TMDLs, Are We There Yet?: The Long Road Toward Water Quality-Based Regulation Under the Clean Water Act

Oliver A. Houck

Editors' Summary: Water quality standards-based regulation has been the "reserve clause" of the Clean Water Act (CWA), intended to clean up waters that remain polluted after the application of technology standards. For 20 years, these provisions lay idle, prodded forward at least by litigation in the early 1990s. Today, they are at the center of nearly two dozen lawsuits, a Federal Advisory Committee Act committee, and a flurry of regulatory guidance. Their implementation presents serious issues of federalism, science, and political will. In a prior Article in ELR—The Environmental Law Reporter, the author discussed the origin of CWA § 303(d), the Act's provision on water quality standards. In this Article, he describes the implementation of § 303(d).

The author is a Professor of Law at Tulane Law School. The research assistance of Brent Walton, Tulane Law School '97, Deborah Clarke, '98, and Shannon Skinner, '98, is acknowledged with gratitude.

[27 ELR 10391]

The central tension of the Clean Water Act (CWA)¹ has always been between state programs based on local water conditions and a federal program based on national standards. The showdown came in 1972 when the Congress opted for nationwide technology standards, over the vigorous protests of states and regulated industry that state water quality standards were up to the job. The states and industry were successful, however, in retaining a water quality-based strategy for upgrading waters that remained polluted after the application of technology standards—§ 303(d).² Regulation of water discharges geared to their impacts on receiving water quality was America's original approach to pollution control, and it has been at issue in nearly every major legislative and administrative review of the Act since 1965.³ Throughout, and to this day, water quality-based regulation has been the avowed process of choice for the majority of states and their associated trade groups. When it came to execution, however, the implementation of this process under § 303(d) languished. Indeed, it was virtually nonexistent until a recent series of citizen lawsuits dragged it back on stage.

An earlier Article in ELR—The Environmental Law Reporter described § 303(d), its evolution from early state water quality standards programs, and the continuing struggle over water quality-based strategies from 1972 to the [27 ELR 10392] present.⁴ This Article describes the implementation of § 303(d) by the U.S. Environmental Protection Agency (EPA) and the states, the recent explosion of litigation surrounding it, and the current state of the program. The picture that will emerge is that of a federal agency moving slowly, pressed by other priorities and shielded by what it considered to be unreviewable discretion; states moving even more slowly, confident that they were not in the line of fire and that this was not a priority for EPA or anyone else, the action nudged forward by infrequent and inconclusive lawsuits, then catapulted forward by a recent wave of suits imposing significant compliance requirements on short deadlines; and a frantic scramble by federal and state agencies both to evade and comply.

As stated earlier, this picture is not yet complete. EPA, states, industry, agriculture, forestry, municipalities, environmentalists, and members of Congress are now in full discussion over the future of the § 303(d) program. The earlier Article described how we got to § 303(d). This Article describes its implementation. A concluding Article will reflect on the consequences of these developments, their possible outcomes, and their significance to environmental law.

Section 303(d): The Avoidance Years

Following the passage of the Federal Water Pollution Control Act Amendments of 1972,⁵ EPA was fully occupied, indeed overwhelmed, in promulgating technology standards for point sources under the Clean Water Act and defending them in court.⁶ The Agency had little inclination, and indeed saw little reason, to implement the "safety net" features of § 303(d) before the technology requirements were in place. After all, water quality upgrading was only required when polluted waters could not be brought up to standard through best available technology requirements. And these requirements were many years away.

EPA, further, had its hands full trying to establish a floor for state water quality standards that would at least preserve the potential for § 303(d) upgrading in the future. The Agency struggled with resisting states over policies requiring that existing water quality be protected,⁷ that achievable water quality be met as well,⁸ that standards be set for all state waters,⁹ that existing water uses be maintained,¹⁰ that uses such as "waste transport" be rejected,¹¹ that the standards of downstream states be respected,¹² and that state criteria for the uses chosen be based on at least arguably sound science.¹³ What even these rather threshold efforts for a national program based on state water quality standards showed is that, although the states were to retain "primary responsibilities" for water pollution control under the Act,¹⁴ EPA was going to have to play a major role in keeping it honest. If states could cut a corner, many would. As for moving forward to identify polluted water bodies and establishing load limits on their own initiative, there is no evidence in the decade following the 1972 Amendments that the states were going to take this bull by the horns.

EPA took § 303(d) by the horns very gently. The section's obligations were to be triggered by EPA's formal identification of pollutants appropriate for water quality analysis and total maximum daily loads (TMDLs).¹⁵ Once these pollutants were identified, the drill began: states had 180 days to submit their lists of water quality limited segments (WQLSs), priorities for cleanup, and TMDLs.¹⁶ In October 1973, rather promptly considering its many duties under the new Act, EPA published a proposed notice of a two-volume set of pollutants appropriate for the § 303(d) process.¹⁷ Then, nothing happened. The identification languished. Lawsuits in the late 1970s attempted to challenge the absence of TMDLs on the Colorado River¹⁸ and on waters in South Dakota¹⁹ and failed for want of the predicate: EPA hadn't started the clock.

Instead, EPA opted to fold the TMDL process into its, in retrospect, overly ambitious regulations²⁰ for basin planning under §§ 106, 208 and 303(e).²¹ Basin plans encompassing all industrial, municipal, and nonpoint source controls would also establish, inter alia, TMDLs and discharge load allocations.²² "Substantial failure" of any plan to conform to § 303(e) (requiring a state "continuing planning process" that included TMDLs) might "indicate" that the process was deficient, leading to disapproval of the process and ineligibility for state delegated programs.²³ These alarming-sounding sanctions notwithstanding, the basin [27 ELR 10393] planning menu attempted too much in the same sitting and would fall of its own weight in the years ahead.

EPA's occupation with basin planning was pushed forward by a court order in 1975, requiring a more comprehensive and accelerated approach to the planning process.²⁴ The Agency's subsequent regulations provided a three-year schedule for the approval of all state plans, and more guidance on the implementation of § 303(d).²⁵ For each water quality segment, TMDLs would include "a total allocation for point sources" and a "gross allotment for nonpoint sources," the combined total of which would not exceed the TMDL.²⁶ Each load allocation would incorporate an "allowance for anticipated economic and population growth over at least a five-year period," and an "additional allowance reflecting the precision and validity of the method" used in calculating these loadings.²⁷

Without EPA's identification of pollutants that would trigger TMDL planning, however, these requirements remained rhetorical. In 1978, with EPA still struggling to make basin planning work, the Agency was brought up short by another court order requiring it to publish a final identification of TMDL pollutants.²⁸ Regrettably, from EPA's perspective, the § 303(d) process would now be set in motion. In the notice of its required regulations, the Agency explained that it did not consider such finalization of the pollutant list "a matter of high priority" because many of the "practical results" of TMDLs were already being accomplished through basin planning;²⁹ right or wrong, this perspective persists to the present day. There may have been another reason for EPA's reluctance as well. The EPA Assistant Administrator for Water had been a close aide to Sen. Edmund Muskie (D-Me.) during his years of combat against water quality-based regulation and for the adoption of new technology standards;³⁰ one gets a strong sense from EPA's new regulations that this was a game its water program administrators did not see as worth playing. They would "provide a phased approach" to establishing TMDLs "consistent with the intent of Congress" (whatever intent Congress may have had other than the explicit language of § 303(d) itself), in order to "ensure" that "current State water quality management programs will not be disrupted."³¹ In not disrupting state programs, they would certainly succeed.

EPA's regulations delayed, soft-pedaled, and understated the § 303(d) requirements to a remarkable degree. Seizing on the statutory language that the states' "first submissions" of polluted waters and TMDLs were due in 180 days, EPA asked states to identify "one or more" water quality limited stream segments, and one or more TMDLs, in those first six months.³² Priority rankings for listed state waters would await future state/EPA agreements.³³ Nonpoint source contributions would not need to be considered in setting these priorities since "the relative significance of point and nonpoint sources will in some cases not be determined until TMDLs are developed,"³⁴ whatever that might mean. Lest even these minimal obligations seem harsh, EPA further stipulated that the "results of past or ongoing efforts" could be submitted for approval.³⁵ In short, one TMDL submission would suffice. When the second one was due was anyone's guess, as was what would constitute a TMDL. The stage was set for inaction.

