Critical Review: “Fish Remain High in Selenium Long After Mountaintop Coal Mines Close”

This above article¹ presents new selenium (Se) data in muscle tissue from three fish species in Crowsnest Lake (i.e., brown trout, lake trout, and mountain whitefish) and concludes that these data represent an environmental concern, based on the fact that the tissue Se concentrations exceed fish tissue guidelines. However, the article’s interpretation, contextual framing, and supporting evidence for these claims raise several scientific and methodological concerns.

The authors posit an apparent paradox, specifically, that low aqueous Se concentrations and elevated fish muscle Se concentrations. This is presented as an alarming ecological signal, and supports the thesis that the Se in these fish has accumulated over many years despite low aqueous concentrations. The assumption here is that Se has potentially cycled within the food web; however, a more fulsome historical understanding of this system is required to make this link. Specifically, were aqueous Se concentrations historically higher, and then declined more recently? If this were the case, comparing current water Se concentrations to current fish tissue concentrations is an “apples vs oranges” scenario. While elevated tissue Se in fish certainly warrants further investigation (i.e., are populations and/or fish health being impacted?), this ‘paradox’ is actually not inherently unusual. Selenium bioaccumulates, particularly in food webs involving algae, invertebrates and fish/birds, and tissue concentrations can remain elevated for years after aqueous exposures have declined, particularly in lentic waters with longer residence times. Without time-resolved aqueous and tissue Se data, it is at best speculative to interpret these findings as alarming, or even ecologically significant. The article does not present any data or evidence of historical Se concentrations in water or biota to support its claims of historical or ongoing accumulation.

The article presents new Se fish tissue data, but fails to reference any previous monitoring efforts in Crowsnest Lake or nearby waterbodies. Consequently, one cannot assess trends in Se (i.e., whether concentrations are increasing, stable, or anomalous). This is significant, since it would need to be demonstrated that high concentrations had been sustained in order for Se to bioaccumulate in aquatic biota tissues. Moreover, there is no attempt to address the mobility patterns of the fish species sampled, which is a significant omission, given the authors’ assertion that the Se in the system originated from the two decommissioned mines¹. For example, species like brown trout and mountain whitefish may use tributaries for spawning or feeding, potentially exposing them to elevated Se concentrations outside the lake. If this is the case, then the lake itself may not have been the primary source of Se accumulation. Without tools like radiotelemetry, tagging, or isotopic forensics, the source attribution to Crowsnest Lake or Tent Mountain is speculative.

The authors themselves acknowledge that the fish populations reported on in the paper (i.e., brown trout, lake trout, and mountain whitefish) in Crowsnest Lake are “self-sustaining”; they then suggest that current tissue Se concentrations may lead to a “population collapse” or “reproductive failure”. This contradiction is neither reconciled nor supported with data on recruitment, deformities, or reproductive metrics. The reference to such a collapse is not only unsupported, but actually conflicts with their own admission of population persistence. The article further suggests that Whirling Disease and Se toxicity could have overlapping symptoms, possibly confounding diagnoses. This suggestion is completely speculative and unsubstantiated. No peer-reviewed evidence to-date links Se-induced pathologies in fish to the clinical signs of Myxobolus cerebralis infection (the cause of Whirling Disease).

A central assumption of the paper is that Se is leaching into the lake from historical waste rock from the closed Tent Mountain mine. However, the authors provide no chemical fingerprinting, hydrological modeling, or sediment Se profiles to demonstrate a link. As such, the attribution again remains speculative and
circumstantial. Phrases such as “devastating consequences,” “complete reproductive failure,” and “acute threat” are used throughout the paper without any corresponding ecological data. These statements are alarmist and do not reflect the chronic, sub-lethal, nature of Se toxicity in aquatic ecosystems; it is well known that Se is not acutely toxic at environmental concentrations and does not cause sudden population collapses in the absence of sustained exposure and bioaccumulation.

Figure 2 compares Se concentrations across fish species and regions, but does not control for species-specific bioaccumulation potential, life stage, or habitat use. Comparing lake trout to other fish in other regions is both inappropriate and misleading. Species-specific comparisons are more relevant, and a consideration of whether the fish inhabit lentic vs lotic systems would make the comparison more valid and informative. Furthermore, no effort is made to explain the geological and ecological differences among the geographically-diverse reference sites in comparison to Crowsnest Lake. While the authors report Se concentrations in fish tissue, key details on analytical methods, quality control, and lab accreditation are not provided in the article. Without such information, confidence in the analytical results is difficult to evaluate fully. This is especially important in the case of Se, due to its low environmental concentrations and the complexity of accurately measuring low concentrations. As suggested, and given the significant implications of these data, independent verification or confirmation through expert review is highly recommended.

It should be noted that the article comprises a couple of important inaccuracies, specifically:

• the article refers to a CCME sediment Se guideline of 2 μg/g, which does not exist; Canada has not adopted sediment guidelines for Se. Additionally, the cited whole-body fish tissue guideline of 6.7 μg/g appears to be based on the ECCC Federal Environmental Quality Guideline (FEQG), and is not a
  CCME guideline. The FEQG applies specifically to egg/ovary Se (14.7 μg/g) and muscle Se only as a screening value. Mis-stating the origin and intent of these values is a significant error that undermines the scientific rigor of the article; and, 
• the article cites the 93% decline in Westslope Cutthroat Trout in the Fording River, but fail to cite the detailed, peer-reviewed Investigation of Cause (IOC) study, which attributed the decline to climatic conditions and not selenium toxicity.
• the article incorrectly implies that a $60 million fine was due to this decline. That fine, in fact, stemmed from broader non-compliance issues from 2012 – long before that decline occurred – and not a direct attribution to the population drop.

Summary

While it is understandable that elevated Se in fish tissue relative to low water concentrations raises a concern, the information provided in this paper fails to establish causality, provide historical or trophic context, or rigorously support its conclusions. The speculative connections to Tent Mountain, Whirling Disease, and catastrophic ecological effects are not supported by specific and relevant evidence. Misuse of terminology, misattribution of guideline values, and emotionally-charged language further erodes its credibility. 

A study of this type initially requires the clear identification and articulation of key research hypotheses (e.g., “is the elevated Se in fish tissue mine-related?”; “how is Se cycled in Crowsnest Lake”), and then determine what data/information are required to address those hypotheses. A multi-disciplinary, scientifically-defensible approach integrating fish movement, food web monitoring, sediment and water Se speciation, and reproductive success metrics, is essential to developing and reporting on scientifically-defensible conclusions.