Mercury

By Lisa Heinzerling, Catherine A. O'Neill, and Rena I. Steinzor

Background

The Issue Whether the United States should impose more stringent pollution controls on the two largest remaining sources of mercury contamination in the human food chain: mercury cell chlor-alkali plants and coal-fired power plants.

Mercury is highly toxic to humans. Exposure to even small amounts of methylmercury, the most toxic form of this elemental metal, can lead to irreversible neurological damage. Methylmercury's neurodevelopmental effects place the developing fetus, children, and adults up to age 20 at particular risk. The most recent data also suggest adverse effects on the cardiovascular systems of adults. Humans are exposed to methylmercury primarily through the consumption of fish that has been contaminated by the deposition of mercury air emissions into surface waters (for example, rivers, lakes, and streams).

Currently, one in six or, by some recent estimates, as many as one in five women of childbearing age in the United States has a blood mercury level that would put a developing fetus at risk. This figure doubles for Native American and Asian American women.

Methylmercury contamination is severe and widespread. Vast expanses of the waters that support fish on which humans rely for food are under fish consumption advisories due to methylmercury contamination. By the most recent tally, forty-five states and several tribes have issued advisories placing some or all of their waters off limits for those who would eat the fish they catch. Mercury advisories blanket the lakes and rivers of the Northeast, the Ohio Valley and the upper Great Lakes. As of 2003, 21 states and one tribe have issued advisories covering the entirety of their lakes and/or rivers. In addition, 100 percent of Lakes Superior, Michigan, Huron, and Erie are under mercury advisories.

Mercury occurs naturally in the environment and is also produced by anthropogenic (man-made) industrial processes. The steady increase in activities such as the combustion of fossil fuels, and the manufacture of chlorine using mercury cells has elevated the volume of mercury released into the environment. Mercury air emissions convert to methylmercury in rivers, lakes, and streams, becoming part of a complex cycle with local, regional, and global components.

What People Are Fighting About

Sources of Pollution

Anthropogenic emissions in the United States are dominated by emissions from mercury cell chlor-alkali facilities and coal-fired utilities. Nine outmoded mercury cell chlor-alkali plants "lose" as much as 65 tons per year of mercury in "fugitive" emissions. EPA also estimates that, as of 1999, coal-fired utilities emitted 48 tons of mercury per year, although this estimate may well under-count such emissions. Other large source categories are industrial and commercial boilers, which contribute, respectively, 9.7 and 6.5 tons per year. Although we have only recently begun to gather emissions data for gold mining and refining, these industrial segments produce the next largest amounts of mercury. A handful of other source categories rounds out the roster. A dec

ade ago, the share of total U.S. mercury emissions contributed by coal-fired utilities was rivaled by emissions from two other source categories: medical waste incinerators and municipal waste combustors. However, sources in each of these other categories have reduced their emissions on the order of 90 percent, as a result of regulations issued by EPA under the federal Clean Air Act (CAA). As of 1999, medical waste incinerators had reduced their mercury emissions to just 2.8 tons per year and municipal waste combustors had reduced their mercury emissions to just 5.1 tons per year. At that time, coal-fired utilities remained unregulated and mercury cell chlor-alkali plants were not subject to meaningful mercury controls.

EPA Acts

In December, 2003, EPA announced its long-awaited proposal for regulating mercury emissions from coal-fired utilities. In March, 2005, EPA issued the final version, the Clean Air Mercury Rule (CAMR). Before the 2003 proposal was issued, EPA had been widely expected to require coal-fired utilities to achieve a 90 percent reduction in mercury emissions - from approximately 48 tons to five tons - by 2007. Instead, EPA promulgated a rule that would permit coal-fired utilities to continue to emit more mercury for a longer time - well into the next decade. This startling departure from more stringent regulation eschews the ordinary process for regulating air toxics under section 112 of the CAA, in favor of a cap-and-trade approach.

As for chlor-alkali plants, the Agency imposed a set of toothless requirements that ask plant personnel to visually inspect giant mercury cell rooms for fugitive emissions, a pointless exercise because mercury vapor is generally invisible to the naked eye.

