Total Maximum Daily Load Litigation: An Opportunity to Augment Instream Flows?

[This article was published in Rivers, Vol. 6, No. 2, 1997, by S.E.L. & Associates. 8 1997.]

Introduction

Instream flows provide a number of benefits to aquatic ecosystems including habitat for aquatic organisms, improved conditions for most riparian vegetation, and improvements in water quality. Stream dewatering degrades overall water quality and aquatic ecosystems as much as dumping garbage into a stream. The water quality provisions of the Clean Water Act, 33 U.S.C. ' 1303 (hereinafter, "' 303(d)"), provide a source of authority for states to implement minimum stream flows, where appropriate, in the context of allocating pollutant loads on water quality limited streams. The focus of this article will be the technical and political challenges faced by states implementing the water quality provisions of the CWA.

Section 303(d) of the Clean Water Act (hereinafter, "CWA") requires states to allocate pollutant loads to remedy streams that are not meeting water quality standards through the use of point source discharge permits. Despite, the plain language of ' 303(d), the water quality provisions of the Act were largely ignored until recently. During the first fifteen years of the Act's existence, from 1972 to approximately 1987, regulators focused on technology-based regulations and discharge permitting. However, a spate of recent litigation regarding the role of states and EPA in implementing the water quality provisions of the Act makes clear that federal courts do not share the states' and EPA's unwillingness to implement water quality provisions.

Ironically, the water quality provisions included in the CWA are a hangover from an earlier era when states were charged with setting and achieving interstate water quality standards. Under the 1948 and 1965 Water Quality Acts (ch. 758, tit. III, ' 303, 62 Stat. 1155 (codified as amended in scattered sections of 33 U.S.C.)) no federal enforcement action was ever brought for water quality violations and the minimal state efforts to set and enforce water quality limits met with little success. States were the lead actors and in effect, these earlier federal statutes provided states with little more than an encouraging nudge in the right direction. In amending the earlier Water Quality Acts in 1972 (resulting in what we know today as the Clean Water Act), Congress was aware of the failures of the earlier approaches. Nonetheless, Congress was also unwilling to rely solely on the federally mandated technology-based regulations that established discharge monitoring permits. Water quality standards were retained as a backstop. Section 303 became the centerpiece of the water quality provisions, incorporating much of the substance of earlier Water Quality Acts with more directive language. Despite the powerful tools ' 303(d) provided states to address intractable water quality problems, states have only recently begun to use ' 303(d) to address water quality problems, and then only under court order.

A brief background of the regulatory structure of the CWA facilitates an understanding of how States can use the water quality provisions to implement minimum stream flows. Part I provides this background. Part II discusses caselaw involving the water quality provisions of the CWA. Part III discusses the technical and political challenges states face in implementing TMDL provisions.

Part I

The foundation of the Clean Water Act is the classification of surface waters. States classify all surface waters, by segment, on the basis of past, present and potential future uses. State classifications are subject to EPA approval. After classification, states then set pollutant levels, often called "stream standards", for the amount of each pollutant each segment can accept and still support its uses, by reference to EPA's pollutant criteria levels. If a stream segment formerly hosted trout spawning, stream standards must be set to ensure future conditions that will support trout spawning. Under the Act, the state must set pollutant levels for all "pollutants" that impact the fishery including sediment and turbidity limits, and minimum stream flows, as improper levels of these (as well as whatever else the state determines it must regulate to bring back the trout fishery) impede trout spawning. As a practical matter, and in spite of the language of the Act, setting stream classifications is essentially a political exercise for states. States prefer less restrictive stream classifications because resource users along a stream find it easier to comply with the regulatory effects of less strict use classifications and the states would rather not fight about it. Further, the CWA's "antidegradation policy" precludes downgrading streams from higher uses to lower ones absent a lengthy administrative process, and subject to EPA veto.

The stream standard/use classification foundation supports the Act's two regulatory strategies. First, the Act requires a discharge permit to discharge pollutants from a point source (basically, the end of a pipe). EPA establishes industry specific discharge limits. The federal presence in effluent discharge limits is probably a main reason for the successes of the technology-based regulations. Battle and Lipeles, 1993. However, technology-based regulation does not remedy all water quality problems. In recognition of this point, Congress included in the Act water quality provisions to be applied to all surface waters for which technology-based regulation proved to be inadequate. Waters not meeting stream standards solely through technology-based regulations are termed water quality limited segments (WQLSs) and are subject to additional reductions in pollutants under the Act's Total Maximum Daily Load (TMDL) program. 33 U.S.C. ' 1303; CWA, ' 303(d).