Inaction occurred. A few states submitted a few lists. Most states submitted nothing at all. The question became: what then? EPA's answer, based on a literal reading of the Act, was: nothing.³⁶ Section 303(d) required only that the Agency approve or disapprove state submissions. If the states submitted something unacceptably minimal, EPA could correct the problem with its own lists and TMDLs. But if the states submitted nothing at all, there was this unfortunate disconnect: there was nothing the Agency could do.

The Constructive Submission Theory: Avoidance Might Not Work

By the early 1980s, then, the next round of citizen suits under § 303(d) had to establish EPA's duty to act. It was not an easy road. In Scott v. City of Hammond,³⁷ an Illinois district court found that, although neither Indiana nor Illinois had submitted TMDLs for Lake Michigan, EPA had no power either to require the state to do so or to promulgate its own. At which point, § 303(d) was a thoroughly dead letter.

On appeal, however, the Seventh Circuit reversed,³⁸ reasoning that the "prolonged failure" of a state to submit anything could amount to a "constructive submission" to EPA of no TMDLs, triggering EPA's duty to act.³⁹ In so holding, the court relied on the apparent expectation of [27 ELR 10394] Congress that the § 303 process would actually take place to rebut EPA's insistence that Congress had nowhere given it explicit authority to intervene: "We think it unlikely that an important aspect of the federal scheme of water pollution control could be frustrated by the refusal of the states to act."⁴⁰ Under these circumstances, EPA's inaction was "tantamount to approval of state decisions that TMDL's are unneeded."⁴¹ Remanding for the district court to decide whether in this case the states had in fact "determined not to submit TMDLs," the court made clear that this decision was not to be based on a state's motive but, rather, on whether there were sound reasons justifying the states' failure, and "persuasive evidence" that they would be moving quickly.⁴² If not, EPA was going to be forced, over its obvious reluctance, off of the sidelines and into the game.

For the remainder of the 1980s, EPA treated Scott as an aberration and stayed with its (non)game plan. So long as the states did nothing, it could do nothing too. In 1985, the Agency consolidated its water quality management program regulations in the form that they appear today.⁴³ The regulations continued their emphasis on basin planning and presented the § 303(d) process as a related, but separate, drill⁴⁴—a drill, however, without a deadline. The states would submit water quality limited segments and TMDLs, per the statute, "from time to time"; the actual schedules for these submissions would be worked out between EPA regional administrators and the states.⁴⁵ If there is any lesson to be learned across the spectrum of environmental law, it is that drills without deadlines are not performed. Its regulations written, EPA walked back off the field. It would not be able to stay there for long.

To EPA's dismay, Scott did not go away. The constructive submission theory took hold. A follow-up suit in Oregon, Northwest Environmental Defense Center v. Thomas,⁴⁶ led to a consent decree in 1987 with a timetable for federal action if Oregon did not make its, by then several years overdue, submissions of impaired waters. Starting in 1991, a series of Alaska cases⁴⁷ took the next step of requiring EPA to promulgate TMDLs for state waters, noting that no TMDLs had been submitted, nor even "attempted," nor even "promised" to be attempted.⁴⁸ The district court found EPA's "reassurances" that it had discretionary authority to act "not particularly comforting" in light of the fact that EPA had "failed to take action on this matter for over ten years";⁴⁹ nor was the court persuaded by testimony on "EPA's other worthwhile water quality programs," and its plea for the ability to "respond to future environmental crisis by shifting available resources away from other tasks."⁵⁰ As for judicial deference to agency interpretations of law, the only interpretation EPA had shown with regard to the Clean Water Act's TMDL requirements "has been to ignore them."⁵¹

The message was coming home that state failure to act would bring in EPA, however reluctantly. These cases begged the question, however, of whether whatever a state might submit would suffice to forestall federal intervention. The next round of rulings seemed to indicate that virtually any state action got EPA off the hook. Following the remand in Scott, the states of Illinois, Indiana, and Michigan submitted rather surprising determinations that TMDLs were unnecessary for Lake Michigan; Wisconsin was a little more responsive, identifying four lake sectors for the development of wasteload allocations. EPA approved the submissions, and was then challenged in National Wildlife Federation v. Adamkus.⁵² The Federation claimed that these minimal submissions constituted, in effect, a submission of no TMDLs, triggering EPA's duty to promulgate the load allocations on its own. In 1991, the Illinois district court rejected the argument,⁵³ finding that since the states had submitted something and EPA had approved it, the statute was fulfilled.

In 1993, a district court in Minnesota made a similar ruling on EPA's approval of that state's progress on TMDLs in a case that previews what a grinding process full TMDL development nationwide is going to be. In 1987, Minnesota had submitted a list of five WQLSs, one of which carried a TMDL. In 1992, the state raised its list to seven river segments, and scheduled TMDL development for them over the following 10 years. Citizens brought suit in Sierra Club v. Browner,⁵⁴ noting that Minnesota's semiannual report to EPA listed 1,116 waters as not meeting state standards, an assessment that included less than one-twentieth of the state's water bodies;⁵⁵ clearly, in the Sierra Club's view, more WQLSs and TMDLs were called for. EPA responded by publishing its own list of 447 water quality limited segments for Minnesota, and went on to approve 43 state national pollutant discharge elimination system (NPDES) permits conditioned on wasteload allocations as TMDLs.⁵⁶ On these facts, the court could not find that there had been a constructive submission of no WQLSs or TMDLs.⁵⁷ The Sierra Club was forced to argue that the TMDL schedule was too slow, and that the permits, without nonpoint allocations, were not TMDLs at all, arguments this court was not ready to accept.⁵⁸ At trail's end for the constructive [27 ELR 10395] submission theory, EPA could be required to persuade the states to act or to act itself. If states did something, however, albeit minimal and on whatever schedule, they just might escape further review.

Slow Motion

In 1989, the U.S. General Accounting Office (GAO) issued a report on TMDLs entitled More EPA Action Needed to Improve the Quality of Heavily Polluted Waters.⁵⁹ It was not a pretty picture. EPA Region X, the EPA office examined in depth, had received and approved only one TMDL for the 602 listed WQLSs in that region.⁶⁰ Region II had approved 4 of 168.⁶¹ Alaska and other states had told the GAO with more candor than caution that they had no particular plans to set TMDLs, and regarded the process as a waste of time.⁶² As for EPA's efforts, the report's section headings included such indictments as "EPA Action on TMDLs Was Compelled by Lawsuits,"⁶³ "EPA Has No Way of Assessing TMDL Implementation Nationwide,"⁶⁴ "EPA's Current Systems Do Not Measure TMDL Compliance,"⁶⁵ and "TMDLs Not Integrated Into Reporting Requirements."⁶⁶ The report concluded that EPA had no program for TMDLs, no schedule for implementation, and no way even of knowing what was going on.⁶⁷

One would think that this level of criticism would prompt a national response. It did not. Region X, stung most directly by the GAO analysis and by the Oregon litigation which had now spread to the state of Washington as well, began to make noises advancing the cause of TMDLs. In a series of small articles, the Region's Chief of Water Planning met the issue with remarkable optimism and candor.⁶⁸ TMDLs were "one of the most powerful, but also one of the most under-utilized" tools of the Clean Water Act, he said.⁶⁹ They were also, to many, one of the "most frightening" because of their reach to nonpoint sources of pollution.⁷⁰ Actually, TMDLs were a "very simple and surprisingly logical problem solving process."⁷¹ Shortage of information on water quality impacts? Congress said that "ignorance is no excuse for inaction. Just add a margin of safety to compensate for the lack of knowledge and keep moving."⁷² Lack of controls over nonpoint sources? A TMDL was still useful to "force managers to define both the actual amount of pollution reduction needed and the actions necessary" to achieve it, ensuring "more effective NPS [nonpoint source] programs"; indeed, "there may actually be many more control options available than are first apparent."⁷³

Whatever Region X was saying, however, EPA headquarters had yet to catch the wave.⁷⁴ It was not until April 1991 that EPA began publishing guidelines for state implementation of § 303(d),⁷⁵ and October 1992 that it finally set a deadline for the submission of state WQLS lists.⁷⁶ EPA's biennial reports to Congress, meanwhile, continued to paint nationwide progress on water quality improvement in glowing terms—and to say nothing, not even by way of reference, on the process or progress of TMDLs.⁷⁷ On the ground, the Agency was committing itself to the less potentially confrontational tasks of distributing grant money for state nonpoint programs and promoting census-based "watershed planning."⁷⁸ Section 303(d) with its more objective and enforceable requirements remained, at the federal level, a voluntary program, one more thing the states might also do. It remained that way until a third wave of lawsuits jolted the Agency back onto the field.