Downsides of Mercury Trading

EPA's attempt to apply the tool of cap-and-trade to the problem of mercury has met with a raft of criticism. The clamor over the rule is not surprising given the high stakes for public health. EPA's interpretation of its authority under the CAA is also unprecedented, and is contradicted by its interpretations over the previous decade. Finally, investigations by the EPA Inspector General and the news media have revealed procedural irregularities that cast the Agency's political leadership in a negative light, including summary orders to career staff to cut off the analysis of the costs and benefits of more stringent, facility-specific pollution controls.

Critical attention is warranted as well because cap-and-trade, like other regulatory tools, has strengths and weaknesses that make it well suited for some environmental problems but ill-suited for others. Because mercury is so toxic, and is also very persistent in the environment -- that is, it does not break down into less dangerous components but instead stays around for a very long time -- the unrestricted buying of allowances to pollute could easily lead to "hot spots" of pooled contamination that will threaten specific communities. (See CPR's Perspective on Emissions Trading for further analysis of these problems).

Another weakness of EPA's cap and trade rule is the actual level of control selected. Whether one employs a cap-and-trade approach or a technology-forcing standard, the key to how much protection is provided is the level of control or emissions set by the agency. EPA has set overall caps (or limits) to be imposed on total emissions at 38 tons per year (TPY) by 2010 and 15 TPY by 2018. The Agency has allocated "allowances" authorizing sources to emit up to the aggregate level set by the cap. Sources are then authorized to trade allowances among themselves - buying and selling in order to ensure that each source holds allowances sufficient to cover its relatively larger or smaller quantities of mercury emissions.

A close look at EPA's approach reveals that the caps EPA has set are extraordinarily weak. The first-phase cap, to be met by 2010, will require no reductions beyond those already realized as "co-benefits" of other regulatory controls sources must already comply with. And the second-phase cap, effective in 2018, allows sources to "bank" (or save) allowances, enabling aggregate emissions to exceed the cap if sources decide to hoard allowances. EPA models project a mere 50 percent reduction in emissions by 2020. Based on past agency practice, many observers had expected EPA to require a 90 percent reduction by 2007. These very modest - and delayed - reductions in a toxic metal known for persistence in the environment translate into an unjustified windfall for coal-fired utilities, which have recently announced plans to build dozens of new plants across the country as an alternative to continued dependence on "foreign oil." Given methylmercury's neurodevelopmental effects, this reprieve and the dramatic expansion of coal-fired plant construction threatens future generations in a manner that directly violates the CAA goals.

Environmental Justice: Hot Spots and Risk Avoidance

Of particular moment, EPA's approach raises a host of environmental justice issues. As framed, the cap-and trade proposal would disproportionately burden various fishing tribes and indigenous peoples, other communities of color, and low-income communities that depend on fish. As amply demonstrated by the National Environmental Justice Advisory Council, the fish consumption practices of these groups differ considerably from those of the "typical U.S. consumer" that is the apparent focus under EPA's approach. Members of these groups consume more fish, at greater frequencies, contaminated at higher levels - with the result that they are more highly exposed than members of the general population. Members of these groups are thus the ones relatively likely to suffer the adverse health effects of increased mercury emissions over a longer period of time. For some groups, the harms resulting from the delayed and limited emissions reductions may also be felt along interrelated economic, social, cultural, spiritual, and political dimensions. This is the case, for example, for the various Ojibwe tribes and other fishing tribes of the upper Great Lakes.

As mentioned earlier, EPA's approach introduces the problem of "hot spots" - local or regional instances of relatively concentrated emissions and, ultimately, relatively high exposure. EPA's cap-and-trade approach expresses the relevant caps in terms of total mercury emissions and allows mercury to be traded freely among sources anywhere in the United States. By design, it says nothing about how the emissions are distributed within this nationwide boundary. The potential for perpetuating or exacerbating hot spots has long been recognized as the Achilles heel of cap-and-trade approaches. The likelihood that such hot spots will coincide with areas that are home to tribes and indigenous peoples, communities of color and other low-income communities has been more recently acknowledged.

