Total Maximum Daily Loads, or TMDLs, serve a profoundly important function under the federal Clean Water Act. They provide regulators with a systematic and comprehensive mechanism for identifying all sources and causes of water quality impairment, and then calculating the reductions needed to address the impairment in an equitable manner. But for TMDLs to be effective, they must be derived in a legally and technically defensible manner.

Under section 303(d)(1)(C) of the Clean Water Act, TMDLs must be established "at a level necessary to implement the applicable water quality standards with seasonal variations and a margin of safety which takes into account any lack of knowledge concerning the relationship between effluent limitations and water quality." TMDLs are typically expressed as the sum of wasteload allocations assigned to point sources (e.g., regulated industrial and municipal facilities), load allocations assigned to nonpoint sources and natural background, and a margin of safety to account for uncertainty.

TMDLs are more than simply an equation. Although the U.S. Environmental Protection Agency (EPA) commonly refers to TMDLs as "planning" or "informational" tools, they have important regulatory consequences. Once a TMDL has been established, National Pollutant Discharge Elimination System (NPDES) permits for existing municipal, industrial, and construction point sources must contain limits that are consistent with the assumptions and requirements of any available wasteload allocation in the TMDL. 40 C.F.R. § 122.44(d)(1)(vii)(B). And NPDES permits for new sources are prohibited unless: (1) there are sufficient remaining pollutant load allocations to allow for the discharge, and (2) the existing dischargers are subject to compliance schedules designed to bring the receiving water into compliance with applicable water quality standards. 40 C.F.R. § 122.4(i). EPA has determined that "all pollutants, under the proper technical conditions, are suitable for the calculation of TMDLs." Mercury, however, presents special challenges.

First, the "applicable" water quality standards for mercury are in a state of flux. For TMDL purposes, the applicable standards are those adopted by states and approved by EPA under Clean Water Act section 303(c). Most of the applicable standards are based on water column values (i.e., waterbody concentrations). EPA issued recommended standards based on the amount of mercury in fish tissue as opposed to the water column in 2001. Many states have not formally adopted EPA’s recommendations, yet in a number of recent proceedings states have opted to use ad hoc values derived using EPA’s fish tissue approach (often considerably more stringent that the water columnbased standards on the books).

Second, in many waterbodies, an important contributor of observed mercury is atmospheric deposition. However, neither the TMDL program in particular, nor the Clean Water Act in general, provides any direct authority over atmospheric deposition sources.

Third, the science is imperfect and evolving, as explained in greater detail below. While it is important to make progress in the face of uncertainty, sometimes the level of uncertainty is too overwhelming to permit good decisions about how to proceed.

Over the past several years, a number of states have attempted to develop mercury TMDLs. Most of these proceedings have been driven by litigation-based deadlines for states to either establish their own TMDLs or cede that authority to EPA. Beginning with Minnesota in 2007, states moved from waterbody-specific TMDLs to more ambitious state-wide or even multi-state TMDLs, in effect compounding the uncertainties and complications associated with mercury. The two most recent states to enter the fray are North Carolina and Florida, each of which initiated a state-wide mercury TMDL proceeding in the spring of 2012. The issues raised in these proceedings are both cautionary and illustrative of the challenges inherent in developing technically sound and legally defensible TMDLs.

Mercury and the mercury inventory

Mercury is a naturally occurring element with a very complex biological/geological/ chemical cycle. The complexities lie not only in determining sources of mercury but also in the interactions of mercury in air, water, soil, and biota. In spite of these complexities, EPA has approved mercury TMDLs based on assumptions of direct and linear relationships between mercury transport through air, water, and soil and the ultimate bioaccumulation in higher trophic level fish (e.g., largemouth bass). This regulatory approach is overly simplistic given the complexity of a chemical such as mercury.

The first complexity in developing defensible mercury TMDLs is a reliable mercury loading inventory for the TMDL area. States have determined with few exceptions that 95–100 percent of the mercury loading to a waterbody is from the deposition of atmospheric mercury released from anthropogenic and natural sources and the re-emission of previously deposited mercury, not from point sources discharging wastewater directly into the waterbody. To account for the remaining 0–5 percent of the loadings, states typically develop an inventory of NPDES-permitted wastewater discharges. Because a number of studies have shown that deposited mercury is predominately from global sources, mercury inventories should also be global.

While research varies on the numbers, mercury emitted from natural sources is estimated at 2000 metric tons per year (Mg/yr). Beginning with the industrial revolution, anthropogenic mercury emissions increased dramatically, although they have recently declined because of modern environmental controls. In addition to natural and anthropogenic mercury emissions, fractions of the natural and anthropogenic mercury that have deposited on land and in water are re-emitted into the atmosphere. Global mercury emissions are estimated at 6000 Mg/yr, approximately equally divided between natural, anthropogenic, and re-emission sources. The United Nations Environment Programme (UNEP) has estimate mercury emissions in the United States at 670 Mg/yr; 19 percent anthropogenic, 48 percent natural, and 33 percent re-emissions.