The Roof Falls In

The most recent litigation would challenge the quality of the state submissions under § 303(d), and the adequacy of EPA's response. It would revive the question of whether "anything" as a TMDL would do. Suits filed in several venues from New York to the state of Washington began to close in, but the first to reach judgment were from Idaho and Georgia. From EPA's perspective, the roof fell in.

Idaho Sportsmen's Coalition v. Browner⁷⁹ began with the simple request that EPA develop WQLSs for a state that had submitted none until 1989, and then, in 1992, a grand total of 36.⁸⁰ On a "constructive submission" theory, this submission may have sufficed but the court went on to [27 ELR 10396] find EPA's approval to be arbitrary and contrary to law.⁸¹ Under court order, EPA ultimately approved a list identifying 962 Idaho WQLSs,⁸² a quantum leap in the identification of polluted waters that should shake the confidence of those reporting steady progress in the condition of the nation's waters. Phase two of the case led to another order, in 1994, that EPA proceed, in cooperation with Idaho, to develop a schedule for the development of TMDLs.⁸³ Phase three, in 1996, challenged the adequacy of that schedule.⁸⁴ In its 1996 decision, the court minced few words. Under the schedule offered by EPA and the state, TMDLs would not be completed for another 25 years, if then.⁸⁵ The proposed schedule set no deadlines, only "expected" dates and "targets," and assumed that the actual need for TMDLs was going to be far less than the 962 WQLSs, a conclusion to be borne out by post hoc evaluations.⁸⁶ At Idaho's proposed rate, the court noted, "the twenty-five years could easily turn into fifty orseventy-five"; "nothing in law could justify so glacial a pace."⁸⁷ The court remanded the decision to the Agency, with the "suggestion" that a completion date of five years would be "reasonable."⁸⁸

Before EPA could even begin to eliminate this precedent on appeal, it received the same verdict, and more, in Georgia. Sierra Club v. Hankinson⁸⁹ challenged everything about the Georgia program from the identification of WQLSs, to water quality monitoring, to the prioritization of state waters for TMDLs, to the number, adequacy, and pace of development for the TMDLs themselves.⁹⁰ The state had, in fact, made its first submission of of 123 WQLSs in 1992, augmented by the time of trial to 340 waters.⁹¹ For these waters, the state had submitted two TMDLs, was working on two more, and was projecting a total of 28 over the next 10 years.⁹² On cross-motions for summary judgment in March 1996, the court found itself unable to determine whether the WQLS list was adequate, setting it down for the unsettling prospect of trial on the merits.⁹³ The TMDL schedule, however, was found insufficient.⁹⁴ At Georgia's proposed pace, the court noted, it would take over 100 years to prepare TMDLs for only the waters currently on the list.⁹⁵ The few TMDLs underway, furthermore, were inadequate on their face as not considering load allocations from nonpoint sources or water conditions at other than lowest flows.⁹⁶ EPA went on to settle the WQLS issue with a new schedule agreeable to plaintiffs.⁹⁷ Then the second shoe dropped.

On August 30, 1996, the Hankinson court issued its order on TMDLs, requiring—where the Idaho court had only suggested—their completion within five years on a prescribed schedule, basin by basin.⁹⁸ Clear Water Act NPDES permits would be revised or terminated within one year following each new TMDL, and permits for new discharges into designated WQLSs would be accompanied by TMDLs to achieve water quality standards.⁹⁹ Should the state fail to comply, EPA would revise the state's delegated NPDES program to require it.¹⁰⁰ If the state further refused, EPA would "withdraw certification of the State NPDES program."¹⁰¹ The court, finally, would retain jurisdiction over the case and would receive a detailed report on Georgia's TMDL progress, annually.¹⁰²

EPA appealed the Georgia decision, but it could no longer afford to ignore the game. Lawsuits were pending in more than a dozen states and notice letters of intent to sue from citizen groups in a half dozen more.¹⁰³ In early 1996, EPA called for final 1996 state WQLS lists by April 1 of that year.¹⁰⁴ In February, the EPA Assistant Administrator for Water wrote directly to state environmental agency heads stating that TMDLs were a "critical step" and urging them to "support the TMDL program."¹⁰⁵ The response was not enthusiastic. By July 26, nearly four months past deadline, 34 states had not yet submitted final WQLS lists and 17 had submitted no lists at all.¹⁰⁶ EPA moved the deadline to October,¹⁰⁷ then to December.¹⁰⁸ EPA regional administrators were instructed to expedite their approval or disapproval of the state submissions, and where necessary to begin preparing their own.¹⁰⁹

When the dust had settled, in early 1997, all states and territories but three were in with something called a list.¹¹⁰

[27 ELR 10397]

How adequate they were, what kinds of TMDLs would follow, and in what time frame remained an open question. But at last, 25 years after the passage of § 303(d), the TMDL process had actually begun.

EPA Takes the Lead

Up to this point, EPA's response to § 303(d) was basically driven by lawsuits, court orders, and consent decrees. During the late fall of 1996, as the latest rulings from Georgia and Idaho were coming in, EPA's § 303(d) program director was heard to wish out loud: "If only we could just win one of these cases!"¹¹¹ To which the author asked why he would want to win: What else would move the Agency or the states to do TMDLs? The question was rhetorical, but the challenge to EPA at the start of 1997 was—having put on a full-court press to obtain state WQLS lists and blunt the most immediate thrust of litigation—whether it could get ahead of the curve and map an agenda for the implementation of § 303(d) that would begin to meet the statute's requirements and put some distance between itself and the courts.

To its credit—and overlooking the lateness of the hour—EPA launched several initiatives of its own, the most proactive of which were new policy guidance for its regions and the states, and a Federal Advisory Committee Act (FACA) committee. The thrust of the guidance was to try to move the process forward, ahead of the pursuing litigation. The thrust of the FACA committee was to try to achieve consensus among the states, and environmental groups, and potentially affected point and nonpoint source dischargers, over where and how the program would go. Each set a new tone for the debate.

In November 1996, impelled by a meeting with environmental litigants, EPA issued a draft TMDL Program Implementation Strategy¹¹² updating its 1991 guidelines and explaining the "vision, priorities and steps" it would take to "help States meet" TMDL program requirements;¹¹³ if at all possible, this would remain a state game. EPA's "vision" continued to emphasize "watershed approaches,"¹¹⁴ but for which it now recognized TMDLs were the "technical backbone";¹¹⁵ this would remain nominally a watershed game as well, but § 303(d)'s required load allocations were now the driver. The Strategy offered to ease state burdens by extending the frequency of reporting requirements,¹¹⁶ consolidating report categories,¹¹⁷ and providing more assistance on monitoring and assessments and protocols for the calculation of TMDLs.¹¹⁸ EPA also offered a TMDL SWAT team to provide quick assistance on TMDL development,¹¹⁹ and identified financial assistance available both from EPA sources and the programs of other agencies.¹²⁰

This said, the Strategy noted but left largely unresolved several thorny policy issues of TMDL implementation, among which were the eligibility of waters for listing, the degree of certainty required for TMDL calculations, the pace for TMDL development, and "reasonable assurances"¹²¹ that abatement strategies, particularly for nonpoint sources, would actually take place.¹²² The document would answer few of these questions and raise few hackles. The tough ones would be deferred to subsequent guidance and to a committee the agency was, at the same time, in the process of convening.

Also in November 1996, EPA appointed an advisory committee of 20 individuals with near-balanced representation of states, user groups, and the environmental community.¹²³ The stated purpose of the committee was to "provide consensus recommendations" on virtually every aspect of the TMDL program.¹²⁴ To arrive at this consensus, "constructive and substantive discussion" was called for among the stakeholders.¹²⁵ That discussion has begun, and will continue at least until mid-1998. Many of the issues were technical, and susceptible to relatively smooth resolution; others were more fundamental, and would take more time. Key among these were the scope of eligible waters, scientific uncertainty, implementation, and nonpoint sources.