The first step to implementing the TMDL provisions is to identify and prioritize water quality limited segments (WQLSs) of surface waters. States use a variety of methods to identify water quality limited segments. Colorado, for instance, developed its most recent draft list of WQLSs and sent them to interested parties prior to the formal notice and comment period. Thus, the state has the opportunity to correct any glaring omissions or excesses prior to formal comment.

The second step is to develop total maximum daily loads (TMDLs), which quantifies the total pollutant load a particular stream segment can handle and still meet the stream standards. Pollutant loads include an assessment of the wasteload allocation from point sources and the load allocation, attributable to nonpoint sources as well as natural background levels of the pollutant. The load allocation must include considerations of seasonal variations and margins of safety. Thus, if data are not available to make a reasonable assessment of the sources or amounts of pollutant loads, the state was charged with making a best guess--but a very protective best guess to improve the quality of the water. 33 U.S.C. 1303(d)(1)(C); CWA, ' 303(d)(1)(C). The first two steps present many political and fiscal hurdles for states. In addition to these practical problems, states are not subject to any federal standards in determining water quality limited streams or TMDLs. EPA must approve states' actions in steps one and two, but the fact remains that the statute allows fifty different water quality standards.

The third step, allocating pollutant loads amongst dischargers and nonpoint sources on a stream, is rife with more technical and political challenges and is taken up in greater detail in Part III. As a practical matter, only a handful of TMDLs have progressed this far and the process of actual pollutant allocation is somewhat speculative. Nonetheless, courts are forcing the issue on states in their decisions regarding environmental plaintiffs' challenges to inadequate implementation of ' 303.

Part II

Neither EPA nor the states have come to the water quality table willingly. The Clean Water Act was passed in 1972 and it was not until 1979, at the behest of the District of Columbia federal district court, that EPA issued regulations regarding which pollutants were subject to the TMDL requirements. (Board of County Commissioners 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)). States had 180 days in which to submit some of their water quality limited segments and some TMDLs; regulations on priority rankings would be established in the future and nonpoint sources did not need to be considered initially. 43 Fed. Reg. 60664, 60666. States were unenthusiastic. In 1981, environmental plaintiffs brought suit in federal district court for the failure of Illinois and Indiana to submit lists of water quality limited surface waters. EPA argued that as the language of ' 303(d) does not affirmatively impose a duty on the agency to act in the absence of state action, it need not act to implement TMDLs either. (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 [7th Cir. 1984]). The district court agreed, but the Seventh Circuit reversed, finding that no submission of TMDLs was a constructive submission of nothing and that EPA had to respond. If state submissions were inadequate (or zero), EPA had to act in the state's place. Scott was the first of a number of "constructive submissions" suits.

The second generation of TMDL suits involves environmental plaintiffs suing EPA and states for inadequate compliance with ' 303(d). Currently, approximately twenty-five lawsuits involving EPA's failure to correct states' inadequate implementation of ' 303(d) are ongoing or have concluded. In Idaho and Georgia, the respective federal district courts found the states' actions under ' 303(d) to be grossly inadequate. The Georgia court imposed a five year time limit for the conclusion of the entire TMDL process on all water quality limited stream segments in the state, including those water quality limited by nonpoint sources. (Sierra Club v. Hankinson, 939 F. Supp. 865 [N.D. Georgia, 1996]). Georgia had completed two TMDLs in the sixteen years since the statute required completion of TMDLs. At this rate, the court noted, "Georgia will take over one-hundred years to complete TMDLs for the 340 WQLS identified on the 1994 WQLS list." Id. at 871. As the state admitted that the 1994 list did not contain all WQLSs in the state, the court noted that the actual time could be far longer than one hundred years. Furthermore, the state's wasteload allocations for nonpoint sources were not TMDLs because they did not provide for "daily loads" nor did they take into account seasonal and daily variations. The federal court, ruled that Georgia had five years in which to comply with the water quality provisions of the Act. The court retained jurisdiction and scheduled annual reporting dates. The state's failure to comply would result in EPA revising Georgia's NPDES Program to require it. Continuing noncompliance would result in EPA withdrawing its certification of the state program.