EPA claims in the preamble to the proposed rule that it "does not expect any local or regional hot spots" under this national cap-and-trade approach. EPA reaffirms this claim in the final rule, but without addressing any of the substantive arguments raised by commenters. Indeed the potential for hot spots was among the most controversial and commented upon issues, with even the Government Accountability Office and EPA's Inspector General questioning EPA's claim on this point. Moreover, EPA's own models suggest reason for concern. For example, available data point to the possibility of local and regional hot spots in the upper Great Lakes states of Michigan, Minnesota, and Wisconsin. These hot spots may affect even members of the general population in this region, given their higher than average rates of fish consumption. And they would impose a profound burden on the fishing tribes, whose practices place them among the most highly exposed.

Additionally, EPA's rule relies heavily on risk avoidance - that is, it asks those who will bear the risks of unabated mercury contamination to forego eating fish in order to protect themselves. EPA concedes in the preamble to the proposed rule that those who regularly consume fish - who it unflinchingly acknowledges will be "Native Americans, Southeast Asian Americans and lower income subsistence fishers" -- particularly children up to age 20 and women of childbearing age - are at greater risk than the "typical U.S. consumer." "In response," EPA suggests that they take on the responsibility to avoid these risks. It refers them to advisories that ask them to eat less fish or different species or to stop consuming fish altogether - despite the widely recognized health benefits of eating fish. This choice of risk avoidance - in lieu of reducing the risks by preventing mercury contamination in the first place - is likely to burden disproportionately fishing tribes and other higher consuming subpopulations. It is to these groups that EPA addresses this "response." And while giving up fish for some individuals may mean finding less palatable substitutes in terms of protein and other nutrients or expending more time and money to travel to less contaminated waters to fish, giving up fish for other individuals may be impossible.

For example, members of the various Ojibwe and other fishing tribes of the upper Great Lakes value fish and fishing for physiological, economic, social, political, cultural and spiritual reasons. The burden imposed by contaminated fish is thus not only different in degree, but also different in kind, from the burden imposed on the general population.

Distorted "Costs" and "Benefits"

EPA attempts to justify its final mercury rule by arguing that the costs of obtaining mercury reductions from coal-fired utilities outweigh the benefits of doing so. EPA comes to this conclusion by narrowly defining the benefits of mercury emissions reductions, by invoking assumptions about exposure that systematically undercount those affected, and by discounting all benefits realized in the future. While these infirmities are typical of cost-benefit analysis in other contexts, a few analytical distortions warrant particular note. (For more information on this troubled process, see CPR's Perspective on Cost-Benefit Analysis.)

First, EPA makes several crucial cuts in its benefit calculus: it counts only benefits to human health (and so excludes ecological health and other benefits, including, e.g., economic, social, political, cultural and spiritual well-being for the fishing tribes), and of these human health benefits, it counts only neurodevelopment effects and so excludes cardiovascular and other effects. It further winnows the benefits, by counting only neurodevelopment effects that are captured by IQ decrements. Various independent analyses have demonstrated that quantifying these excluded benefits would reverse the conclusion of the analysis, with the benefits of regulation dwarfing the costs.

Second, EPA narrowly circumscribes the exposed population, counting only prenatally exposed individuals whose mothers eat freshwater fish caught by recreational anglers on inland U.S. lakes. Among the exclusions are all those exposed during childhood, all those exposed via ingestion of freshwater fish caught commercially on inland U.S. lakes, and all those exposed via ingestion of non-freshwater fish caught recreationally or commercially in coastal or other waters

Third, EPA makes an extraordinary judgment regarding social good: it counts losses in future earnings as a cost of mercury contamination, but it counts as a benefit of mercury contamination the money the government saves because children with lower IQs (as a result of mercury exposure) need fewer years of school. Under this startling approach, a weak rule is preferable, because it preserves the "benefits" of lower IQs due to mercury contamination. Then, using a discount rate of three percent, EPA determines that each IQ point decrement avoided will net a mere $8,807 in benefits (a figure that is reduced to $1,580, it points out, if a seven percent discount rate is used).

Legal Challenges

Several states, tribes, and environmental groups have filed lawsuits challenging the final rule. These lawsuits are based on the existing language of the Clean Air Act, which EPA ignored in adopting a cap-and-trade approach. As a hedge against an adverse judicial decision, conservatives in Congress have introduced legislative versions of the rule's cap-and-trade approach, which would revise the statute to reflect the policies pursued by the Agency. These provisions are part of President Bush's "Clear Skies" initiative.