Eligible Waters

Section 303(d) requires TMDLs for all waters for which effluent limitations are not sufficient to meet water quality standards.¹²⁶ EPA's regulations require states to prepare TMDLs for all waters that are not "expected" to meet water quality standards,¹²⁷ a criterion more in the eye of the beholder.¹²⁸ Under this guidance, some states have decided [27 ELR 10398] that below-standard waters did not need TMDLs because "other kinds of activities" were "planned or underway" to restore them.¹²⁹ To these states, this approach represented a reasonable allocation of resources. To environmentalists, it represented an end run around the load allocations of § 303. On its face, the statute does not provide the grace of exempting below-standard waters because of the expected effects of other abatement programs, or, for that matter, the expected effects of TMDLs. Like so many jurisdictional questions in environmental law—e.g., whether an area is a "wetland," whether a species is "endangered," whether a toxin is listed for public disclosure—the scope of a program goes a long way to determining its success. What remains unlisted, remains largely unremedied.

Scientific Uncertainty

The Achilles' heel of water quality standards-based regulation has always been the difficulty of ascribing and quantifying environmental effects for particular discharge sources. There is always another possible source, or another possible reason, that the fish in Lake Pontchartrain are dying. There is always an arguable threshold level for pollutants that may not harm fish, or for oxygen levels below 5 milligrams per liter. And when we come to more complex biological impacts such as the fate and effects of nutrients,¹³⁰ particularly those effects hundreds of miles downstream, we are beyond any pretense of precise mathematics for cause and effect decisions. The question is whether we are also, for these same reasons, beyond the reach of law.

We should not be. Section 303(d) itself speaks directly to the issue in requiring a margin of safety in its TMDLs in order to accommodate the uncertainty of its underlying science.¹³¹ Other major provisions of the Clean Water Act—and the Clean Air Act, hazardous waste laws, and wildlife and endangered species laws as well—require similar, "best guess" judgments when we are at the far edge of science and decisions need to be made.¹³² Reviewing courts have been generally quite tolerant of the mix of science and best guesswork that composes most risk-based environmental decisions.¹³³ "Good science"—a catch phrase of the 1990s—does not mean precision; it means the best science can do at the time.

This said, legal challenges are already rising over the degree of science necessary to support load calculations and their allocations to particular sources.¹³⁴ As these challenges mount, it will be important first to distinguish between the allocations and the calculations themselves. Allocation of loadings to particular sources in the TMDL process is entirely political, as it is in the analogous state implementation plan process of the Clean Air Act;¹³⁵ the mix of reductions from point and nonpoint sources a state may choose to meet its ambient standards is a matter for the state to legislate, negotiate, or otherwise determine.¹³⁶ As for the underlying calculations, it is reasonable that they be rationally derived from the best available data; it is unrealistic to require more. Indeed, it would be fatal.

In the most significant challenge to TMDL calculations to date, involving dioxin loadings on the Lower Columbia River, the calculations were upheld over a variety of objections to the methods by which they were derived.¹³⁷ If the courts continue to understand the limits of science and its role in TMDL decisionmaking, as Congress did, and as the Lower Columbia court did, the process will move forward toward its abatement goals. If, on the other hand, courts begin to require the hypertechnical, isolated, cause- [27 ELR 10399] and-effect kinds of proof that are emerging from cases in hazardous waste regulation¹³⁸ and toxic torts,¹³⁹ TMDLs will never get off the ground. Science, in this area of the law, will never deliver precision.

Implementation

Implementation is of course where all public laws live or die, and environmental laws have, over time, accreted layers of overlapping responsibilities and enforcement mechanisms to ensure that, at day's end, something at least remotely like what Congress intended has happened. Section 303(d) muddies the issue because Congress plainly intended states to implement the program, and for EPA to backstop it only where the states failed to do their jobs. Two implementation issues arise. The first is the time frame within which TMDLs are to be established. The second is the mechanism by which they are to be carried out. Unfortunately, for these are critical issues to the success of the program, the language of the statute is not very helpful.

With regard to the time frame, § 303(d) provides only that states shall submit their WQLSs and TMDLs to EPA "from time to time,"¹⁴⁰ with the first submission due within six months of EPA's identification of the applicable pollutants, or June 26, 1979.¹⁴¹ It is obvious that Congress presumed a certain amount of good-faith effort here, a presumption that was not entirely borne out in EPA's regulations that called for one state submission and little more,¹⁴² or in state performance. As it turned out, this dilatory pace prompted courts to begin to impose their own deadlines, as short as five years for TMDLs on all of a state's listed waters.¹⁴³ EPA has subsequently proposed policy guidance that would require the submission of all TMDLs, from all states, within 8 to 13 years.¹⁴⁴ Whether even this pace can be maintained, and whether it will produce load allocations and plans of sufficient quality to be effective, are legitimate and difficult questions. Within them lie several issues of assessment and monitoring technology and of the accommodation of uncertainty. These issues pale, however, before the root question of this issue: do the resulting load allocations have to be implemented at all?

Once WQLSs and TMDLs are prepared, the language of § 303(d) ends. Section 303(e) proceeds to require a "continuing planning process" (CPP) with "plans" that "include" § 303(d)'s TMDLs.¹⁴⁵ While these sections authorize EPA to approve or disapprove a CPP on the basis, inter alia, of TMDLs, they do not authorize the Agency to implement them. The question is, at this point, has the statute run its string? Does all the work of TMDLs and their load allocations wind up as references in state plans, implemented if and as the states may wish? Or does the TMDL itself have to include the means of its own implementation in order to receive EPA's approval? EPA's authority to review and reject TMDLs¹⁴⁶ may succeed in securing the inclusion of those steps and commitments necessary to implement them, retaining some meaningful outcome for the process. Further, if these measures are inadequate, EPA may reject a TMDL and may then promulgate measures of its own in a federal TMDL.¹⁴⁷ But then what? For point sources, the Agency may ensure that those additional limitations imposed by a TMDL are actually implemented through its supervision of discharge permits under the NPDES program.¹⁴⁸ But for nonpoint sources, here is the rub: there are no federal controls over nonpoint sources under the Clean Water Act.¹⁴⁹ For these sources, the § 303(d) program leads, ultimately, to a state prerogative.¹⁵⁰ If, as will next be discussed, it is found to cover point sources at all.

Nonpoint Sources

The big enchilada. As described in an earlier Article,¹⁵¹ nonpoint source pollution has become the dominant water quality problem in the United States, dwarfing all other sources by volume and, in conventional contaminants, by far the leading cause of nonattainment for rivers, lakes, and estuaries alike. It is no secret to any observer of the Clean Water Act that the primary reason for this mushrooming problem is the fact that while other sources have been abated through required controls and their enforcement, no comparable controls or enforcement have been applied to agriculture, silviculture, and the rest of the nonpoint world.¹⁵² Enter, now, TMDLs, with the potential for specific, quantified load allocations (i.e., reductions) from nonpoint sources. The nonpoint world quakes. And reacts.

From the outset of the FACA committee, agricultural interests made their view clear that § 303(d) does not apply to nonpoint sources.¹⁵³ In their reading of the Act, TMDLs are to be set after the exhaustion of emission limitations, and since emission limitations are only set for point sources, TMDLs are restricted to point sources as well.¹⁵⁴ An earlier Article described the genesis of § 303(d) and its language and concluded that, while the section is entirely silent as to whether it applies to only point, only to nonpoint, or to both point and nonpoint sources, the members of the House Public Works Committee, where this section was born, were [27 ELR 10400] well aware that nonpoint sources contributed significantly to the failure to attain water quality standards, and that the most logical reading of the process they arrived at is to read nonpoint sources as included.¹⁵⁵ Indeed, in both the context of that time and the present, TMDLs for point sources alone make no pollution control sense at all.

This reading of the statute is bolstered by the fact that EPA has given the statute the same reading, consistently, from the issuance of its first regulations. As early as 1975, the Agency was calling for the allocation of nonpoint source loadings within TMDLs.¹⁵⁶ It has continued to do so in its regulations, guidelines, and draft strategy through 1996, 20 years of consistent agency interpretation.¹⁵⁷ Under the U.S. Supreme Court's ruling in Chevron, U.S.A., Inc. v. Natural Resources Defense Council,¹⁵⁸ and since,¹⁵⁹ on deference to an agency's interpretation of its statutory mandates, this regulatory history should be dispositive. Within the FACA committee, however, EPA has recognized the existence of the opposing view.¹⁶⁰ The opposing view, indeed, has taken the counter-offensive.