The Idaho court in (Idaho Sportsmen's Coalition v. Browner, 951 F. Supp. 962 [W.D. Wash. 1996]), allowed the state and EPA a "reasonable" amount of time to complete TMDLs on the 962 water quality limited segments in the state, but rejected EPA'S proposed twenty-five year time limit. The court found two legal flaws in EPA's proposed schedule. First, "extreme slowness". Browner at 966. The court noted the role of TMDL development to impose pollution control measures where technology-based point source regulation is inadequate. The court declined to allow EPA to create a process, which the statutory language required to be completed in 1973, which could be completed in twenty-five, fifty, or seventy-five years.

The second flaw noted by the court was the failure of the schedule to provide for TMDL development for all WQLSs. EPA and the state had assumed that the 962 segment identified was an overestimation, and that monitoring and evaluation would reduce it. The court rejected the EPA's attempts to "treat the hoped-for results of state programs as a substitute for CWA compliance". Id. The court remanded the case to EPA for a filing within six months of a reasonable schedule for developing TMDLs for all WQLS waterbodies in the state. "Nothing in the law could justify so glacial a pace." Browner at 967.

The cases demonstrate the characteristic ping-pong flavor of TMDL litigation. Plaintiffs file in federal district court; EPA stonewalls; the court grants plaintiffs summary judgment on their first motion; EPA drags its feet and grudgingly complies; plaintiffs go back to court, etc. Ongoing TMDL lawsuits in Colorado and Wyoming present similar facts including: state failures to identify all WQLSs, EPA failures to approve or disapprove state submissions in a timely fashion, "TMDLs" submitted by the states and approved by EPA that fail, in derogation of the regulatory definition, to allocate pollutant load between point and nonpoint sources, and "TMDLs" submitted by states and approved by EPA that fail to include margins of safety and seasonal variation levels of the pollutant allocation. (See Complaint and Answer, Colorado Environmental Coalition v. Browner, Case No. 97-S-1841 (D. Colo., August 25, 1997); Complaint and Plaintiff's Brief in Support of Summary Judgment, Wyoming Outdoor Council v. Browner, 96-S-2831, (D. Wyo., December 11, 1996)). Judging from the facts alone, EPA will likely receive another court order to step in and correct Colorado's and Wyoming's inadequacies in implementing the TMDL provisions. Eventually, even EPA will have to admit it has a duty to implement ' 303(d) in the absence or inadequacy of state action. This will undoubtedly spawn a third generation of TMDL litigation which will involve environmental plaintiffs suing states and EPA for failure to comply with courts' orders or for failure to allocate wasteloads under TMDLs or for arbitrary and capricious approval of inadequate TMDLs. However, in the lull before that particular legal storm, state's authority to undertake the technically and politically challenging task of implementing TMDLs can be examined.

Part III

The lesson of the TMDL litigation is that, despite the arguments of EPA, federal courts will enforce CWA provisions requiring EPA, in the absence of adequate state action, to implement TMDL provisions. States should address their TMDL problems not only as a matter of law, but also as a matter of protecting their water resources. But states need not await a court order. Under the CWA, states have ample authority to describe and impose TMDLs on the regulated community. Further, states can use their authority, where appropriate, to achieve additional minimum stream flows.

While states ought to be embracing their authority to correct water quality problems within their borders, most states lack the most elementary tools to begin to develop TMDLs. Few states have conducted adequate and thorough water quality data collection. Part of the reason is likely historical: the CWA allowed states to continue to use the stream standards established under the earlier 1965 Water Quality Act. Thus, states had little incentive to begin any comprehensive data collection. Another reason for the lack of data is fiscal. Few state water resources departments have adequate personnel and funding to undertake massive water quality data collection. Wyoming, for instance, has planned a three to seven year data collection effort prior to the identification of water quality limited stream segments. Skeptics may view these efforts as stonewalling and foot-dragging on the part of states, but proper allocation of pollutant loads relies on good science and adequate data. Further, to the extent that state decisions regarding TMDL allocation are immune to court challenges, those concerned with actual improvements in water quality should insist states undertake comprehensive and thorough data collection.

