Download PDF
May 1, 2013

NAS Committee Releases Report Recommending Changes in Process Used by EPA and the Services to Assess Risks to Endangered and Threatened Species from Pesticides

The ACTA Group

In a widely anticipated and potentially influential report released on April 30, 2013, a committee of the National Research Council of the National Academy of Sciences (NAS) made detailed recommendations concerning potential revisions to the process by which the U.S. Environmental Protection Agency (EPA) and the Fish and Wildlife Service or the National Marine Fisheries Service (the Services) assess risk during consultation under Section 7(a)(2) of the Endangered Species Act (ESA) for specific pesticide registration actions taken by EPA under the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA). EPA and the Services asked the NAS committee to provide recommendations on specific technical questions, as the agencies have approaches to risk assessment that “differ because their legal mandates, responsibilities, institutional cultures, and expertise differ,” and because the agencies “have been unsuccessful in reaching a consensus regarding their assessment approaches.” Although the NAS committee’s report is only advisory, EPA and the Services are likely to make a significant effort to review the recommendations and to implement those they deem feasible. The report, entitled “Assessing Risks to Endangered and Threatened Species from Pesticides,” is available online.


As requested by the agencies, the report contains detailed recommendations on certain constituent elements in the risk assessment process, including identification of the best available data, use of exposure modeling, use of geospatial and other geographic data to evaluate species habitat, identification of those effects that should be deemed relevant to jeopardy determinations, risks from mixtures, and the use of uncertainty factors. Nevertheless, because many view the current ESA consultation process for pesticide registration actions as poorly coordinated and the available agency resources for consultation to be overburdened, officials in the respective agencies as well as stakeholders in industry and non-governmental organizations (NGO) are likely to focus in particular on recommendations that either address the respective roles of the agencies in the risk assessment process, or that recommend changes in existing risk assessment methodologies or procedures practices that might alter the number of formal consultations or the scope of recommended mitigation measures. The NAS committee did not recommend any changes to the basic steps in the current consultation process, but the committee did suggest some significant changes in existing risk assessment methodologies and procedures.

NAS divides the process for an ecological risk assessment (ERA) undertaken by EPA and the Services for pesticide registration decisions into three basic steps. In Step 1, EPA determines whether use of a pesticide “may affect” endangered or threatened species. This question is interpreted broadly using screening level analysis, and the answer is typically affirmative for almost any pesticide applied outdoors. Based on a “may affect” determination, EPA may decide to initiate formal consultation immediately, or to proceed to Step 2, in which EPA decides whether the pesticide is “likely to adversely affect” endangered or threatened species. If EPA decides in Step 2 that the pesticide is not likely to adversely affect endangered or threatened species, and the relevant Service concurs, the consultation ends. Step 3 involves formal consultation, in which the Service determines whether the pesticide registration action is “likely to jeopardize the continued existence” of any endangered or threatened species.

In 2004, EPA and the Services issued joint “counterpart regulations” under which EPA would have been allowed to take responsibility for Step 3 in this process, but these regulations were partially invalidated during judicial review in 2006. In subsequent years, it has become standard practice for EPA to go directly from the broad screening determination in Step 1 to formal consultation in Step 3, and this shift of the responsibility for risk assessment to the Services has contributed to the burden on available resources and personnel. In its report, the NAS committee appears gently to criticize this practice, stating that “it might be more efficient in many cases to conduct a Step 2 analysis before deciding to enter formal consultation,” and that “Step 2 would filter out some actions, and fewer biological opinions would be needed.”

The NAS committee seems also to suggest a greater potential role for EPA in risk assessment through a key recommendation for greater continuity in the risk assessment process. The committee states: “If the Services can build on the EPA assessment conducted for Steps 1 and 2 rather than conducting a completely new analysis for Step 3, the ERA will likely be more effective and scientifically credible.” Of course, this recommendation presumes that current differences in risk assessment approaches between the agencies will be resolved in a manner that allows the recommended continuity to occur. In addition to specific technical recommendations, the NAS committee suggests that EPA and the Services can consult during “problem formulation,” a step where key effects and potentially affected species are identified, or that the agencies can work out the risk assessment methodologies to be used in “technical working groups.”