On April 29, 1997, the U.S. Forest Service wrote EPA to protest, and reject, the application of § 303(d) to nonpoint sources.¹⁶¹ The section "was written with point sources in mind," it said, and EPA's approach "may not satisfy the requirements of the Act."¹⁶² Getting to the point, the Forest Service wrote that EPA was "exposing all Federal and State agencies to additional litigation" over nonpoint source controls.¹⁶³ So much was, in fact, true; siltation from forest roads and washouts from clearcuts had eliminated trout and salmon from the then one mountain river¹⁶⁴ and, in the spring of 1997, wiped out quite a few human habitations as well.¹⁶⁵ The Forest Service, as many states, believed in the application of management practices without specific limits on pollutants.¹⁶⁶ In support for this argument, the Forest Service pointed to § 319¹⁶⁷ of the Clean Water Act, adopted in 1987 specifically to address nonpoint sources of pollution.¹⁶⁸ Section 319, it wrote, was and should be the exclusive remedy.¹⁶⁹

The argument is essentially political. As a legal argument, it is thin. In enacting § 303(d), Congress, at best, said nothing about whether nonpoint sources were in or out, and would have had to have been insane to, on the one hand, spell out the TMDL process, and on the other, exclude those nonpoint sources it recognized at the time were so much the cause of the problem.¹⁷⁰ In enacting the 1987 Amendments with § 319, Congress likewise said nothing about TMDLs, pro or con, and whether nonpoint sources were included or excluded.¹⁷¹ EPA's TMDL regulations including nonpoint sources had, by that time, been on the books for 11 years. The notion that Congress repealed a federal program without referring to it by enacting another is novel to environmental law. This argument has been considered and rejected with regard to a quite similar claim that § 303(d)'s application to toxic dischargers was preempted by Congress' subsequent enactment, also in the 1987 Amendments, of a separate water quality-based program for toxins.¹⁷² In fact, Congress has, year after year, thrown program after program into the breach against specific environmental problems, often overlapping or duplicative programs, often without thinking through how the new and the old would mesh.¹⁷³ As of 1997, for example, it had enacted no fewer than six different strategies under the Clean Water Act to abate toxic pollution.¹⁷⁴ All remain viable. There is rarely only one way in environmental law.

In practice, §§ 303(d) and 319's approaches to nonpoint source pollution are compatible and, indeed, mutually supportive. Within the statutory scheme, § 319 is the carrot, funding state programs for nonpoint source abatement statewide, for all waters whether they are currently above standard or below.¹⁷⁵ In keeping with its broad sweep, § 319's provisions are voluntary. States may choose to participate or not; participating states may choose regulatory approaches or not.¹⁷⁶ Section 303(d), on the other hand, addresses a narrower and more nasty job: the chronically [27 ELR 10401] polluted waters of the United States. For this problem zone, enter a stick: quantified pollution load allocations. The nature of the allocations and of the implementing controls remain up to the states, but states do have to come up with them. The Clean Water Act is, after all, about restoring the nation's waters. In this context, the existence of both §§ 303 and 319 makes sense. Or at least as much sense as the multiple programs of the Clean Water Act, the Clean Air Act, and similar pollution control laws often do.

Nonetheless, given the stakes at issue and the vehemence with which the U.S. Department of Agriculture and its nonpoint source constituents will fight quantified abatement requirements, this question is far from resolved in the federal family. To say nothing of what a constituency-oriented Congress may do. At this juncture, the only safe observation from the fact and scale of nonpoint source pollution and its effect on achieving water quality standards is that, unless TMDLs include quantified restrictions on nonpoint sources, they are wasting everyone's time.

But Will Anybody Follow?

The Clean Water Act is an experiment in cooperative federalism. No such experiments are easy. Section 303(d) is a holdover from an earlier day of state supremacy, with a veneer of federal supervision and control. It is also a holdover from an earlier philosophy that pollution could be abated by calibrating the effects of individual discharges, a proposition that has yet to be proven in air, water, or any medium. The Congress that enacted § 303(d) was equally suspicious both of state enthusiasm for the hard work of pollution control and of the water quality standards method of regulation. But it was willing to give them a shot. Twenty-five years later, that shot seems to be ready to launch.

The effort, over more than a decade, to bring the states' responsibilities under § 303(d) into play, the targeting of polluted waters, and the load reductions that Congress prescribed, should be a little sobering to those who champion the devolution of environmental responsibilities away from federal mandates. The states have been all in favor of the responsibility for regulating water pollution through their water quality standards, right up to the point that they had to do it. When the curtain was finally opened on the number of state waters below standard and without cleanup plans, the news was jolting. And when it finally became time to prepare TMDLs for these waters, all of a sudden this method of regulation was, according to many states, too complex and unwieldy. Load calculations could not be made for even the most common pollutants. Load allocations could not be made for nonpoint sources. We just could not get there from here.

Whether we actually will get there from here remains an open question. But the difficulties surfacing—once again—in water quality standards-based regulation should also throw new light on the relative effectiveness of technology-based standards. Over the past 20 years, it has been increasingly fashionable to criticize technology-based regulation as dictatorial, innovation-stifling, wasteful, and excessive. When one, now, begins to appreciate the full scale of the water quality-based alternative—perhaps 50,000 WQLSs eligible for TMDLs at an estimated $ 1 million per study¹⁷⁷ and an order of magnitude times that amount more for implementation, with no assurance of real load reductions at the far end—technology standards may begin to look like quite a bargain.

There is yet one more lesson here. The same effort, over more than a decade, to make the program Congress envisioned in § 303(d) happen should also be informative to those who seek to alter the role of citizen groups in the implementation of the Clean Water Act and other laws. So long as EPA's responsibilities were viewed as discretionary and unreviewable, virtually nothing happened. Few WQLSs. Fewer TMDLs. On this provision of law that was vital, indeed pivotal, to the enactment of the statute as we know it today—the very passage of the 1972 Act hinged on it—Congress had in effect wasted its time. The litigation that began to effectuate § 303(d) was no grand conspiracy of the environmental movement. The plaintiffs in Scott, the Alaska cases, and Northwest Environmental Defense Center were unrelated and indeed unaware of each other. But in their individual and entrepreneurial fashion, they put the law to work. The genius of American public environmental law—and the reason American laws work where the similar and often stronger-looking laws of other countries do not—is brought home again by this experience. We have a three-part government, and the American public can go to court.

But at bottom, the courts can only go so far. Indeed, under § 303(d), EPA can only go so far. At some point, through leverage, funding and hard negotiation, the states are going to have to buy into the program. Even if they do, it is not yet clear that the resulting TMDLs will be what Congress anticipated, the fallback engine for restoring the nation's waters. A final Article in this series will review those TMDLs that are coming out of the system, and reflect on the prospects for fulfilling the expectations of Congress and of a surprising, to some, number of Americans for whom clean water is a primary concern.

1. 33 U.S.C. §§ 1251-1387, ELR STAT. FWPCA §§ 101-607. Technically the "Federal Water Pollution Control Act," in 1977 Congress ceded to common usage and recognized the "Clean Water Act" as well. See Clean Water Act of 1977, Pub. L. No. 95-217, § 2, 91 Stat. 1566 (1977) ("This Act may be cited as the 'Federal Water Pollution Control Act' (commonly referred to as the Clean Water Act).").

2. Section 303 of the Federal Water Pollution Control Act Amendments of 1972 retained the water quality standards-based regulatory approach of earlier federal legislation, with significant strengthening provisions, chief among which was § 303(d). 33 U.S.C. § 1313(d), ELR STAT. FWPCA § 303(d). Section 303(d) requires states to (1) identify waters that will remain in violation of water quality standards after the imposition of technology controls (e.g., water quality limited segments, or WQLSs), id. § 1313(d)(1)(A), ELR STAT. FWPCA § 303(d)(1)(A), (2) prioritize these waters for additional cleanup measures, id.; (3) identify load allocations that will reduce discharges to meet water quality standards (e.g., total maximum daily loads, or TMDLs), id. § 1313(d)(1)(C), ELR STAT. FWPCA § 303(d)(1)(C); and (4) incorporate these load allocations into discharge permits and state water quality plans, id. § 1313(e)(3)(C), (F), ELR STAT. FWPCA § 303(e)(3)(C), (F).

3. See Oliver Houck, TMDLs: The Resurrection of Water Quality Standards-Based Regulation Under the Clean Water Act, 27 ELR 10329 (July 1997), and sources cited therein. The description that follows is taken from this Article.

4. Id.

5. Pub. L. No. 92-500, 86 Stat. 816 (codified as amended in scattered sections of 33 U.S.C.).

6. For a description of this task and its surrounding litigation, see Oliver Houck, The Regulation of Toxic Pollutants Under the Clean Water Act, 21 ELR 10528, 10537-39 (Sept. 1991), and sources cited therein.

