DERR-2024-007802

Implications of Selenium in Proposed Wastewater Discharges to Great Salt Lake
prepared by
A. Dennis Lemly, Ph.D.
Senior Scientist in Aquatic Toxicology
USDA-Forest Service
Southern Research Station
Coldwater Fisheries Research Unit
Blacksburg, Virginia
prepared at no cost for
Friends of Great Salt Lake
Salt Lake City, Utah
January 30, 2004
Maximum Acceptable Selenium in Waters of Great Salt Lake
Among aquatic ecosystems, those that are shallow and slow moving or static (lentic) are
most likely to accumulate selenium and experience toxic impacts in fish and wildlife (Lemly
1
2002a). The Great Salt Lake (GSL) fits into this category of ecosystems and thus carries a
relatively high risk of selenium poisoning. It is therefore crucial to set environmentally sound limits
on waterborne selenium for GSL, based on real-world case examples of selenium poisoning, in
order to keep accumulation below levels that threaten aquatic life.
The traditional approach to establish limits on waterborne selenium has been to simply
adopt the USEPA national freshwater criterion (5 _g/L, parts-per-billion). The environmental basis
for this number comes from Belews Lake, NC, where fish populations in areas of the lake with
waterborne selenium >5 _g/L suffered major toxic impacts (Lemly 1985, USEPA 1987). At the
time this criterion value was established (mid-1980's), little data were available from other aquatic
systems to evaluate how broadly applicable the criterion actually was. Since that time, a growing
body of scientific information has emerged, including additional information from Belews Lake,
which indicates that toxic impacts to aquatic life can occur when selenium levels approach 2 _g/L
(Frankenberger and Engberg 1998, Skorupa 1998a, Lemly 1997, Hamilton and Lemly 1999;
references by Skorupa and Hamilton included as attachments). This lower toxic threshold is
especially likely if waterborne selenium is predominantly in the selenite form or has been
biologically processed (i.e., selenate incorporated into the food chain, deposited as detritus, recycled
into water), or if the contaminated water enters a river backwater, wetland, pond, lake, reservoir, or
other impoundment (Lemly 2002a). Because of these findings, a value of 2 _g /L has been
recommended by several selenium experts as the concentration limit necessary to protect fish and
wildlife (Peterson and Nebeker 1992, Maier and Knight 1994, Skorupa 1998b, Hamilton and
Lemly 1999, Lemly 2002a, Hamilton 2004), and USEPA has begun a review/revision process for
their national freshwater criterion (USEPA 1998, Hamilton 2003). Moreover, based on broad
experience dealing with a variety of selenium contamination issues, the U.S. Fish and Wildlife
Service and a number of state water quality agencies have adopted a value of 2 _g/L or less as their
management or regulatory standard (see Engberg et al. 1998, Skorupa 1998b, Hamilton and Lemly
1999, USFWS 2003). An especially pertinent case example to draw upon for guidance is the
Salton Sea, CA (a saline ecosystem quite similar to GSL), where only 1.5 _g/L waterborne selenium
accumulated to toxic concentrations in the aquatic food chain (Skorupa 1998a). I recommend that 2
_g /L be adopted as the maximum acceptable concentration of selenium in ambient waters of GSL,
and that this value be designated as a site-specific standard applicable to GSL and adjoining
wetlands, irrespective of standards that may be in place for other surface waters of the state (e.g., 5
_g/L). Further, I recommend that this same concentration limit be imposed on all wastewater and
drainage discharge entering GSL, with no provision that mixing/dilution be allowed in order to meet
the 2 _g/L standard in ambient waters.
