³ or any spill that flows off lease.

Spills of tailings are reported under an ad hoc variety of names, such as fluid fine tailings, mature fine tailings, coarse sand tailings, non-segregated tailings, bitumen slurry, bitumen froth, process water, and residual bitumen. The environmental toxicity of tailings is well established. Tailings are a significant source of air pollution, pose environmental hazards to birds and wildlife, contaminate groundwater, and pose risks of catastrophic dike failures.

Bitumen tailings are the liquid and solid residues remaining after bitumen has been removed from mined deposits. Tailings vary in their chemical and physical composition. Most commonly, they are mixtures of sand, silt, clay, water, and inorganic and organic compounds such as salts, gypsum, naphthenic acids, bitumen, volatile organic compounds, polycyclic aromatic hydrocarbons, and other chemicals either present in the mined deposit or added during processing.

Given the AER’s documented failure to detect, monitor, and mitigate the effects of crude oil and saline spills in Alberta (Timoney 2021), and following the public outcry in response to two major tailings spills at Imperial Oil Kearl lease in 2022 and 2023, a study of the AER tailings spills management is timely.

In April 2023, after a pilot study of tailings spills in the AER’s Field Inspection System (FIS) database, I observed that the reported data were, in short, too good to be true.

I therefore submitted a freedom of information request (FOIP) to the AER for all documents relating to the 514 tailings spills listed in the FIS record (Jan 2014 to Apr 2023). To facilitate analyses, Ecojustice Canada extracted the relevant spill information into a single excel spreadsheet.

The study answers several important questions. Are the AER’s claims of no environmental impacts from all tailings spills supported by evidence? If not, what is a plausible rate of harm?  Are claims of perfect spill recovery in 75 % of tailings spills supported by evidence? Are the reported tailings spill footprints credible? How many spills and what spill volumes are attributed to each of the companies?  Do the reported number of spills and total spill volumes accurately describe the actual spill rates? Are spill locations and incident dates accurate? Is the AER claim of routine inspections of spills supported by the data? 

This article summarizes the results of a detailed study of technical and often erroneous and conflicting information reported by industry to the AER. Readers wishing to examine the data to reach their own conclusions can request all the information via email. 

Substances released and recovered

Twelve substances were reported as released in association with the tailings spills. By number of spills and total volume released, tailings comprise the majority; process water and bitumen slurry were the second and third most numerous and largest releases.

Reported recovery rates were exceptionally high. Spill recovery followed an all or none pattern (24.2 % no recovery, 75.4 % perfect recovery). In practice, most spills would experience partial recovery (neither zero nor perfect recovery), but no partial recoveries were reported. The bimodal distribution for spill recovery demonstrates that the values are subjectively chosen, not the result of measurement.

Discrepancies in spill and recovery volumes between publicly available FIS data and internal documents were common. The most striking example of volume under-reporting is incident 20183682, reported in FIS as a tailings spill of 44,596 m³. The FOIP spill volume was 100 times larger: 4,459,680 m³, a volume that accumulated over a 326-day event. That single spill resulted in a release volume 15 times larger than the total reported FIS volume for all 514 tailings spills. The reason given by the AER for reporting the incorrect FIS volume to the public was: “The volumes tab does not allow a number that high to be captured.” The AER’s data files contain numerous other examples of erroneous data. 

Spill number and volume and companies responsible

Based on the FIS data, there were 77.4 reported spills per year along with an average annual spill volume of 41,990 m³. These totals are underestimates of the true rates and do not include releases due to tailings pond seepage.

The AER’s reported spill volumes, which serve as the official record, understate the true values. Spill volumes in the FOIPed documents are often higher than the spill volumes reported in the FIS record. Furthermore, the FOIPed documents record far more spills than revealed in the FIS data.

Many of the 514 FIS spills proved to be multiple spills subsumed under a single spill incident number. In large part the difference in totals stems from the AER’s poor bookkeeping and to its ad hoc manner of recording spill constituents. If all tailings spills are included, FIS failed to report 456 tailings spills (454 + 2 unreported Kearl spills leaked to the media).

Extraction of company names from the FOIPed documents allowed me to assign companies to the FIS tailings spills. Suncor and Syncrude were responsible for the most spills and the largest spill volumes. Suncor was responsible for 32 % of all tailings spills during the reporting period and 49 % of the total spill volume. Syncrude was responsible for 23 % of spills and 20 % of the spill volume. The FOIPed documents revealed that company names, although known by the AER, are withheld from the public record.

Incorrect spill dates

Incorrect, arbitrarily chosen occurrence dates are common. In 27 % of spills, the date was given as day 1 of the month. The likelihood of this pattern occurring by chance is vanishingly small. In 21 % of spills there was a 29-30 day delay between occurrence and notification, indicating that spills are not being reported to the AER until the end of the month. Because dates of spill occurrence are not credible, the questions arise: how long were the spills occurring before they were discovered, and what were the true volumes of the releases?

Incorrect spill locations

A precise location is the first requirement for monitoring and assessing the impact of spills. Due to inaccurate and imprecise locations, in no spills were the locations adequate to relocate the spills on the ground. In some cases, the spill locations were inaccurate by many kilometers.

Furthermore, discrepancies between spill locations reported in FIS and in FOIP are common. Three examples: Incident 20153108, FIS/FOIP location discrepancy: 9.8 km; incident 20221107, discrepancy 6.7 km; incident 20231024, discrepancy 5.5 km. These are large errors that the regulator should have detected.