7. Jeffrey M. Gaba, Federal Supervision of State Water Quality Standards Under the Clean Water Act, 36 VAND. L. REV. 1167, 1189 (1983) (discussing EPA's antidegradation policy). EPA's current water quality standards policies are codified at 40 C.F.R. §§ 131.10, .11, and .12 (1996).

8. Gaba, supra note 7, at 1194-95.

9. Id. at 1194.

10. Id. at 1190.

11. 2 WILLIAM H. RODGERS, ENVIRONMENTAL LAW: AIR & WATER 289 n.3 (1986).

12. William F. Pedersen Jr., Turning the Tide on Water Quality, 15 ECOLOGY L.Q. 69, 99, 102 n.155 (1988).

13. Gaba, supra note 7, at 1210 (citing Mississippi Comm'n on Natural Resources v. Costle, 625 F.2d 1269, 10 ELR 20931 (5th Cir. 1980)).

14. 33 U.S.C. § 1251(b), ELR STAT. FWPCA § 101(b).

15. Id. § 1313(d)(2), ELR STAT. FWPCA § 303(d)(2). EPA's identification of pollutants was to be done pursuant to § 304(a)(2)(D). Id. § 1314(a)(2)(D), ELR STAT. FWPCA § 304(a)(2)(D).

16. Id. § 1313(d)(2), ELR STAT. FWPCA § 303(d)(2).

17. 38 Fed. Reg. 29646 (Oct. 26, 1973).

18. Environmental Defense Fund v. Costle, 657 F.2d 275, 294-95, 11 ELR 20459, 20469-70 (D.C. Cir. 1981) (EPA did not identify salinity as a pollutant until December 28, 1978 and thus the states' duty to submit TMDL calculations did not arise until June 28, 1979; The Environmental Defense Fund's (EDF's) claim was premature since EPA did not have the occasion to approve or disapprove the state TMDL submissions before the EDF filed its motion for summary judgment.).

19. Homestake Mining Co. v. EPA, 477 F. Supp. 1279, 1288 (D.S.D. 1979) (EPA had not identified the pollutants at the time of the Cheyenne River Basin Plan's adoption; although South Dakota had not established TMDLs as required by § 303(d), the court held that they were not yet required until 180 days after EPA's identification of pollutants.).

20. Preparation of Water Quality Management Basin Plans, 39 Fed. Reg. 19634 (June 3, 1974) (codified at 40 C.F.R. § 130).

21. 33 U.S.C. § 1256, ELR STAT. FWPCA § 106; id. § 1288, ELR STAT. FWPCA § 208; id. § 1313(e), ELR STAT. FWPCA § 303(e).

22. Determination of Total Maximum Daily Loads, 39 Fed. Reg. 19641 (June 3, 1974) (codified at 40 C.F.R. § 131.304(a)).

23. Prohibition of Approval of Certain Planning Processes; Withdrawal of Process Approval, 39 Fed. Reg. 19639 (June 3, 1974).

24. Natural Resources Defense Council v. Train, 396 F. Supp. 1386, 1392-93, 5 ELR 20405, 20407 (D.D.C. 1975).

25. Preparation of Water Quality Management Plans, 40 Fed. Reg. 55344 (Nov. 28, 1975) (codified at 40 C.F.R. § 131(1)).

26. Plan Content, 40 Fed. Reg. 55346 (Nov. 28, 1975) (codified at 40 C.F.R. § 131.11(f)(3)(ii)).

27. Plan Content, supra note 26, at 40 C.F.R. § 131.11(g)(3).

28. Board of County Comm'rs v. Costle, No. 78-0572, slip op. (D.D.C. June 20, 1978), (cited in Total Maximum Daily Loads Under Clean Water Act, 43 Fed. Reg. 42303 (Sept. 20, 1978).

29. Total Maximum Daily Loads Under Clean Water Act, 43 Fed. Reg. 42303 (Sept. 20, 1978).

30. Mr. Thomas Jorling was Minority Counsel to the Senate Committee on Public Works and, subsequently, EPA Assistant Administrator for Water Waste Management, 1977-79.

31. Total Maximum Daily Loads Under the Clean Water Act, 43 Fed. Reg. 60664 (Dec. 28, 1978) [hereinafter Total Maximum Daily Loads]. What EPA may have intended by its allusion to congressional intent was the expectation that technology standards take priority over water quality standards in the implementation of the 1972 Amendments. See Houck, supra note 3, at 10337-38, and the referenced comments of Senator Muskie.

32. Total Maximum Daily Loads, supra note 31, at 60666.

33. Id.

34. Id.

35. Id. at 60662.

36. Scott v. City of Hammond, 530 F. Supp. 288 (N.D. Ill. 1981), aff'd in part, rev'd in part, 741 F.2d 992, 14 ELR 20631 (7th Cir. 1984). EPA argued that Congress did not intend that EPA establish TMDLs if the states chose not to act. To support its position, EPA pointed out several instances wherethe CWA explicitly requires EPA to intercede in the absence of state action; thus, by negative implication, congressional intent was to rely exclusively on the states to set the TMDL machinery in motion because under § 303(d) there is no explicit CWA requirement for EPA to act in the absence of state action. The argument leads to the anomalous conclusion that EPA intervention is called for in response to inadequate state performance, but not in response to no state performance.

37. 530 F. Supp. at 290.

38. 741 F.2d 992, 14 ELR 20631.

39. Id. at 996, 14 ELR at 20632.

40. Id. at 997, 14 ELR at 20633.

41. Id. at 998, 14 ELR at 20634.

42. Id. at 997 n.11, 14 ELR at 20633 n.11.

43. Water Quality Standards Regulation, 48 Fed. Reg. 51400 (Nov. 8, 1983) (codified at 40 C.F.R. §§ 35, 120, 131).

44. Id. at 51404.

45. State Review and Revision of Water Quality Standards, 48 Fed. Reg. 51400, 51407 (Nov. 8, 1983) (codified at 40 C.F.R. § 131.20).

46. No. 86-1578PA (D. Or. complaint filed Dec. 12, 1986); see also Northwest Envtl. Defense Ctr. v. Thomas, No. 86-1578 BU (D. Or. consent decree filed June 3, 1987).

47. Alaska Ctr. for the Env't v. Reilly, 762 F. Supp. 1422, 21 ELR 21305 (W.D. Wash. 1991); Alaska Ctr. for the Env't v. Reilly, 796 F. Supp. 1374, 22 ELR 21204 (W.D. Wash. 1992), aff'd sub nom. Alaska Ctr. for the Env't v. Browner, 20 F.3d 981, 24 ELR 20702 (9th Cir. 1994).

48. 762 F. Supp. at 1425, 21 ELR at 21306.

49. Id. at 1428, 21 ELR at 21307.

50. 796 F. Supp. at 1379, 22 ELR at 21206 (citing Affidavit of Robert Burd at 7, Exhibit 3, Defendants' Opposition to Plaintiffs' Motion to Compel).

51. 796 F. Supp. at 1379, 22 ELR at 21206.

52. No. 87 C 4196, 1991 WL 47374, at *1 (N.D. Ill. Mar. 28, 1991).

53. Id. at *5.

54. 843 F. Supp. 1304, 24 ELR 21006 (D. Minn. 1993).

55. Id. at 1308, 24 ELR at 21007.

56. Id.

57. Id. at 1314, 24 ELR at 21010.

58. Id. ("Although Minnesota and the EPA may not be implementing TMDLs as quickly as plaintiffs would like, the Act does not set deadlines for the development of a certain number of TMDLs. The Act instead requires the development of TMDLs 'in accordance with the priority ranking' of the WQLS list.").

59. GAO, WATER POLLUTION: MORE EPA ACTION NEEDED TO IMPROVE THE QUALITY OF HEAVILY POLLUTED WATERS, GAO REPORT TO THE CHAIRMAN: SUBCOMMITTEE ON REGULATION AND BUSINESS OPPORTUNITIES COMMITTEE ON SMALL BUSINESS HOUSE OF REPRESENTATIVES (Jan. 1989).