Guidance for Implementing a 2 _g/L Selenium Limit
USEPA provides states the option of developing site-specific standards and regulations
governing their implementation if adequate justification is given. Establishing standards that are
sensitive to selenium’s local variations in aquatic cycling and toxic effects should be the goal of
2
Utah’s regulatory agencies regarding GSL. Evidence now available for selenium bioaccumulation
and impacts in western wetlands and saline lakes similar to GSL indicates that setting 2 _g/L as the
standard would be a prudent step to take. However, in addition to “setting a number” it is also
important to consider how standards will be implemented and enforced — i.e., will there be
provisions for averaging concentrations over several days in order to meet standards, will periodic
exceedance of standards be allowed, and will mixing zones be used to dilute selenium inputs and
achieve standard limits. The state may seek to follow the example set by USEPA (e.g., Stephan et
al. 1985, USEPA 1985, USEPA 1987), which allows averaging, exceedances, and mixing zones to
factor into the implementation of standards. However, the guidance provided by USEPA is generic,
and it would greatly reduce the effectiveness of a site-specific selenium standard for GSL. There
are several important precautions that should be taken by water quality regulators if they intend to
rely on USEPA guidance documents to implement site-specific selenium standards.
USEPA guidelines allow the national criterion for chronic exposure to be exceeded
periodically (once every three years, on average) as long as the four-day average concentration is 5
_g Se/L or less (USEPA 1987). During exceedances, the permissible ambient (ecosystem-wide)
waterborne concentration can be as high as 20 _g Se/L (Figure 1). Stephan et al. (1985) gives the
rationale for this approach: “the averaging period of four days was selected by the USEPA on the
basis of data concerning how rapidly some aquatic species react to increases in the concentrations
of some pollutants, and three years is the Agency’s best scientific judgement of the average amount
of time aquatic ecosystems should be provided between excursions”. The wording of the statement
reveals that this is a generic model for contaminant exposure-response and associated derivation of
criteria–i.e., the words “some aquatic species” and “some pollutants”. The model was developed
in the early 1980's when there was relatively little field data on selenium cycling and
bioaccumulation in aquatic systems, and no attempt has been made by USEPA to test it’s
assumptions using selenium data that have become available since that time.
3
Figure 1. Contrasts between the existing USEPA national water quality criterion for selenium (chronic
exposure, A) and the process recommended for setting a site-specific standard for GSL (B). A principal
difference is that in implementing USA national criteria, USEPA guidelines allow four-day averages and
exceedances up to the Criterion Maximum Concentration (CMC, 20 _g Se/L). This can offset the
protection to aquatic life that is afforded by the Criterion Continuous Concentration (CCC, 5 _g Se/L). To
provide full protection, the site-specific standard for GSL should be set using a biologically-based
concentration limit (2 _g/L) as the CMC, and not allow averages, exceedances, or mixing zones.
There are 4 specific flaws that invalidate the USEPA model when it is applied to selenium:
(1) The USEPA guidance document clearly indicates that the process for USA national criteria is
molded to fit publicly owned wastewater treatment facilities (POTWs) that discharge a point source
into a flowing receiving water (Stephan et al. 1985, pp. 11-12). However, the most widespread
threats of selenium poisoning in aquatic habitats are in lentic systems (reservoirs, wetlands, and off-
channel bays and impoundments), and are due to power plant discharges, agricultural irrigation, and
other sources, not POTWs. The environmental dynamics of selenium in lentic ecosystems is quite
different than the riverine conditions used for the USEPA model.
(2) The
four-day
average is
based on
organism
responses
to
waterborne exposure alone. However, food chain bioaccumulation and dietary intake are more
important in causing chronic selenium toxicity to aquatic life (Lemly 1985, Lemly 1997). This
component of selenium cycling is overlooked in the USEPA model. Moreover, exposure-
bioaccumulation-response times for selenium in fish and aquatic birds (waterborne or dietary
intake) are on the order of weeks or months rather than four days (e.g., Lemly 1982, Heinz,
Hoffman and Gold 1988, Coyle et al. 1993, Heinz and Fitzgerald 1993)–the USEPA model
assumptions are not correct.