The integrity of the data upon which TMDLs are based is critical because the remainder of TMDL implementation involves state decisions susceptible to political maneuvering. After pollutant sources and quantities are determined, the state must prioritize the water quality limited segments to reflect which segments will receive TMDL allocations first. Then, the state must assign reductions in pollutant discharges to each source on a stream. Finally, states must choose to either impose wasteload allocations or to negotiate amongst pollutant sources the allocation of pollutant loads. All of these decisions are likely to be driven, at least in part, by the mix of pollutant sources on the stream. Despite political considerations, ' 303(d) in tandem with the state's authority to issue discharge permits, provides adequate authority for states to include TMDLs in permits for point source dischargers. Section 303(d) by its terms, provides states with authority to negotiate TMDLs with nonpoint sources and, under the state's ' 401 certification authority, to impose pollutant reductions on dischargers that conduct activities that require a federal permit.

Point sources are an obvious target for states allocating TMDLs. The discharger's NPDES permit reflects his privilege to discharge into a surface water. The state can condition repermitting on the point source further reducing his discharges. As the discharger's permit is his only legal means of discharging in the first place, the state has considerable power over the discharger in allocating further reductions, as long as the additional reductions are not arbitrary. If a stream is water quality limited on the basis of point source discharges (a determination that itself may be politically driven), a TMDL may have the majority of the pollutant reductions allocated to point sources. But focusing on point source dischargers, though attractive is entirely inconsistent with ' 303(d). After all, what triggers a state's duty under ' 303(d) is waters for which discharge permitting is insufficient to achieve compliance with water quality standards. Further, EPA regulations require the pollutant allocation to be between point and nonpoint sources, as well as the recognition of background levels of the pollutant that may be present. Point sources may need to make additional reductions in pollutant discharge, but the focus of TMDLs is to change the discharges of nonpoint sources into the stream.

This is one reason for plaintiffs' challenge to EPA's approval of Wyoming's TMDLs. Many of Wyoming's TMDLs are merely discharge permit limits reformulated into TMDL language. They contain neither nonpoint source nor natural background allocations and the allocations are made on the basis of the discharge permitted 30-day period, rather than the daily allocations required under the CWA. Despite these legal deficiencies, point sources are an attractive target for states. In addition to a state's existing regulatory relationship with point sources through the NPDES permit, nonpoint sources are associated with constituencies, such as agriculture, grazing, timber and construction, that have rarely been held accountable for water quality problems. Estimates of nonpoint source contributions to water pollution are between fifty and eighty-five percent of all water pollution, and seventy-five percent of all water quality limited waters. (Houck, 1997; Wyoming 305(b) Report to EPA on the Status of State Waters, 1996). Despite these numbers, states are unenthusiastic about taking on such politically powerful groups, particularly in the west.

This political reluctance is bolstered by the lack of explicit nonpoint source regulatory language in the CWA. The two sections that deal explicitly with nonpoint sources, Sections 208 and 319 provide non-mandatory measures to be taken by states to characterize nonpoint source problems on a watershed-wide basis (33 U.S.C. ' 1288, ' 1329; CWA ' 208, ' 319), and provides federal dollars to remedy some of the problems (33 U.S.C. ' 1329; CWA ' 319). Thus, while the language of ' 303(d) does not qualify its directive to clean up WQLSs on the basis of the pollution source, the descriptive rather than remedial nature of '' 208 and 319 have been the basis of EPA and state arguments that nonpoint sources are not covered by ' 303(d). As discussed above, courts have disagreed with this reading of the provision (see, e.g., Sierra Club v. Hankinson, 939 F. Supp. 865 [N.D. Georgia, 1996]), and only through regulation of nonpoint sources can water quality limited segments reach mandated stream standards, if effluent limitations on point sources are inadequate regulation. However, as a political matter, state allocation of pollutant loads between point and nonpoint sources will likely favor the nonpoint source.