60. Id. at 4.

61. Id. at 24.

62. Id. at 23.

63. Id. at 6.

64. Id.

65. Id.

66. Id.

67. Id. at 34-36.

68. Tom Wilson, Taking the Fear Out of TMDLs, NONPOINT SOURCES NEWS—NOTES, Oct. 1990, at 19-21.

69. Id. at 19.

70. Id.

71. Id. at 20.

72. Id.

73. Id. at 20, 21.

74. The minimalist nature of EPA's implementation of § 303 is reflected in a 1996 memorandum from EPA Region VIII to the state of Wyoming stating that "past decisions have allowed for 'functional equivalent TMDLs' when it is shown that actions on the part of the state met the substantive intent of the Act (reference: Sportsmen's Clubs of Texas v. Layton, No. CA 3-86-0121-R N.D. Texas)." Letter from Carol Campbell, Ecosystems Protection Programs, Region VIII, EPA, to Gary Beach, Administrator, Water Quality Division, Department of Environmental Quality, Wyoming (June 26, 1996) (on file with author). To the contrary, a review of the docket of the captioned case shows no decision, order, or other resolution on "functional equivalency," or any other substantive issue in the case. Up to June 1996, EPA Regions were still finding slender, if not inventing, reasons to avoid promulgating TMDLs.

75. EPA, Pub. No. 440/4-91-001, GUIDANCE FOR WATER QUALITY-BASED DECISIONS: THE TMDL PROCESS (Apr. 1991).

76. Memorandum from Geoffrey H. Grubbs, Director, Assessment and Watershed Protection Division, EPA, to Water Quality Branch Chiefs, Regions I-X; Approval of [§ [1303(d) Lists: Promulgation Schedules/Procedures, Public Participation (Oct. 30, 1992) (on file with author).

77. EPA, NATIONAL WATER QUALITY INVENTORY, 1994 REPORT TO CONGRESS (Dec. 1995).

78. Memorandum from Robert Perciasepe, Assistant Administrator for Water, EPA, A Healthy Watershed Strategy (Aug. 9, 1996) (on file with author).

79. 951 F. Supp. 962, 964 (W.D. Wash. 1996) (describing procedural background of case).

80. Id.

81. Id. at 969.

82. Id. at 964.

83. Id.

84. Id. at 962.

85. Id. at 964.

86. Id. at 967.

87. Id.

88. Id. at 969.

89. No. 1:94-cv-2501-MHS, 1996 WL 534909, at *1 (N.D. Ga. Mar. 25, 1996).

90. Id. at *3.

91. Id.

92. Id.

93. Id. at *6.

94. Id.

95. Id. at *1.

96. Id. at *6.

97. Sierra Club v. Hankinson, 939 F. Supp. 872, 873 (N.D. Ga. 1996).

98. Sierra Club v. Hankinson, No. 1:94-cv-2501-MHS, 1996 WL 534914, at *1 (N.D. Ga. Aug. 30, 1996).

99. Id. at *2.

100. Id.

101. Id.

102. Id.

103. As of December 12, 1996, EPA was under separate court orders to propose TMDLs in Oregon, Alaska, and Georgia; litigation to compel WQLS lists and/or TMDLs were pending in Idaho, New York, Georgia (WQLSs), New Jersey, Pennsylvania, Delaware, West Virginia, Louisiana, New Mexico, Kansas, California (two cases), Washington, and Oregon; notices of intent to sue had been filed in Alabama, Florida, Mississippi, North Carolina, Wyoming, and Arizona. EPA, TMDL Litigation (Dec. 2, 1996) (unpublished memorandum on file with author). EPA may have overlooked its still-pending case in Texas. See supra note 75.

104. Memorandum from Robert Perciasepe, Assistant Administrator for Water, EPA, to Regional Administrators, EPA Regions I-X, EPA Action on 1996 Lists, Priority Rankings and TMDL Targeting Plans Submitted by States Under [§ [1303(d) of the Clean Water Act (Aug. 9, 1996).

105. Memorandum from Robert Perciasepe, Assistant Administrator for Water, EPA, to State Environmental Commissioners and Regional Administrators, Total Maximum Daily Loads: A Key to Improving Water Quality (Feb. 26, 1996).

106. Memorandum from Robert Perciasepe, supra note 104.

107. Water Pollution: 'Staged Approach' Outlined by EPA for Dealing With Listings of Polluted Waters, Setting of TMDLs, [27 Current Developments] Env't Rep. (BNA) 925 (Aug. 23, 1996).

108. Memorandum from Robert Perciasepe, supra note 104.

109. Id.

110. Oral presentation of Geoffrey H. Grubbs, Director of EPA Water Office Assessment and Watershed Protection Division, Federal Advisory Committee Meeting on TMDLs in Galveston, Texas (Feb. 19-21, 1997).

111. Conversation with Geoffrey H. Grubbs, EPA, in Washington, D.C. (Dec. 5, 1996).

112. EPA, Draft TMDL Program Implementation Strategy (Nov. 18, 1996) [hereinafter Draft Strategy].

113. Id. at 3.

114. Id.

115. Id.

116. Id. at 10.

117. Id. at 12.

118. Id. at 8.

119. Id. at 16.

120. Id. at 18, 19. EPA identified CWA § 319 (nonpoint pollution), State Revolving Funds, and Farm Bill conservation programs. For example, the 1996 Farm Bill's Environmental Quality Incentives Program provides over $ 2 billion available annually which, if appropriately directed, could provide major assistance in implementing TMDLs. Id.

121. Id. at 22.

122. Id. at 18. On the last and critical point, the Agency would "carefully consider" the "potential consequences to a pollution source in the event that implementation does not occur," such as penalties or sanctions.

123. EPA Press Release (Nov. 19, 1996) (on file with author). The committee includes four state representatives, four representatives of environmental organizations, and at least four representatives of point and nonpoint source discharges. Id. at 3.

124. Id. at 2. The only subject areas off limits to the FACA committee discussions are legislation, appropriations, and litigation. Wisely so, if the committee was to have any hope of achieving consensus. Introductory Remarks of Geoffrey H. Grubbs, Assessment and Watershed Protection Division, Federal Advisory Committee on TMDLs (Nov. 19-21, 1996).

125. Id.

126. 33 U.S.C. § 1313(d)(1)(A), (C), ELR STAT. FWPCA § 303(d)(1)(A), (C).

127. 40 C.F.R. § 130.7(c)(ii) (1996) ("TMDLs shall be established for all pollutants preventing or expected to prevent attainment of water quality standards.").

128. EPA has recognized the softness of this interpretation, and has indicated its intention to revisit the question. See Draft Strategy, supra note 112, at 12 ("EPA will further explain the conditions under which a TMDL is not necessary and when a water does not need to be listed.").

129. Draft Strategy, supra note 112, at 7.

130. For the difficulties in modeling the effects of nutrient loadings, see EPA, Background Paper No. 4: Science and Technology, FACA Committee Meeting (Nov. 19-21, 1996).

Certain types of water quality problems and processes are still poorly understood. For example, excessive nutrients in rivers can cause the growth of attached algae. The relationship between nutrient loading and attached algal growth is very difficult to quantify and therefore definition of an acceptable load and determination of a TMDL is problematic.

Id. at 4. These difficulties noted, the downstream effects of nutrient loadings can be enormous. Louisiana is currently experiencing an 8,000 square mile "dead zone" of anaerobic water at the mouth of the Mississippi River, due largely to nutrient runoff from farms in the Midwest, over 1,000 river miles away. ELIZABETH COLEMAN, THE PERSISTENT ENEMY: HYPOXIA ON LOUISIANA'S CONTINENTAL SHELF, COAST AND SEA: MARINE AND COASTAL RESEARCH IN LOUISIANA UNIVERSITIES 7 (1992). More than 70 percent of the total nitrogen delivered to the Gulf of Mexico by the Mississippi River originates above the confluence of the Ohio and Mississippi. Richard Alexander et al., The Regional Transport of Point and Nonpoint Source Nitrogen to the Gulf of Mexico, EPA Gulf of Mexico Hypoxia Management Conference (Dec. 5, 1995).

131. "Such load shall be established at a level necessary to implement the applicable water quality standards with seasonal variations and a margin of safety that takes into account any lack of knowledge concerning the relationship between effluent limitations and water quality." 33 U.S.C. § 1313(d)(1)(c), ELR STAT. FWPCA § 303(d)(1)(c).

132. 42 U.S.C. § 7409(b)(1), ELR STAT. CAA § 109(b)(1); 33 U.S.C. § 1317(a)(4), ELR STAT. FWPCA § 307(a)(4); 42 U.S.C. § 9621(d)(1), ELR STAT. CERCLA § 121(d)(1); 16 U.S.C. § 1536(a)(2), ELR STAT. ESA § 7(a)(2); for a discussion of its "best guess" standard, see Oliver Houck, The "Institutionalization of Caution" Under Section 7 of the Endangered Species Act: What Do You Do When You Don't Know, 12 ELR 15001 (Apr. 1982).

