Technical Overview of Ecological Risk Assessment: Risk Characterization

About Risk Characterization

Risk characterization is the final phase of the ecological risk assessment. The risk characterization:

• integrates the analyses from the exposure characterization and ecological effects characterization;

• describes the uncertainties, assumptions, and strengths and limitations of the analyses; and

• synthesizes the overall conclusion about risk that is used by risk managers in making risk management decisions.

Risk characterization has two major components: risk estimation and risk description. Risk estimation compares exposure and effects data, considers integrated exposure and effects data in context of Levels of Concern (LOCs), and states the potential for risk. The risk description interprets risks based on assessment endpoints. In interpreting the risk, the risk assessor evaluates the lines of evidence supporting or refuting risk estimates in terms of the following factors:

• Adequacy and quality of data

• Degree and type of uncertainty

• Relationship of evidence to risk assessment questions

For a risk characterization to be useful to risk managers, it must be transparent, clear, consistent, and reasonable (the TCCR principles). Once the risk characterization is finalized, it may be used as the basis for producing fact sheets, press releases, technical briefings, and other communication products.

Deterministic Approach

Calculation of Risk Quotients

For most risk assessments, EPA uses a deterministic approach or the quotient method to compare toxicity to environmental exposure. In the deterministic approach, a risk quotient (RQ) is calculated by dividing a point estimate of exposure by a point estimate of effects. This ratio is a simple, screening-level estimate that identifies high- or low-risk situations.

Calculation of risk quotients are based upon ecological effects data, pesticide use data, fate and transport data, and estimates of exposure to the pesticide. In this method, the estimated environmental concentration (EEC) is compared to an effect level, such as an LC₅₀ (the concentration of a pesticide where 50% of the organisms die.)

RISK QUOTIENT = EXPOSURE / TOXICITY

Terrestrial Animals

For screening-level risk assessments, the following toxicity endpoints are routinely used as inputs to the Risk Quotient (RQ) method for expressing risk to terrestrial animals:

Other toxicological endpoints may be used in the risk characterization if they can be linked to assessment endpoints in a reasonable and plausible manner.

Using the model T-REX, EPA calculates acute and chronic avian and mammalian risk quotients (RQs) for spray, granular, and seed treatment applications of pesticides.

Spray applications

For spray applications, T-REX calculates two types of risk quotients:

• dietary-based RQs

• dose-based RQs

Acute and chronic dietary risk quotients are calculated using the following formulas:

• Acute dietary RQ = EEC/LD₅₀

• Chronic dietary RQ = EEC/NOAEL

Dietary risk quotients for spray applications are calculated by directly comparing the concentration of an administered pesticide (or estimated to be administered) to the LC₅₀ or NOAEL (mg/kg-diet) values from acceptable or supplemental toxicity studies. These risk quotients do not account for the fact that smaller sized animals need to consume more food relative to their body weight than larger animals or that differential amounts of food are consumed depending on the water content and nutritive value of the food.

Dose-based risk quotients, on the other hand, do account for these factors. Acute and chronic dose-based risk quotients for spray applications are calculated using the following formulas:

• Acute dose-based RQ = Ingestion rate-adjusted EEC / weight class-scaled¹ LD₅₀

• Chronic dose-based RQ = Ingestion rate-adjusted EEC / weight class-scaled NOAEL

Dose-based risk quotients incorporate the ingestion rate-adjusted exposure from the various food items for the different weight classes of birds and mammals and the weight class-scaled toxicity endpoints. The dose-based EECs are compared with the LD₅₀ or NOAEL values from acceptable or supplemental toxicity studies that are adjusted for the size of the tested animal compared with the size of the animal being assessed.

In dose-based risk quotients, the exposure values are presented as mass of pesticide consumed per kilogram body weight of the animal being assessed (mg/kg-bw). EECs and toxicity values are relative to the animal's body weight since consumption of the same mass of pesticide residue results in a higher body burden in smaller animals compared with larger animals. For birds, only acute values (LD₅₀s) are adjusted since dose-based risk quotients are not calculated for the chronic risk estimation. An acute dose-based RQ is derived from 20-, 100-, and 1000-gram birds based on allometric equations for the LD₅₀ value. Adjusted mammalian LD₅₀s and reproduction NOAELs (mg/kg-bw) are used to calculate dose-based acute and chronic risk quotients for 15-, 35-, and 1000-gram mammals.