Chlor-alkali Plants

Chlorine manufacturing also contributes significantly to mercury emissions in the United States. As explained above, mercury cell chlor-alkali plants "lose" as much as 65 tons of mercury each year during the manufacturing process. This missing mercury is not included in estimates of total U.S. mercury emissions. The industry claims that the missing mercury condenses and accumulates in pipes, tanks, and other plant equipment; environmentalists contend that the missing mercury escapes as fugitive emissions. EPA concedes that "the fate of all the mercury consumed at mercury cell chlor-alkali plants remains something of an enigma."

EPA declined to address this source of mercury when it issued a "maximum achievable control technology" (MACT) standard for chlorine manufacturers under section 112 of the Clean Air Act in December 2003. Instead, it issued a rule that bestows special treatment on the nine remaining mercury cell chlor-alkali plants within this source category, fails to address the missing mercury other than by "work practices" requirements that are likely impossible to enforce, and permits these antiquated facilities to continue to operate indefinitely with only the tiniest of reductions - approximately three tons, by EPA's estimate - in mercury emissions. Meanwhile, EPA knows that these nine facilities buy, store, and use between 38 and 219 tons per year of pure mercury, based on 2003 and 2004 data. Yet the agency makes no real attempt to address what basic math reveals to be a glaring shortcoming in its regulatory effort: the 65 "lost" tons. Beyond visual inspections of invisible emissions, EPA has asked the nine chlor-alkali plants to report the quantities of virgin mercury added to the mercury cells in the five years preceding the rule's compliance date of December 19, 2006. But it makes no promises to do anything with this information.

As unfortunate, EPA knows that the use of mercury in the manufacture of chlorine and caustic is entirely unnecessary in the first place. Unlike coal-fired utilities, which emit mercury as an unintended but inevitable byproduct of burning coal, mercury cell chlor-alkali facilities employ an outmoded process that intentionally adds large amounts of mercury to an electrolytic cell. Thirty-two of the approximately 43 facilities that manufacture chlorine and caustic have eschewed this process in favor of diaphragm or membrane cells, which are significantly more energy efficient and do not involve mercury. Indeed, the process used by the nine remaining mercury cell chlor-alkali plants is so outdated that a new facility of this sort has not been built in 30 years. Yet EPA failed to implement a MACT requirement to provide incentives for these nine plants to modernize, preferring instead to prop them up for the foreseeable future by treating them as a separate "subcategory" with more lenient requirements than the other sources in the chlorine manufacturing category. EPA's chlorine manufacturing rule is the subject of a pending lawsuit.

Gold Mines

Gold mines, which emit mercury to the air primarily in the process of roasting the ore, collectively account for a modest fraction of mercury emissions. However, as in the case of mercury cell chlor-alkali plants, emissions from individual facilities may dwarf those from individual coal-fired power plants. The average coal-fired power plant emits approximately 250 pounds of mercury per year. Yet a single mine in Nevada emitted 9,400 pounds of mercury in 1998; at another, mercury emissions were 2,200 pounds. The magnitude of these contributions has only relatively recently come to the fore, since the Toxic Release Inventory (TRI) rules required gold mines to begin reporting their discharges of mercury in 1998.

EPA's response to these TRI data is still taking shape. Among the options is a MACT standard under section 112 of the CAA, which, as noted above, is the standard approach to regulating emissions of hazardous air pollutants, including mercury. A different option has been employed by EPA's Region Nine and the Nevada Department of Environmental Protection for the four largest emitters in Nevada: a "voluntary mercury air emissions reduction program" or "VMRP." According to the TRI data, these sources together emitted 13,560 pounds of mercury, accounting for roughly 90 percent of mercury emissions from Nevada gold mines. Under the VMRP, these four sources undertook process changes which, according to EPA, resulted in a 40 percent reduction in emissions from the baseline by the end of 2002 and a 75 percent reduction by the end of 2003. Such figures are indeed impressive, and the relatively short order in which the reductions were achieved a cause for hope in efforts to protect against the harms of mercury contamination. In addition, EPA cites the cost-effectiveness of such a flexible approach.