4
(3) The concentrations of waterborne selenium allowed by USEPA during exceedances (up to 20
_g Se/L) are not environmentally acceptable for lentic systems or lotic systems that will deliver
selenium into off-channel bays, wetlands, lakes, reservoirs, or other down-gradient lentic systems.
Studies such as those by Cumbie and Van Horn (1978), Bryson et al. (1984), Lemly (1985),
Gillespie and Baumann (1986), and Hamilton et al. (1996) show that concentrations of 10-20 _g
Se/L can quickly reach dietary levels that are toxic to fish and aquatic birds. Consider, for example,
a scenario in which an exceedance causes waterborne selenium in a lake or wetland to reach 15 _g
Se/L–an acceptable concentration in the USEPA model (Figure 1). By the time ambient locations
reach this level, the entire “bioaccumulation engine” of the ecosystem will have been fueled by the
influx of new selenium, which substantially escalates the toxic threat to aquatic life (Lemly 1985).
(4) The three year period between excursions (exceedances), although perhaps reflecting the best
scientific judgement available for some pollutants in the early 1980's, is not appropriate for
selenium given today’s knowledge of the environmental dynamics and cycling of this trace element.
Once an aquatic ecosystem has captured the selenium dose delivered by an exceedance, it can
continue to cycle it tightly within the system for many years. For example, studies show that the
recovery period for reservoirs contaminated by 10 _g Se/L selenium could be >10 years, perhaps
several decades, due to recycling of selenium from sediments into benthic-detrital food chains and
associated dietary and reproductive toxicity to fish (Garrett and Inmann 1984, Lemly 1997).
Similar problems are evident with the use of dilution or mixing zones, which are areas
exempt from ambient standards. This concept was developed for application to flowing waters
(Stephan et al. 1985). It has no credible basis for application to selenium in lentic/wetland systems
because the “dilution zone” may constitute the entire body of open water (Lemly 1985, 1997).
Even in riverine habitats, the notion of mixing zones is a totally artificial process because USEPA
has not referenced data verifying that a mixing zone can effectively dilute a selenium-laden effluent
and also be environmentally compatible with fish and wildlife habitat uses, which it must be under
federal statutes in the USA, such as the Migratory Bird Treaty Act and the Endangered Species Act
(Margolin 1979). Selenium strongly bioaccumulates in food organisms and makes the dilution
zone an area of extremely high exposure for fish and wildlife. Several case studies show that using
mixing zones to dilute seleniferous water creates more biological hazards than it resolves (e.g.,
Skorupa 1998). The apparent benefits gained by achieving target concentrations in a mixing zone
may be more than offset by detrimental effects that are caused by other aspects of the selenium
cycle. The threat of toxic impacts overrides the need to attempt “dilution as a solution”,
particularly in shallow, saline habitats such as GSL.
Given these flaws, it is important to closely examine the rationale for, and distinction
between, national criteria and site-specific standards. USEPA criteria are intended to provide
protection for most aquatic species most of the time, not everything all of the time (Stephan et al.
1985). Because of this basic caveat, as well as the fact that there are differences in ecosystem and
aquatic species sensitivity to selenium, there may be a plausible argument for allowing some leeway
in meeting the national criterion–i.e., a reasonable averaging of concentrations over time if
5
monitoring indicates that there are no biological effects (but not 20 _g Se/L exceedances). However,
at a local level, the national criterion’s intent to protect “most species” still leaves large gaps that
could lead to substantial inconsistencies (toxic effects at or below the criterion level, lack of effects
above the criterion). Site-specific water quality criteria, i.e., the standard for GSL, should reflect the
sensitivity of local biota and close the gaps. If full protection of aquatic life is desired, then there
should be no provision for averages, exceedances, or mixing/dilution zones in the implementation of
standards.