¹ Weight class-scaled = LD₅₀ times the body weight of the animal

Granular applications

For granular and liquid banded applications and granular and liquid broadcast applications, T-REX uses the LD₅₀/ft² method to calculate RQs.

• Acute dietary RQ = (mg a.i./ft²) / LD₅₀

• Chronic dietary RQ = (mg a.i./ft²) / NOAEL

• Acute dosed-based RQ = (ingestion rate-adjusted mg a.i./ft²) / weight class-scaled¹ LD₅₀

• Chronic dosed-based RQ = (ingestion rate-adjusted mg a.i./ft²) / weight class-scaled NOAEL

With this method, acute RQs are based on a toxicity (LD₅₀) and exposure (milligram active ingredient/ft²) value. As with spray application, the dosed-based RQ incorporates the ingestion rate-adjusted exposure from the various food items for the different weight classes of birds and mammals and the weight class-scaled toxicity endpoints.

Seed Treatment

For seed treatment, T-REX calculates two acute RQs and one chronic RQ for birds and mammals.

• Acute RQ = (mg a.i./ft²) / weight class-scaled LD₅₀

• Acute RQ = (mg a.i./body weight/day) / LD₅₀

• Chronic RQ = (mg a.i./kg seed) / NOAEC

The first acute RQ for seed treatment is calculated by dividing the EEC (mg a.i./ft²) by the LD₅₀ adjusted for body weight of the animal. A second RQ is calculated by dividing the EEC (mg a.i./body weight/day) by an unmodified LD₅₀. The chronic RQ is based on the EEC (mg a.i./kg seed) divided by the NOAEC value.

To determine the confidence of risk conclusions that are based on RQ calculations for pesticide treated seeds, EPA/OPP has developed a series of risk characterization steps that can be found at: Refinements for Risk Assessment of Pesticide Treated Seeds – Interim Guidance.

Additional information concerning the model T-REX can be found on OPP's Models for Pesticide Risk Assessment website.

Aquatic Animals

For screening-level risk assessments, the following toxicity endpoints are routinely used as inputs to the Risk Quotient (RQ) method for expressing risk to aquatic animals:

Other toxicological endpoints may be used in the risk characterization if they can be linked to assessment endpoints in a reasonable and plausible manner.

Acute risk quotients in fish and invertebrates are calculated using the following formula:

• Acute RQ_{fish and invertebrates} = Peak water concentration / Most sensitive organism LC₅₀ or EC₅₀

Chronic risk quotients in invertebrates are calculated using the following formula:

• Chronic RQ_{invertebrates} = 21-day average water concentration / Aquatic invertebrate chronic toxicity NOAEC

Chronic risk quotients in fish are calculated using the following formula:

• Chronic RQ_fish = 56-day or 60-day average water concentration / Fish early life-stage or full life-cycle toxicity NOAEC

Plants

Terrestrial Plants

EPA uses the model TerrPlant to calculate risk quotients for non-target monocots and dicots (terrestrial plants) inhabiting dry and semi-aquatic areas.

For terrestrial non-endangered plants, TerrPlant uses the lowest EC₂₅ values from seedling emergence and vegetative vigor studies for both monocots and dicots.

For terrestrial endangered plants, TerrPlant uses the NOAEC or EC₀₅ associated with the lowest EC₂₅ from seedling emergence and vegetative vigor studies for both monocots and dicots.

Other toxicological endpoints may be used in the risk characterization if they can be linked to assessment endpoints in a reasonable and plausible manner.

For runoff to dry and semi-aquatic areas, TerrPlant, compares the combined deposition estimates from runoff and spray drift to effects or toxicity levels measured in seedling emergence studies.