The challenge of allocating pollutant load amongst nonpoint sources is unavoidable if the stream segment is water quality limited solely on the basis of nonpoint sources. Again, the TMDL must include point and nonpoint source components, as well as the natural background levels of a limiting pollutant. But some western streams, in particular, have no point source discharges at all; thus, as a practical matter, a TMDL that complied with the regulatory definition would have daily allocations of pollutant loads amongst nonpoint sources and zeroes for daily allocations amongst point sources. In addressing nonpoint source problems, states must decide, in allocating pollutants to nonpoint sources, whether they will use carrots or sticks. The carrot approach is a negotiated agreement, whereby water users agree to modify their management practices, including water use, to accommodate pollutant allocations. States might use federal monies at their disposal through ' 319 of the Act as additional persuasion for reluctant water users. The state might also use a "stick": withholding CWA ' 401 certification to prevent a landowner from receiving a necessary federal permits, such as grazing permit unless he agrees to practices that will reduce pollutant levels.

The Deep Creek TMDL, developed under the auspices of the Montana Department of Environmental Quality, is an example of a negotiated solution. Deep Creek is located south of Townsend, Montana, and is a major tributary of the Missouri River. Deep Creek is one of the few spawning stream available for salmonids in between Toston Dam and Canyon Ferry Reservoir. The upper reaches of the creek supports a resident trout fishery. The entire creek was designated water quality limited because of its heavy sediment load, assumed high turbidity levels, and degraded habitat. As the focus of an effort to improve the trout fishery in this region of the Upper Missouri, the DEQ chose Deep Creek for its first TMDL effort.

The Deep Creek TMDL demonstrates the challenges facing all states that are only now beginning to address ' 303(d): little was known about the water quality conditions of the stream. Practically, Deep Creek was known to play an important role in the trout fishery of the Upper Missouri, and the creek was widely believed, based on visual examinations and the opinions of those living nearby, to be water quality limited. Biologists from Montana State University, in conjunction with DEQ and Montana Fish and Game scientists began to collect data on the relevant parameters impacting the trout fishery. They obtained aerial photos of the entire stream, made a stream bank inventory, collected water quality data, substrate cores, and surveyed fish and macroinvertebrate communities. As Montana does not have numeric stream standards for sediment load or concentration, scientists were forced to examine physical parameters, including cobble embeddedness (meaning how easily a female fish could push the bottom of the stream around to create a nest in which to deposit her eggs), and surface fines (meaning the size of the sediment particles); very small particles adhering to the bottom of the stream make oxygen exchange difficult and result in poor trout egg survival. The Natural Resources Conservation Service then performed a watershed inventory in 1991, describing in broad terms the geology, topography, soils and land uses of the watershed. The state had previously divided the Creek into 11 "reaches", or segments divided by physical characteristics or property lines. (Endicott and McMahon, 1996.)

What three years of study and a year of analysis revealed was a creek that required physical intervention to correct the deterioration of the riparian and aquatic ecosystems. While the CWA does not authorize states to correct all riparian problems, to the extent that riparian deficits harm aquatic communities, they fall within the statute. In the case of Deep Creek, that meant providing habitat for spawning beds through streambank manipulations and re-establishment of lost channel length. It also meant changes in grazing practices, to eliminate some of the fine sediments currently being deposited in the creek. As fish survival is poor without sufficient water in the stream, the TMDL requires minimum stream flows. However, none of this would have been possible without the cooperation of the adjacent landowners.

Montana gathered its "stakeholders", those in the local community who as adjacent landowners were impacted by management decisions regarding Deep Creek or those who, while not adjacent landowners, derived some enjoyment of the stream and wanted to be a part of the negotiated process to achieve improved water quality. A series of meetings were held over the course of a year to achieve consensus on the "allocation" of the TMDL. The results were spectacular--on paper at least. Local irrigators agreed to modify their water uses to keep water in the stream: a minimum of 3 cfs at the lower reaches and 9 cfs in the upper reaches. Two years of monitoring reflected a success at this voluntary instream flow. Further, a number of the ranchers on the stream agreed to modify their grazing practices, including fencing out cattle along a number of properties, and pumping water to separate tanks so that cattle were not in the stream itself to drink. These changes alone will improve temperature conditions for trout and reduce the amount of sediment introduced into the stream. In addition to physical changes to the stream banks and the addition of substrate to improve riparian communities (and subsequently, trout habitat), these management practice changes will likely greatly improve the water quality of Deep Creek.