133. See Lead Indus. Ass'n v. EPA, 647 F.2d 1130, 10 ELR 20643 (D.C. Cir. 1978), cert. denied, 449 U.S. 1042 (1980) (affirming national ambient air quality standards for lead based on a series of conservative assumptions on risk and exposure, over a host of industry objections); Hercules, Inc. v. EPA, 598 F.2d 91, 8 ELR 20811 (D.C. Cir. 1978) (water quality standards); Environmental Defense Fund v. EPA, 598 F.2d 62, 8 ELR 20765 (D.C. Cir. 1978) (polychlorinated biphenyl standards); Ethyl Corp. v. EPA, 541 F.2d 1, 6 ELR 20267 (D.C. Cir. 1976) (lead levels in gasoline).

134. Lewis County Util. Corp. v. Washington Dep't of Ecology, No. 96-043, 1997 WL 240790 (Wash. Pollution Control Bd. Apr. 18, 1997) (challenging denial of water withdrawal permit, alleging a "lack of evidence that the amount of water requested would have a significant or measurable impact"); see also Statement on Behalf of Chemical Manufacturers Association, FACA Meeting (Nov. 19-21, 1996), in EPA, DRAFT SUMMARY OF MEETING ONE 6 (urging that no TMDL restrictions be imposed until the science of each assessment is corroborated). The degree of supporting science is clearly where the challenges to TMDLs will lie. These challenges were, of course, instrumental in emasculating water quality-based regulation under the Federal Water Pollution Control Act of 1965. See Houck, supra note 3, at 10330 n.10.

135. 42 U.S.C. § 7410, ELR STAT. CAA § 110; see WILLIAM H. RODGERS JR., Implementation Plans: Procedure and Evaluation, in ENVIRONMENTAL LAW 196-202 (1994).

136. Union Elec. Co. v. EPA, 427 U.S. 246, 6 ELR 20570 (1976).

137. Dioxin/Organochlorine Ctr. v. Rasmussen, 37 ERC 1845 (W.D. Wash. 1993), aff'd sub nom. Dioxin/Organochlorine Ctr. v. Clarke, 57 F.3d 1517, 25 ELR 21258 (9th Cir. 1995). Among the points at issue were EPA's assumptions of safe levels of exposure, and its margin of safety at 22 percent of total loadings.

138. United States v. Ottati & Goss, Inc., 900 F.2d 429, 20 ELR 20856 (1st Cir. 1990) (reversing Superfund cleanup decision).

139. See Brock v. Merrell DOW Pharmaceuticals, Inc., 874 F.2d 307 (5th Cir. 1989) (epidemiological evidence insufficient to show causation); Turpin v. Merrell DOW Pharmaceuticals, Inc., 959 F.2d 1394 (6th Cir. 1992) (animal studies evidence insufficient proof); Sorensen v. Shaklee Corp., 31 F.3d 638 (8th Cir. 1994) (ingestion of chemically treated substances insufficient evidence of causation).

140. 33 U.S.C. § 1313(d)(2), ELR STAT. FWPCA § 303(d)(2).

141. See supra text accompanying note 31. EPA issued its identification on December 28, 1978.

142. See supra text accompanying note 33.

143. See Sierra Club v. Hankinson, No. 1:94-cv-2501-MHS, 1996 WL 534909, at *1 (N.D. Ga. Mar. 25, 1996).

144. Draft Memorandum of Robert Perciasepe, Assistant Administrator for Water, EPA, New Policies for Developing and Implementing Total Maximum Daily Loads (TMDLs) 2 (Mar. 21, 1997).

145. 33 U.S.C. § 1313(e)(1), ELR STAT. FWPCA § 303(e)(1).

146. Id. § 1313(d)(2), ELR STAT. FWPCA § 303(d)(2).

147. Id.

148. Id. § 1342(i), ELR STAT. FWPCA § 402(i) (EPA review of NPDES permits).

149. See Houck, supra note 3, at 10342.

150. EPA, of course, holds residual leverage that could be important in influencing a state to regulate nonpoint pollution, including direct funding assistance, related point-permit review, and approval of state-delegated programs. This leverage and its limits will be further explored in an ensuing Article.

151. Houck, supra note 3, at 10342-43.

152. Id., and sources cited therein.

153. Conversation with Richard Parrish, FACA committee member (Mar. 11, 1997).

154. Id.

155. Houck, supra note 3, at 10337.

156. Preparation of Water Quality Management Plans, supra note 25, at 55346, § 131.11(f)(3) ("for each water quality segment, a total allocation for point sources of pollutants and a gross allotment for nonpoint sources of pollutants"). These gross allocations were then to lead to point source load allocations, id. at 55346, § 131.11(g), and nonpoint controls, id. at 55346, § 131.11(j).

157. See Draft Strategy, supra note 112, at 20.

158. 467 U.S. 837, 14 ELR 20507 (1984) (deference to EPA's somewhat fluid policy on nonattainment under the Clean Air Act).

159. See, e.g., Babbitt v. Sweet Home Chapter of Communities for a Great Oregon, 115 S. Ct. 2407, 25 ELR 21194 (1994) (deference to U.S. Department of the Interior's interpretation of § 9 of the Endangered Species Act).

160. EPA, Background Paper No. 2: Criteria for EPA Approval of State, Tribal TMDLs, FACA Committee Meeting 2 (Nov. 19-21, 1996) ("It has been suggested that since TMDLs do not have a direct regulatory effect (except through NPDES permits), they should not be a high priority, especially in watersheds impaired mainly by nonpoint sources. On the other hand, it has been argued that nonpoint source management programs should be guided by TMDLs and load allocations.").

161. Letter of Arthur Bryant, Director, Watershed and Air Management, U.S. Forest Service, to Geoffrey H. Grubbs (Apr. 29, 1997).

162. Id. at 1.

163. Id. at 2.

164. See National Wildlife Fed'n v. U.S. Forest Serv., 592 F. Supp. 931, 14 ELR 20755 (D. Or. 1984) (describing destruction of salmon streams by clearcutting and resultant landslides).

165. See Hal Bernton, Local Stories of Clear-cuts & Mudslides, PORTLAND OREGONIAN, Feb. 2, 1997, available in 1997 WL 4139560; see also Jeffrey St. Clair, U.S. Environment: Logging Practices Unleash Lethal Landslides, INTER PRESS SERV., Jan. 6, 1997, available in 1997 WL 7073076.

166. Letter of Arthur Bryant, supra note 161.

167. 33 U.S.C. § 1329, ELR STAT. FWPCA § 319.

168. Letter of Arthur Bryant, supra note 161.

169. Id.

170. See Houck, supra note 3, at 10337 (discussing the House committee report on TMDLs and nonpoint source pollution).

171. See SENATE COMM. ON ENVIRONMENT & PUBLIC WORKS, 100TH CONG., 2D SESS., A LEGISLATIVE HISTORY OF THE WATER QUALITY ACT OF 1987 (1987).

172. Dioxin/Organochlorine Ctr. v. Rasmussen, 37 ERC 1845, 1848 n.3 (W.D. Wash. 1993), aff'd sub nom. Dioxin/Organochlorine Ctr. v. Clarke, 57 F.3d 1517, 25 ELR 21258 (9th Cir. 1995).

173. For a discussion of the choice among several strategies under the Clean Air Act to control human exposure to lead, see FREDERICK R. ANDERSON ET AL., ENVIRONMENTAL PROTECTION: LAW AND POLICY 174-181 (2d ed. 1990).

174. See Houck, supra note 6 (describing CWA toxics regulation through separate programs based on human health standards, technology standards, water quality criteria, toxic water quality criteria, whole effluent testing, and biological criteria).

175. 33 U.S.C. § 1329, ELR STAT. FWPCA § 319; see Houck, supra note 3, at 10342.

176. Houck, supra note 3, at 10342.

177. EPA surveys estimate the cost of TMDL development from $ 4,039 to $ 1,023,531. EPA, TMDL DEVELOPMENT COST ESTIMATES: CASE STUDIES OF 14 TMDLs http://www.epa.gov/owow/tmdl/part23t.html. A representative of the Georgia Department of Natural Resources has reported TMDL development costs in that state from $ 75,000 to $ 5 million. Presentation of Alan Hallom, Georgia Department of Natural Resources, EPA Summary, FACA Committee Meeting 1 (Nov. 19-21, 1996).