• Acute RQ for runoff to dry and semi-aquatic areas = EECs from runoff + spray drift / EC₂₅ seedling emergence study

TerrPlant also calculates risk quotients for spray drift alone and compares the spray drift deposition value to the more sensitive measure of effect, either seedling emergence or vegetative vigor.

• Acute RQ for spray drift = EEC from spray drift / EC₂₅ in seedling emergence or vegetative vigor study

The selection of the more sensitive measure of effect is automated by the model and is defined as the lowest EC₂₅ (effective concentration that affects 25% of the exposed plants). The associated NOAEC value is selected to define RQ values for listed species exposed to drift. In cases where the EC₂₅ values for seedling emergence and vegetative vigor are equal, the lowest NOAEC value is selected for calculating the RQ value.

• Listed RQ = EEC/NOAEC

The results of these calculations are RQ values for non-listed and listed monocots and dicots inhabiting adjacent, dry, and semi-aquatic areas and RQ values for plants exposed to drift only. Additional information concerning TerrPlant can be found on OPP's Models for Pesticide Risk Assessment website.

Aquatic Plants

For aquatic vascular plants and algae, the screening-level risk quotient is routinely based on the lowest EC₅₀ for vascular plants and algae. Other toxicological endpoints may be used in the risk characterization if they can be linked to assessment endpoints in a reasonable and plausible manner.

The following formula is used to calculate the risk quotient for non-listed aquatic plants:

• RQ = EEC / EC₅₀

For listed² aquatic plants, the following formula is used in calculating the RQ:

• RQ = EEC / NOAEC

² Listed refers to threatened or endangered species.

For multi-use risk assessments, EPA scientists may calculate multiple RQ values for each labeled rate, method of application, interval, or formula. This allows the risk managers to see what uses offer the highest risk potential and those that do not offer high risk potential.

Comparison of Risk Quotients to Levels of Concern

After the risk quotient(s) is calculated, it is compared to the Agency's Level of Concern (LOC). An LOC is a policy tool that the Agency uses to interpret the risk quotient and to analyze potential risk to non-target organisms and the need to consider regulatory action. Several ecological LOCs, which are used in regulatory decision-making, are listed below:

Risk Presumptions for Terrestrial Animals

Risk Presumptions for Aquatic Animals

Risk Presumptions for Plants

Probabilistic Approach

Although the deterministic or risk quotient method is useful for screening purposes, it provides only one point estimate of environmental risk. Following the recommendations of the FIFRA Scientific Advisory Panel and the Ecological Committee on FIFRA Risk Assessment Methods (ECOFRAM), EPA has moved towards developing more sophisticated methodologies for conducting probabilistic or refined ecological risk assessments.

• In a refined risk assessment, EPA uses probabilistic tools and methods to estimate the variability and/or uncertainty in factors that influence risk and to express risk in terms of the probability and magnitude of adverse effects.

• A refined or probabilistic risk assessment produces a distribution or range of values instead of one fixed value. The assessment becomes more refined as the levels increase, presenting increasingly focused effects and exposure scenarios. Because the results of the refined risk assessment show the range of possible environmental impacts and which ones are most likely to occur, they provide the risk manager with a flexible tool for making decisions regarding the level of certainty needed for a particular situation.

EPA's draft probabilistic ecological methodologies and approach were presented at several scientific meetings and workshops. Based on the input from the SAP, ECOFRAM, and scientific meetings, EPA developed refined (Level II) terrestrial (TIM, MCnest) and aquatic models for conducting probabilistic ecological risk assessments for pesticides.

Endangered Species Effects Determinations

Under the Endangered Species Act (ESA), all federal agencies are required to ensure that their regulatory actions are not likely to jeopardize the continued existence of threatened or endangered species (listed species) or destroy or adversely modify their critical habitat. The ESA is administered by the U.S. Fish and Wildlife Service and NOAA Fisheries Service (together referred to as the Services).

When registering or reassessing the potential ecological risks from use of a currently registered pesticide, EPA uses risk assessment procedures to determine whether individuals of a listed species have the potential to be harmed by a pesticide, and if so, what specific protections may be appropriate. These risk assessment procedures can be found in the Overview of the Ecological Risk Assessment Process.

