On August 13, EPA finalized Clean Water Act (Act) Section 304(a) recommended criteria for phosphorus and nitrogen in lakes and reservoirs. The new recommendations incorporate scientific models that states and tribes with treatment as state status can use to establish numeric water quality criteria for phosphorus and nitrogen, and they mark an important milestone in EPA’s long-running war on excess nutrients in the nation’s surface waters. In the new recommended lakes criteria, EPA is embracing a stressor-response approach to managing nutrients, instead of the least-disturbed reference method. This is a significant move for the agency and has the potential to set a positive precedent going forward for EPA and state and tribal regulators to re-evaluate the utility of the reference method approach in other contexts.

Nutrients are necessary for a properly functioning aquatic ecosystem, but excess nutrients in the system can cause adverse effects. The stressor-response method focuses on the actual linkage between the concentration of nutrients in surface water and the ecological health of the water body. The goal is to identify a nutrient concentration that ensures attainment of the designated use of a water body, be it primary or secondary contact recreation or drinking water, where the introduction of nutrients beyond that ecological target is expected to cause adverse adjustments in the aquatic environment, such as an overgrowth of phytoplankton. By comparison, the least-disturbed reference method uses a statistical methodology to presume adverse ecological effects for a subset of waters within a particular region. The reference method involves identifying a subset of the healthiest water bodies (like lakes) in a designated area (such as an ecoregion or very large watershed), identifying the ambient nutrient concentrations in those waters, applying a margin of safety, and setting a numeric value above which adverse ecological responses are presumed for all waters in that region.

One problem with the least-disturbed reference method is that the triggering threshold with a margin of safety ends up classifying some of the healthiest waters in the area as “impaired.” For states trying to convince members of the public to implement expensive nutrient reduction strategies, passing out failing grades to the healthiest waters seriously undermines confidence in the regulators. Another problem is that the impairment above the numeric threshold is presumed without consideration of the actual response of the aquatic ecosystem to the nutrient concentration. In short, the stressor-response methodology is better science. The downside is that it usually involves more work to derive numeric criteria and generally requires more robust data sets. The new recommended criteria for lakes are designed to streamline some of that work, while promoting the development of more defensible and targeted criteria.

The use of scientific models that allow customization, as opposed to specific numeric criteria that EPA has typically established under Section 304(a), should also provide states and tribes with additional flexibility when developing water quality criteria for phosphorus and nitrogen. EPA traditionally has pressured states to establish numeric criteria for phosphorus and nitrogen, primarily because monitoring, conditions assessments, and Clean Water Act permitting decisions become more formulaic with numeric benchmarks. Although some states were early adopters of numeric criteria, others have faced challenges establishing single value, categorical numeric water quality criteria. This is, in part, because of the concerns associated with the real-world accuracy of the least-disturbed reference method described above. Other states continue to rely on numeric translations of narrative nutrient criteria to evaluate water quality and establish TMDLs, which presents its own challenges in implementation.

The Clean Water Act requires EPA to establish Section 304(a) recommendations based on the latest scientific knowledge. By contrast, states and tribes are required to promulgate water quality criteria that are based on sound scientific rationale and scientifically defensible methods and that are protective of the designated use. These are two very different legal and technical standards, and EPA’s announcement and final recommendation material do a nice job of explaining that states and tribes are not required to adopt Section 304(a) recommendations. Rather, states are free to adopt the recommended model, another version of the recommended model, or use an entirely different method to establish water quality criteria, as long as the method is scientifically defensible and the resulting criteria are protective of designated uses. In other words, states and tribes are provided flexibility to apply their water quality expertise, make policy choices, and establish priorities within their water quality standards programs. This flexibility and the deference owed to states and tribes is foundational to the co-administration of the Clean Water Act. EPA is responsible for reviewing and approving submittals of state water quality criteria, but at times, it has attempted to impose what it believes is the latest scientific knowledge rather than focusing on whether a state’s approach is scientifically defensible. This lack of deference to state expertise has contributed to programmatic problems for EPA, including the inability to timely act on state Clean Water Act submittals.

EPA’s announcement also explains — as it must — that the new recommended criteria do not “as a general matter” require a state or tribe to revise existing EPA-approved nutrient criteria. However, the agency notes that if a state or tribe chooses not to adopt the new recommendations as part of its triennial review of existing water quality standards, they are required to explain why.

Going forward, states and tribes that develop criteria using the recommended models should expect the biggest challenges to be in selecting, and defending their selection of, data used to run the model to establish water quality criteria, particularly if they customize the model runs. Most states likely have a large body of information and data on the presence of phosphorus and nitrogen in their local lakes and the corresponding ecological conditions of those waters; however, it is unlikely that all that information and data is appropriate or relevant for establishing water quality criteria. States must do their part to maintain a proper balance in the co-administration of the Clean Water Act, including using technically sound methodologies and data to earn the EPA deference that is owed under the Act. States and tribes should therefore select the data that most appropriately reflects their resources and program priorities, and if certain data or information is excluded, they should be sure to provide a technical rationale to explain why it was excluded. Similarly, states and tribes that make changes to the model or model implementation based on local information or use a different model or method, should be prepared to provide a technical basis to demonstrate that their approach is scientifically defensible and that the resulting criteria are protective of designated uses.

Finally, states that choose not to update existing water quality criteria during their next triennial review must develop a thoughtful rationale and supportive factual record. The recent occurrences and renewed awareness of harmful agal blooms, along with the Biden administration’s focus on climate change, are likely to bring heightened attention to existing lake nutrient criteria. This means EPA could work directly with individual states to encourage them to update existing criteria, or EPA could pursue more coercive techniques, potentially resulting in an “administrator’s determination” that the state must update its water quality criteria. It is also quite possible that environmental advocates may petition states or EPA in its oversight role to force states to adopt more stringent criteria. To be clear, using the new Section 304(a) recommendations do not guarantee the adoption of more stringent criteria, as the focus is and must be on what is necessary to attain established water quality standards, but the working assumption will invariably be that less is more when it comes to nutrient concentrations in the water column.