Amazingly, spill locations use the location of the nearest facility, not the actual location of the spills. This practice means that many spill locations are in fact areas of multiple spills, but that no spills may have occurred at the reported location! Of the 514 spills, FIS provided only 115 unique locations; 12 sites were responsible for 326 spills. 

The degree to which the AER’s data can deviate from reality is illustrated in Figure 2. The FIS data state there was “no affect” [sic] on wildlife; the area affected </= 100 m². A FOIPed photograph allowed me to pinpoint the actual spill location. Plotting of the footprint within a GIS indicated an affected area of ~5388 m², larger than the FOIP-reported 2320 m², and more than 50 times larger than the FIS-reported footprint.

The FOIP photos and documentation show extensive damage and contamination of soil and vegetation and the finding of two dead birds. No information was provided as to the results of the Syncrude investigation into the bird deaths. To summarize, a serious spill occurred without an inspection by the AER; the spill volume, location, footprint, and effects on the environment were erroneous.

Spill footprints

The AER’s default value for spill footprints is “100 square meters or less”, and this is used even when documents show the footprint to be larger. Despite assigning a spill footprint for all spills, in 93 % of spills the affected area was not measured.  Based on the relationship between spill volume and footprints, it is probable that at least two-thirds of tailings spills exceed the reported spill footprint.

The claims that in 511 of the 514 spills the tailings spill footprint was </= 100 m² are not supported by credible evidence. The FOIP reported 35 spills with footprints of 150 to 30000 m². For those spills, the FIS-reported footprints of </= 100 m² are contradicted by the AER’s own documents (Figure 3).

Environmental impacts and inspections

The regulator’s FIS data on wildlife and habitat effects report no environmental impacts at all 514 tailings spills. However, AER spill reports lack scientific rigor, as they are produced by company staff trained in bitumen production, not by environmental experts. Spill reports do not assess ecosystems, biota, or chemistry.

The AER’s claim of no environmental impacts is unsubstantiated by evidence. Of 126 spills with photographic evidence, 41-54% showed harm. If this harm rate applies to spills without photos, environmental damage is inferred in 23-36% of those. Many spills occurred in areas already impacted by industrial activities, making it hard to distinguish new harm.

The AER’s failure to assess environmental effects is illustrated by four spills.

Incident 20211344 (Figure 4): A 6500 m³ spill at Imperial Oil Kearl was reported as having no environmental impact. The spill’s footprint was 27 times larger than reported by FIS. Vegetation died and soil was contaminated. No enforcement action was taken despite a six-month reporting delay. The reported perfect recovery was not credible.

Incident 20221305 (Figure 5): At Syncrude Mildred Lake, photos showed clear environmental impacts. Despite FIS reporting a single spill with minimal footprint, FOIP documents listed 13 spills with a volume 12 times larger than reported. Poor record-keeping and unacknowledged impacts are evident. 

Incident 20222103 (Figure 6): A 2700 m³ spill at Syncrude Aurora North was reported with perfect recovery, but the actual footprint was >17 times larger. Photos show significant damage to streams, vegetation, and soil. Despite claims of wildlife deterrents, no evidence was found.

Incident 20230872 (Figure 7) was a spill of 200 m3 tailings with “no affect” [sic], a footprint <100 m2, and “perfect recovery” at CNRL (Figure 6). FOIP documents the spill as 2000 m3, not 200 m3.  Photos document extensive environmental damage yet no environmental data were gathered. Claims of perfect recovery, no environmental effect, and small footprint were not credible. There was no evidence of bird deterrents provided but CNRL stated deterrents were deployed. “Reclamation” involved the obliteration of the entire wetland (6586 m2) with heavy machinery.

The AER claim of routine inspections of spills is not supported by data. Only 3.2 % of tailings spills are inspected by the AER.

Conclusion

The FOIP documents contain all available information on spill volumes, recovery, footprints, locations, dates, and environmental impacts unless materials were unlawfully withheld. The research reveals that the AER lacks the data needed to assess spill effects. Its claims of no environmental impacts are unsupported.

The AER’s failure to ensure companies gather credible environmental data, conduct monitoring and inspections, and protect ecosystems has resulted from regulatory capture. This failure reflects a system designed to maintain the appearance of control while prioritizing fossil fuel production over ecosystem protection.

Despite claiming to protect the environment, the AER has acted as a captured regulator serving the fossil fuel industry. For decades, Alberta’s energy regulators have failed to provide reliable, truthful, or timely information and have neglected monitoring, enforcement, and environmental protection duties.

The AER’s monitoring and assessment of tailings spills lack scientific credibility. Although significant environmental harm from spills is evident, the lack of environmental data means that the true magnitude of the ecological impacts of tailings spills may never be known.

Recommendations

The Alberta Energy Regulator should be replaced and staffed with competent environmental scientists, engineers, and managers, funded independently of the fossil fuel industry, and overseen by an elected public body with no ties to industry. 

The Responsible Energy Development Act needs to be repealed and replaced with new legislation that will protect the public and the environment rather than serve the interests of the fossil fuel industry.

The United Nations, environmental justice organizations, and First Nations need to become better informed about the AER’s and industry’s absymal record of monitoring and management of environmental impacts. They should work together to replace the AER. 

The Canadian federal government should enforce its Competition Act to curtail the false environmental claims made by the fossil fuel industry and the Alberta Energy Regulator.

🔬🧫🧪🔍🤓👩‍🔬🦠🔭📚

Journal reference

Timoney, K. P. (2025). Regulatory failure to monitor and manage the impacts of tailings spills, Alberta, Canada. Environmental Monitoring and Assessment, 197(2), 125. https://doi.org/10.1007/s10661-024-13416-1