Concentration versus Loading Considerations
Because of selenium’s propensity to accumulate in sediments and recycle into the aquatic
food chain, actions to regulate this pollutant must take loading into account (the total mass of
selenium) as well as the waterborne concentration (amount per unit volume). Thus, in addition to
the recommended 2 _g/L waterborne standard, a limit should be set on the loading of selenium to
GSL. The key factor in setting a proper loading limit, or Total Maximum Daily Load (TMDL) is to
determine the ability of the ecosystem in question to take up waterborne selenium and hold it in
biota or detritus/sediments. This is known as the system’s retention capacity (RC). The more that
selenium is held in the system, the higher the RC. It is necessary to know the RC in order to
develop an environmentally sound TMDL because the higher the RC, the lower the TMDL has to
be to prevent toxic threats to fish and wildlife. A published procedure is available for determining
RC based on primary productivity, water flow regime, and sediment type (Lemly 2002b; reference
provided as attachment). Applying this procedure to GSL is a simple process. Without TMDLs,
seemingly harmless waterborne concentrations of selenium (i.e., those that meet regulatory
standards) may build up in sediments, recycle into aquatic food chains, and cause unforseen toxic
impacts (Lemly 2002c).
Regardless of what the waterborne standard and TMDL are, it is important to monitor
selenium concentrations to make sure these regulatory steps are keeping selenium concentrations
below levels of concern in fish and wildlife. Monitoring provides a critical feedback loop that is
needed for on-going assessment of the health of the GSL ecosystem. The following concentration
limits are suggested as guidelines for interpreting the monitoring data (Lemly 2002a).
Table 1. Toxic effect thresholdsa for selenium in aquatic ecosystems.
_________________________________________________________________________________
Selenium source Selenium Effect
concentrationb
_________________________________________________________________________________
Water
Inorganic selenium 2 _g/L Food-chain bioaccumulation

and reproductive failure
in fish and aquatic birds
6
Organic selenium < 1 _g/L Food-chain bioaccumulation
and
reproductive failure
in fish
and aquatic birds
Sediments 2 _g/g Food-chain bioaccumulation
and reproductive failure
in fish
and aquatic birds
Food-chain organisms 3 _g/g Reproductive failure in
fish and
aquatic birds
Fish tissues
Whole-body 4 _g/g Mortality of juveniles and

reproductive failure
Skeletal muscle
(skinless fillets) 8 _g/g Reproductive failure
Liver 12 _g/g Reproductive failure
Ovary and eggs 10 _g/g Reproductive failure
Aquatic Birds
Liver 10 _g/g Reproductive failure
Eggs 3 _g/g Reproductive failure
_________________________________________________________________________________
a These are levels at which toxic effects begin to occur in sensitive species of fish and aquatic
birds. They are not levels that signify the point at which all species die from selenium poisoning.
b Selenium concentrations in parts per billion for water; parts per million on a dry weight basis for
sediments, food-chain organisms, and fish and bird tissues.
Implications for Wastewater Proposed for Discharge into GSL
The reverse osmosis wastewater produced by treating contaminated groundwater from
Kennecott Utah Copper Corporation activities constitutes a highly hazardous waste with regard to
its selenium content. The concentrations of selenium expected in discharges (7-48 _g/L; State of
Utah, 2003), in combination with the volume of discharge (19.2 MGD), will be a major source of
selenium pollution to GSL. Without question, the discharge will create an area of highly
contaminated water, sediments, and associated aquatic life that will expand over time. The extent of
this expansion will depend on: (1) how long selenium-laden wastewater is discharged, and (2)
limnological factors in GSL such as primary and secondary productivity, wind circulation and
mixing, and biological activity in sediments. Hydrological connections with the main body of GSL
7
could lead to contamination of associated wetlands that are managed for waterfowl or other fish and
wildlife uses. Hydrological mapping would provide clues as to where selenium movement may
occur and allow other habitats that are at risk to be identified (Lemly 2002a).
Discharging the proposed amounts of selenium into GSL carries great environmental risk.