The NAS committee implicitly presumes that analysis of the potential effects on individual organisms can be completed by EPA (in consultation with the Services) in Step 2, and that the Services can then extrapolate these individual effects to population effects during Step 3, when EPA determines that adverse effects on individuals are likely and commences formal consultation. This presumption appears to be closely related to another key committee recommendation — to replace the current wide use of point estimates of risk based on effect or no effect levels combined with uncertainty factors and point estimates of exposure based on worst case assumptions with a “probabilistic approach” that integrates the available information by “using probability distributions rather than single point estimates for uncertain quantities.” Under this recommendation, risk estimates for pesticide use would take into account the actual likelihood of a particular adverse effect at each potential exposure level and the likelihood that each exposure level will actually occur. If such an approach could be implemented, probabilistic risk estimates could then be incorporated by the Services into probabilistic population models. In turn, the NAS committee believes that this would allow selection of mitigation measures by decision makers to be informed by an assessment of how probable it is that a particular pesticide use pattern will jeopardize the population of a species, rather than by jeopardy assessments based on fixed estimates and sometimes implausible worst case assumptions.


Many observers have been hoping that the NAS committee would provide recommendations for material reform of the current ESA consultation process for pesticides, which is widely regarded to be dysfunctional. The committee appears to have deftly avoided the underlying policy issues of whether or not the current framework and interaction between statutes is practical or in fundamental conflict. Of course, it is beyond the purview of any advisory committee convened at the request of administrative agencies to suggest legislative reforms, but the NAS committee not surprisingly accepted basic structural elements in the consultation process as it presently exists. This was predictable from the outset given the questions referred to the committee, but since the process began, many observers both supportive and critical of the current situation had hoped that this report might somehow essentially determine “who is right — the Services or EPA”? Nonetheless, if the NAS committee’s recommendations can be fully implemented, they could materially alter the current allocation of responsibility for risk assessment between EPA and the Services, and they could reduce both the number of formal consultations burdening the Services and the scope of the mitigation measures recommended by the Services to avert potential jeopardy. Before any of these things could happen, however, EPA and the Services must reach a consensus based on the recommendations, an issue that has been difficult, at best, to resolve to date and likely presents a formidable challenge moving forward.

The idea of greater continuity between initial risk analysis by EPA and a subsequent population assessment by the Services is appealing, but such continuity is not likely to be meaningful unless EPA is willing to undertake Step 2 more often, rather than defaulting directly from an amorphous Step 1 screening assessment to a Step 3 formal consultation. This in turn will require that EPA and the Services reach a consensus that resolves critical technical disagreements that caused them to seek the assistance of the NAS in the first place. The recommended coordination between EPA and the Services early in the process cannot be expected to resolve such disagreements on a case-by-case basis. The technical working groups recommended by the NAS committee could be critical in transforming general guidance on technical issues into risk assessment procedures that will allow the recommended continuity. It remains to be seen whether this will be practicable given the differing institutional cultures and statutory mandates of the respective agencies.

The idea of using probabilistic methods for ecological risk assessment has considerable promise as well, but only if the agencies can develop and agree on general methods. Developing and using probabilistic approaches for ERA has been a long-standing recommendation of other review panels such as EPA’s Scientific Advisory Panel. Probabilistic analysis could separate the most significant hazards from hazards that have been presumed to exist solely because of the use of default uncertainty factors and worst case assumptions. The report is perhaps its most biting in its criticism of EPA for the lack of such probabilistic tools, describing their availability and practicality in terms implying that EPA may have been remiss in its hesitation to adopt these readily available statistical and policy tools; the report, for example, refers to “probabilistic methods that… have otherwise appeared often in the technical literature, that are familiar to many risk-assessment practitioners, that can be implemented with commercially available software, and that are most readily explicable to decision-makers, stakeholders, and the public.” Apparently, the committee may believe, regarding the ease of introducing probabilistic assessments, that, effectively, “there is an app for that,” a belief that may not be altogether accurate. In theory, such methods could provide decision makers with a more accurate understanding of how likely adverse effects are both at the individual and population levels. Nevertheless, there will be significant impediments to the full implementation of this concept.

Development and adoption of probabilistic methods will take time. In the short term, rather than reducing the logjam of referrals for consultation, this could be a source of further delay. Once developed and validated, one can imagine the internal debate about the policy choice of where to “draw the line” as a definition of where to protect the probabilistically determined species and habitat: should 99.9% of the population be protected, is 90% sufficient, or is sufficient protection provided at 50% of the calculated distribution. The now resolved debate for dietary residues, using the “99.9 percentile,” is routinely misinterpreted in attempts to explain the meaning of this policy choice (a very mistaken but common assumption is that this means that EPA has chosen to not protect .1% of the population, over 300,000 people). For example, there is likely to be resistance from some stakeholders to using probabilistic analysis for endangered or threatened species based on the thesis that individual organisms should be protected. Even though the ESA allows for incidental takes to occur as long as they do not jeopardize the population, the loss of individuals can be significant if a species is sufficiently endangered. Some stakeholders can be expected to argue that the irreversible nature of extinction warrants use of the most risk adverse assumptions.

We will be watching with interest the response of EPA and the Services to these recommendations, and will follow up with further assessments as warranted.