Biogeochemical cycling of mercury

Assuming one can successfully compile a mercury inventory, the ultimate determination of the TMDL is dependent on how mercury cycles through the environment and the complexity of this analysis remains a fundamental impediment to accurately developing mercury TMDLs. In developing their TMDLs, the States of Florida and North Carolina both undertook extensive scientific efforts to determine the relationships among mercury emissions, deposition, dissolution into waterbodies, and, ultimately, fish tissue concentrations. 

undertaken by EPA and UNEP demonstrates that of the mercury deposited in the United States, a minor fraction is from U.S. anthropogenic sources and the remaining is from natural, re-emissions, and other global anthropogenic sources. For example, UNEP modeling shows that a 20 percent reduction in U.S. anthropogenic emissions will result in only a 3.5 percent reduction in mercury deposition in the United States. Modeling by the State of Florida and the State of North Carolina similarly indicates mercury is predominately from sources outside of those states. All modeling shows a high degree of uncertainty and limited correlation to "in the-field" monitoring data. Mercury deposition is not a simple function of mercury emissions. Once mercury is deposited, field research and aquatic modeling again show, with very few exceptions, no significant correlation to waterbody concentrations or to fish tissue concentrations.

In spite of extensive scientific efforts to develop mercury inventories and scientifically based TMDLs, states have defaulted to a very simple presumptive concept that an X percent mercury emissions reduction will result in an X percent deposition reduction, which then results in an X percent reduction in waterbody concentrations and an X percent reduction in fish tissue concentrations. The reason for this simplified linear assumption is that none of the scientific studies yet supports a more sophisticated or scientifically grounded approach. The fact that EPA continues to approve TMDLs based on this simplified assumption presumably derives from a 2001 EPA report that applied these direct assumed relationships through long-term (100-year), steady-state modeling to demonstrate the direct relation between the deposition of ionic mercury and mercury in fish tissue. This same long-term equilibrium (multiple decades to centuries) is confirmed by other modeling studies.

In cases where a TMDL requires a reduction in mercury loading, some states correctly recognize that only anthropogenic mercury emissions can be controlled. Using the TMDL for Northeast Minnesota as an example, the State of Minnesota determined that a 65 percent mercury load reduction would be required to attain applicable standards, and assumed that anthropogenic emissions contributed 70 percent of the mercury loading. Based on these two factors, the state calculated that anthropogenic mercury had to be reduced by 93 percent. As unrealistic as this may be, and setting aside the fact that the Clean Water Act confers no authority to regulate air emissions sources, this same basic approach has been used in several other mercury TMDLs approved by EPA.

In addition to understanding the overall mercury inventory, regulators must also consider the ratio between anthropogenic and total mercury emissions. Although states like Minnesota have conveniently assumed that anthropogenic mercury constituted 70 percent of total mercury emissions, the available studies suggest that anthropogenic mercury emissions in the United States constitute only 19 percent of total mercury emissions. If applied to the Northeast Minnesota TMDL, this means that anthropogenic mercury would need to be reduced 100 percent, and natural and re-emission sources of mercury would have to be reduced 57 percent. In short, a real-world impossibility.

Legal considerations

Over the past two decades, EPA regions and states have established tens of thousands of TMDLs for a range of different pollutants and waterbodies, from small headwater creeks in Appalachia to the 64,000 square mile Chesapeake Bay watershed, from arroyos in the arid west to the abundant bays and embayments of the Pacific coast, and virtually every kind of lake, river, and stream in between. Even with all of the knowledge and experiences gained through those efforts, much remains to be learned and done. No more so than in the mercury context, where special challenges complicate the development of technically sound and legally defensible TMDLs.

Some states have recently established mercury TMDLs based on

ad hoc values rather than formally adopted "applicable" standards, and states have also opted to rely on simplified assumptions about mercury cycling in the environment even though these assumptions cannot yet be verified or validated. Last but not least, states have elected to proceed with TMDLs where the predominant source of mercury (e.g., 95–100 percent is atmospheric deposition, leading to TMDL equations that (1) "assume" reductions that are beyond the authority of the Clean Water Act or governments (U.S. or international) and (2) place point sources in jeopardy of extremely stringent if not unachievable NPDES permit limits, even though point source contributions are de minimis and will have no practical impact on the TMDL outcome.

Some practitioners derisively refer to TMDLs as "

too many damn lawyers." Unlike the wave of lawsuits that breathed life and energy into the Clean Water Act program 20 years ago, there have been remarkably few cases addressing the scope, contents, and effect of TMDLs. Indeed, most of the fundamentals of mercury TMDLs have been established without scientific basis or even confirmation through field studies. So far, practice has begotten precedent, but the question remains: who’s fooling whom?