Applying an EPA or state standard of 5 _g/L to GSL is a mistake. The number is badly outdated
and the implementation guidance that is available from EPA is fatally flawed. Evidence from similar
ecosystems in the West shows that as little as 1.5 _g/L can cause problems. Using GSL as a
dilution chamber will create areas of high selenium exposure to fish and wildlife. It is crucial to
understand that when selenium contamination begins, a cascade of bioaccumulation events is set
into motion which makes meaningful intervention nearly impossible — once the threshold is
crossed, toxic impacts will occur. However, this cascade of events need not happen in GSL if
adequate foresight and planning are exercised. Regulatory authorities in Utah have the opportunity
to exercise this foresight, which includes consideration of other disposal options (e.g., alternative
discharge locations such as Kennecott Utah property). Prudent risk management based on
environmentally sound hazard assessment and water quality goals can prevent biological impacts.
If other options are not chosen and GSL is used as the discharge location, it is essential that: (1) a 2
_g/L standard be imposed on selenium in the discharge itself, (2) a TMDL for total selenium
entering GSL be derived, and (3) biological monitoring be conducted to make sure selenium
concentrations remain below levels of concern for fish and wildlife.
References
Bryson, W.T., Garrett, W.R., Mallin, M.A., MacPherson, K.A., Partin, W.E., and Woock, S.E.
1984. Roxboro Steam Electric Plant 1982 Environmental Monitoring Studies. Volume 2.
Hyco Reservoir Bioassay Studies. Technical Report. Carolina Power and Light Company,
New Hill, NC.
Coyle, J.J., Buckler, D.R., Ingersoll, C.G., Fairchild, J.F., and May, T.W. 1993. Effect of dietary
selenium on the reproductive success of bluegills (Lepomis macrochirus). Environmental
Toxicology and Chemistry 12:551-565.
Cumbie, P.M., and S.L. Van Horn. 1978. Selenium accumulation associated with fish mortality
and reproductive failure. Proceedings of the Annual Conference of the Southeastern
Association of Fish and Wildlife Agencies 32: 612-624.
Engberg, R.A., D.W. Wescot, M. Delamore, and D.D. Holz. 1998. Federal and state perspectives
on regulation and remediation of irrigation-induced selenium problems. Chapter 1 (pages
1-25) in W.T. Frankenberger, Jr., and R.A. Engberg, editors. Environmental Chemistry of
Selenium. Marcel Dekker, Inc., New York, NY.
Frankenberger, W.T., Jr., and R.A. Engberg. 1998. Environmental Chemistry of Selenium.
Marcel Dekker, Inc., New York, NY.
Garrett, G.P., and Inmann, C.R. 1984. Selenium-induced changes in the fish population of a
heated reservoir. Proceedings of the Annual Conference of the Southeastern Association of
Fish and Wildlife Agencies 38:291-301.
8
Gillespie, R.B., and Baumann, P.C. 1986. Effect of high tissue concentrations of selenium on
reproduction by bluegills. Transactions of the American Fisheries Society 115:208-213.
Hamilton, S.J., Buhl, K.J., Bullard, F.A., and McDonald, S.F. 1996. Evaluation of Toxicity to
Larval Razorback Sucker of Selenium-Laden Food Organisms from Ouray NWR on the
Green River, Utah. Technical Report. U.S. Geological Survey, Midwest Science Center,
Ecotoxicology Research Station, Yankton, SD.
Hamilton, S.J., and A.D. Lemly. 1999. Water-sediment controversy in setting environmental
standards for selenium. Ecotoxicology and Environmental Safety 44: 227-235.
Hamilton, S.J. 2003. Review of residue-based selenium toxicity thresholds for freshwater fish.
Ecotoxicology and Environmental Safety 56: 201-210.
Hamilton, S.J. 2004. Selenium toxicity in the aquatic food chain. Science of the Total
Environment (in press).
Heinz, G.H., Hoffman, D.J., and Gold, L.G. 1988. Toxicity of organic and inorganic selenium to
mallard ducklings. Archives of Environmental Contamination and Toxicology 17:561-568.