Montana's success with the Deep Creek TMDL is commendable. However, in terms of the state's continued compliance with federal law, specifically ' 303(d), the voluntary cooperation of adjacent landowners provides little reason for the EPA, DEQ, or potential environmental plaintiffs to turn their attention to other things. Ranches and farms are sold or change hands over time; each time local land ownership changes, the likelihood is that these changes in management practices will not continue. The language of ' 303(d) is clear: states must remedy water quality limited stream segments, and need not do it through negotiation.

States' ' 401 certification authority provides an expedient means of imposing TMDLs. As with ' 303(d), courts have only recently begun to interpret state powers under ' 401, so the depths of its broad language have not fully been plumbed. However, the United States Supreme Court's decision in PUD No. 1 v. Washington Department of Ecology, 114 S. Ct. 1900 (1994), did nothing to restrict the reach of section 401. PUD No. 1 involved the state's conditioning of a federal permit for a hydroelectric project on minimum instream flows. The Court agreed that the state could so condition a permit under its ' 401 authority. Further, the Court held that states could condition issuance of permits on compliance with state water quality standards, including stream standards.

Section 401 provides states authority to "certify" that activities conducted under federal permits will comply with their water quality requirements. State conditions for certification become conditions of the federal permit; states may veto federal permits entirely by declining to certify the activity. The point of section 401 is to prevent states from being out of compliance with their water quality programs--and in trouble with EPA--due to activities conducted on land or under permitted authority over which they have no jurisdiction. Until recently, state authority was thought to restrict the issuance of only five types of federal permits: dredge and fill permits under ' 404, federal point source discharges under ' 402, permits for activities impacting navigation under the Rivers and Harbors Act, '' 9 and 10, and licenses required to build dams under the Federal Power Act.

However, the statutory language is broad: "any activity that may result in any discharge." CWA ' 401. Commentators have suggested, and the PUD No. 1 decision lends support to their arguments, that any federal permit may be subject to state certification under ' 401. (Donahue, 1996; Ransel and Meyers (1995)). Federal permits that are subject to ' 401 certification likely include: federal grazing permits, Bureau of Land Management special land use permits, permits to conduct activities in the National Forests, timber contracts, contracts to use federal irrigation water, and incidental take permits under the Endangered Species Act. (Donahue, 1996). State authority to restrict federal activities, as well as private activities on federal land such as grazing and logging, would likely survive a court challenge. Thus, for nonpoint sources subject to federal permits, states could also use the ' 401 certification stick to allocate nonpoint pollutants, including minimum stream flows.

In the long run, states need not pick one approach over another. Section 303 by its terms creates a regulatory atmosphere dominated by state decision-making. Skillful negotiation and "stakeholders" who are present at the table absent a court order will likely result in many allocations of nonpoint source TMDLs. Further, as between the state and the EPA, local resource users would probably prefer to deal with the state. However, given the timetable imposed by some courts (five years in Hankinson), ' 401 may provide a better vehicle to allocate pollutants amongst nonpoint sources because it is less time-consuming to draft certification conditions than to negotiate allocation of stream-wide TMDLs. However, states should recognize their duty to act to implement TMDLs and should do so fairly. Despite the easy target provided by point source dischargers, nonpoint sources should not be immune to the effects of ' 303(d), particularly as water quality limited streams are classified because the effects of discharge permitting are insufficient to clean up the stream.

REFERENCES

Battle, J.B. and M.I. Lipeles. 1993. 2 Environmental Law: Water Pollution. Cincinnati: Anderson Publishing.

Houck, O. 1997. TMDLs: The Resurrection of Water Quality Standards-Based Regulation Under the Clean Water Act. Environmental Law Reporter (ELR) 27:10329-10344.

Endicott, C. and T. McMahon. 1996. Development of a TMDL to Reduce Nonpoint Source Sediment Pollution of Deep Creek, Montana. Missoula: Montana Department of Environmental Quality.

Donahue, D. 1996. The Untapped Power of Clean Water Act Section 401. Ecology Law Quarterly 101: 201-301.

Ransel, K and E. Meyers. 1995. The Sleeping Giant Awakens: PUD No. 1 of Jefferson County v. Washington Department of Ecology. Environmental Law 25:255-283.