EPA's conclusion regarding the potential risks a pesticide may pose to a listed species and any critical habitat for the species results in an "effects determination." The effects determination can result in three possible conclusions:

• No effect on the listed species or critical habitat,

• May affect but is not likely to adversely affect the listed species or critical habitat, or

• May adversely affect the species or critical habitat

If the screening-level risk assessment indicates that there are no indirect effects and the levels of concern for listed species are not exceeded for direct effects, then EPA declares there is "no effect" from that pesticide's use on listed species and critical habitat. If, on the other hand, indirect effects are anticipated or exposure may exceed the levels of concern (LOCs) for direct effects, the pesticide's use is usually declared as "may affect" the particular listed species or critical habitat.

In the case where the screening-level risk assessment indicates a pesticide may potentially impact, either directly or indirectly, listed species or critical habitat, EPA performs a more refined species-specific and/or habitat-specific assessment. This refined assessment helps characterize the potential for exposure at the predicted levels and distinguish those actions that "may affect but are not likely to adversely affect" a particular species or habitat from those actions that appear "likely to adversely affect" a listed species or critical habitat.

Guidance for defining the potential area in which a species is likely to be adversely affected or critical habitat may be destroyed or adversely modified can be found at: Guidance for Defining the Potential Area In Which a Species Is Likely to Be Adversely Affected or Critical Habitat May Be Destroyed or Adversely Modified. Search EPA Archive

A variety of information and data sources are used to develop more refined species-specific and habitat-specific assessments:

• One source of information is the DANGER program, a computerized database that contains county-level occurrence information for listed species and county-level information on agricultural crops and their acreage.

• Another source of information for determining the distance between the locations of listed species and the types of land on which pesticides may be used is the FESTF database. FESTF, which is being developed by an industry task force, can be used to generate proximity analyses for listed species effects determinations. Guidance for use of FESTF analyses can be found at the following website: Future Use: Use of FESTF-generated Proximity Analyses for Listed Species Effects Determination Search EPA Archive

• Biological requirements and habits of listed species can be obtained from the Services' species' listing rules and recovery plans, status review background documents, species profiles, and benchmark studies. In addition, information on the biological requirements and habits of listed species can be found in the published literature and in other sources such as The Official World Wildlife Fund Guide to Endangered Species of North America.

• Information on species status and environmental baseline is also an important component of the assessments. The environmental baseline includes:

  1. past and present impacts of all actions and activities in the action area;

  2. anticipated impacts of proposed Federal projects in the action area; and

  3. the impact of actions that are contemporaneous with the consultation process with the Services.

  Guidance for developing the species status and environmental baseline for national-scale pesticide listed species assessments can be found at the following website: Information for Species Status and Environmental Baseline for National-Scale Pesticide Listed Species Assessments.

• Incident information from EPA's Ecological Incident Information System (EIIS) and from the Ecological Pesticide Incident Reporting Portal (http://npic.orst.edu/eco/) may be used to assist in characterizing the potential risk of a pesticide to a listed species or critical habitat. Guidance for using incident data in evaluating listed and non-listed animal species can be found at the following website: Guidance for Using Incident Data in Evaluating Listed and Non-listed Species under Registration Review

• Local use practices, sales and use information, and sub-county commodity information can also be used to inform the risk characterization.

• Information on geographic features in a county can be used to determine whether a pesticide use may occur in proximity to a listed species or its critical habitat.

• Monitoring data can be used to better characterize the potential risk to the listed species. An example of monitoring data is the data generated by the USGS National Water Quality Assessment Program (http://water.usgs.gov/nawqa/).

The information from all these sources can be used to determine whether spatial and temporal overlap in pesticide use and specific activities and habits may result in exposure at a level and duration that produces the effect. In addition, information from the above sources can help EPA evaluate the potential exposure levels and nature and magnitude of effects to listed species and critical habitat.

Previous Section: Analysis Phase - Exposure Characterization