However, there is reason to be cautious. First, the quality of the emissions figures on which these claims are based is unclear. Although the 1998 TRI data show total emissions from the four sources to be 13,560 pounds per year, in calculating the touted VMRP reductions, EPA uses a baseline figure of 21,098 pounds per year, based on revisions to the 1998 data. Several knowledgeable observers have questioned this baseline and, thus, the claims that flow from it. Second, voluntary emissions reductions are just that: voluntary. Without a MACT standard in place, these sources could resume their old, dirtier practices at any time it becomes expedient for them to do so, and other facilities in the source category may simply opt not to participate.

The decision not to promulgate MACT standards means more limited protection for public health. The claimed 75 percent reduction in emissions is more modest than the 90 percent reduction typically required by MACT. Moreover, the process EPA employed fails to incorporate the mandatory residual health-based standard as would a MACT approach. Health and environmental agencies, environmental groups, and others in Idaho and Utah - downwind of the Nevada gold mines - have expressed many of these concerns. Meanwhile, Utah has just issued the first state waterfowl consumption advisory, warning against the consumption of two species of wild duck found to be highly contaminated with mercury.

CPR's Perspective

In addressing mercury emissions from coal-fired power plants and chlor-alkali plants, EPA faced a legally straightforward task. In the case of power plants, section 112 of the Clean Air Act lists mercury as a hazardous air pollutant subject to regulation, and EPA had previously concluded that power plants should be regulated as sources of mercury emissions. Under the clear language of section 112, listing of mercury and of this source category led to an obligation on the part of EPA to regulate mercury emissions by prescribing MACT and requiring its installation. A MACT directive would have required each individual source to reduce its emissions to the level achieved by the best performers in its source category. By many accounts, this would have meant reducing mercury emissions by as much as 90 percent, within the three-year deadline provided by section 112.

EPA's mercury rule cannot be fixed with tinkering. EPA should withdraw the rule and replace it with the legally correct and environmentally preferable MACT regime for mercury. EPA's chlor-alkali rule is also utterly inadequate. EPA cannot simply cite the "enigmatic" nature of the mercury that is unaccounted for in the mercury cell process; EPA must find and reduce, through enforceable MACT requirements, emissions of any "lost" mercury.

In addressing mercury emissions from gold mines, EPA needs to weigh carefully the pros of apparently rapid emissions reductions against the cons of relying on wholly unenforceable measures and working outside the comprehensive MACT-based approach to regulating hazardous air pollutants that Congress devised in section 112 of the CAA. Indeed, given the current extraordinary price for gold, there is cause for concern that sources' voluntary undertakings may give way to expediency in processing the ore in order rapidly to bring it to market. In any event, EPA needs to ensure that the emissions data on which it relies are accurate.

Based on the information to date, CPR concludes that, on balance, a MACT-based approach to gold mines' mercury emissions appears more likely to guarantee that human and environmental health are adequately protected. This conclusion must be tentative at this point, given that information regarding this source category is still forthcoming. Of course, one of the drawbacks of this approach is the time it takes to produce a MACT standard; thus, EPA should act promptly.

Finally, as a general matter, EPA should address the harms of mercury contamination by risk reduction, rather than risk avoidance. That is, EPA needs to require sources to prevent or reduce their mercury releases, rather than ask those exposed to protect themselves by taking steps to avoid exposure - here, by reducing or eliminating fish from their diets. Given what is at stake from mercury contamination and given the serious losses that for many accompany having to avoid consuming fish, risk avoidance is an inappropriate regulatory tool in this context.

For Further Information

Lisa Heinzerling and Rena I. Steinzor, A Perfect Storm: Mercury and the Bush Administration, 34 ENVT'L. L. REP. 10297 (2004)

Lisa Heinzerling and Rena I. Steinzor, A Perfect Storm: Mercury and the Bush Administration, Part II, 34 ENVT'L. L. REP. 10485 (2004)

Catherine A. O'Neill, Mercury, Risk and Justice, 34 ENVT'L. L. REP. 11070 (2004)