Heinz, G.H., and Fitzgerald, M.A. 1993. Overwinter survival of mallards fed selenium. Archives
of Environmental Contamination and Toxicology 25:90-94.
Lemly, A.D. 1982. Response of juvenile centrarchids to sublethal concentrations of waterborne
selenium. 1. Uptake, tissue distribution, and retention. Aquatic Toxicology 2:235-252.
Lemly, A.D. 1985. Toxicology of selenium in a freshwater reservoir: Implications for
environmental hazard evaluation and safety. Ecotoxicology and Environmental Safety 10:
314-338.
Lemly, A.D. 1997. Ecosystem recovery following selenium contamination in a freshwater
reservoir. Ecotoxicology and Environmental Safety 36: 275-281.
Lemly, A.D. 2002a. Selenium Assessment in Aquatic Ecosystems: A Guide for Hazard
Evaluation and Water Quality Criteria. Springer-Verlag Publishers, New York, NY.
Lemly, A.D. 2002b. A procedure for setting environmentally safe Total Maximum Daily Loads
(TMDLs) for selenium. Ecotoxicology and Environmental Safety 52: 123-127.
Lemly, A.D. 2002c. Symptoms and implications of selenium toxicity in fish: The Belews Lake
case example. Aquatic Toxicology 57: 39-49.
Maier, K.J., and A.W. Knight. 1994. Ecotoxicology of selenium in freshwater systems. Reviews
in Environmental Contamination and Toxicology 134: 31-48.
Margolin, S. 1979. Liability under the Migratory Bird Treaty Act. Ecology Law Quarterly 7:989-
1010.
Peterson, J.A., and A.V. Nebeker. 1992. Estimation of waterborne selenium concentrations that are
toxicity thresholds for wildlife. Archives of Environmental Contamination and Toxicology
23: 154-162.
Skorupa, J.P. 1998a. Selenium poisoning of fish and wildlife in nature: Lessons from twelve real-
world examples. Chapter 18 (pages 315-354) in W.T. Frankenberger, Jr., and R.A.
Engberg, editors. Environmental Chemistry of Selenium. Marcel Dekker, Inc., New York,
9
NY.
Skorupa, J.P. 1998b. Selenium. Pages 139-184 in P.L. Martin and D.E. Larsen, editors.
Guidelines for interpretation of the Biological Effects of Selected Constituents in Biota,
Water, and Sediment. National Irrigation Water Quality Program Information Report No.
3. U.S. Department of the Interior, Denver, CO.
State of Utah. 2003. Permit No. UT0025551 – Authorization to Discharge Under the Utah
Pollutant Discharge Elimination System (UPDES). State of Utah, Department of Water
Quality, Division of Water Quality, Salt Lake City, UT.
Stephan, C.E., Mount, D.I., Hansen, D.J., Gentile, J.H., Chapman, G.H., and Brungs, W.A. 1985.
Guidelines for Deriving Numerical National Water Quality Criteria for the Protection of
Aquatic Organisms and Their Uses. Publication PB85-227049. National Technical
Information Service, Springfield, VA.
USEPA (US Environmental Protection Agency). 1985. Technical Support Document for Water
Quality-Based Toxics Control. Publication PB86-150067. National Technical Information
Service, Springfield, VA.
USEPA (US Environmental Protection Agency). 1987. Ambient Water Quality Criteria for
Selenium–1987. EPA-440/5-87-006. USEPA, Office of Water Regulations and Standards,
Washington, DC.
USEPA (US Environmental Protection Agency). 1998. Report on the Peer Consultation
Workshop on Selenium Aquatic Toxicity and Bioaccumulation. Publication EPA-822-R-
98-007. USEPA, Washington, DC.
USFWS (US Fish and Wildlife Service). 2003. UPDES Permit UT0025551, for Jordan Valley
Water Conservancy District. Letter Dated October 23 from USFWS Utah Field Office to
Utah Department of Environmental Quality, Division of Water Quality, Salt Lake City, UT.
10