U.S. Perspective: Probabilistic Risk Assessments and Their Use in Risk Management Decisions
Sunzenauer, I.M., Fite Edward, Gallagher Kathryn, Odenkirchen Edward, Irene Stephanie, Urban Douglas, and James Lin.
On this Page
- Presentation Overview
- Why Revise the Ecological Assessment Process for Pesticides?
- What Are The Questions?
- How Did We Go About It?
- What Does This New Assessment Scheme Look Like?
- Where Are We Now and What Do the Models Look Like?
- Highlights of Aquatic Model
- Example of Aquatic Results
- Highlights of the Terrestrial Model
- Example of Terrestrial Results
- How Are Risk Managers Involved?
- What Did the Regulators Say?
- Comments Related to Decision-making
- Next Steps
- EPA Implementation Team
Why revise the assessment process for pesticides?
How did we go about it?
Highlights of aquatic and terrestrial models
Use in risk management decisions
Why Revise the Ecological Assessment Process for Pesticides?
Provide Agency regulators
answers to risk assessment and risk management questions which could not be addressed in the past, and
improved and expanded risk characterizations, focusing on magnitude, probability and certainty of predictions.
What Are The Questions?
What are the magnitude and probability of effects and how certain are you?
Are these effects seen across different species?
Are there population or community impacts? Will the effects influence the density and diversity of the species?
How Did We Go About It?
Began initiative by forming stakeholder workgroup (1997)
Held peer review workshops to review stakeholder recommendations (1999)
Formed EPA Implementation Team
Developed a plan to incorporate probabilistic tools into eco assessment process
Included a 4 level refined risk assessment scheme, which was peer reviewed and strongly supported (2000)
What Does This New Assessment Scheme Look Like?
Level I: Screening Level Assessment
Level II: Preliminary assessment of probability and magnitude of effects
Levels III and IV: More refined assessments, representing increasingly focused biological and exposure scenarios
Where Are We Now and What Do the Models Look Like?
Developed pilot aquatic and terrestrial models using carbofuran as a case study
Peer reviewed and supported as "state-of-the-art" (2001)
Finalizing models, which will be used in Level II assessments
Highlights of Aquatic Model
Aquatic Organisms and Endpoints
Freshwater Fish and Invertebrates
Effects: Distributions based on D/R curve from lab acute toxicity data
Exposure: PRZM/EXAMS or monitoring data used to generate a distribution of exposures
Risk: Joint probability function
Example of Aquatic Results
On average, 11% mortality is estimated, confidence limits of 2 - 27%. Infrequently (5% of time), Daphnia would be expected to incur ∼ 41% mortality (5 - 89%).
|5th percentile||2 (<1-7)||94 (86-99)|
|Median||7 (1-22)||98 (92-100)|
|Mean||11 (2-27)||96 (89-98)|
|95th percentile||41 (5-89)||100 (99-100)|
Highlights of the Terrestrial Model
Acute mortality to avian species
Oral exposure from food and water
Spatial scale at the field level
Temporal scale is for exposure at and following a single application
Uses distributions based on the dose/response curve to estimate and address
Intra-species toxicity variability
Inter-species toxicity variability
Exposure model: Incorporates distributions for parameter values, expressing exposure as a probability curve
Residue variables considered:
Application rate and method
Distribution of residues and degradation rates
Behavior variables considered include the distribution of
what, how much, and where the bird eats
how much and where the bird drinks
how much the bird breathes
how much foliage the bird contacts
Exposure and effects distributions combined using Monte Carlo sampling techniques
Example of Terrestrial Results
How Are Risk Managers Involved?
To ensure assessment results can be used in regulatory decision-making, an ongoing dialogue is needed:
Senior Risk Manager Briefings
Risk Manager Workshops
Focus of this dialogue is to ask
Is the risk characterization presented clear?
Does the information regarding magnitude, probability, and certainty of risk provide you with information that will help you make a regulatory decision?
What Did the Regulators Say?
Strongly supported moving toward probabilistic analyses
Appreciated scientific advancement
Requested more training and continuing and frequent dialogue to ensure they understand the models, assumptions, and limitations
Encouraged including information in the training sessions and in the risk characterization on why ecological problems should be addressed and how they tie into human health
Recommended results be presented in various formats (graphs, charts, etc.)
Comments Related to Decision-making
Encouraged expanding assessments to evaluate mitigation options:
Range of application rates, such as high, low and typical
Timing and frequency of application methods, alternative methods, and changing pesticide formulation
Providing comparative analyses of alternatives using Level II assessments
Expressed concern for the use in registration decisions, where species tested are very limited
Encouraged providing "context" for results such as
Indications of the importance of species impacted
A connection with reproductive impacts, such as whether birds reproduce quickly and can recover from acute impacts
Reference point, such as a comparison to "normal" background mortality
Estimates of population impacts
Complete Level II model development, conduct peer review, and implement
Develop guidance to move between the levels
Continue risk manager training sessions
Provide more "advanced" training sessions
EPA Implementation Team
|Chair||Ingrid Sunzenauer, M.S.|
|Aquatic Team||Donna Randall, M.S. (Lead)|
Tim Bargar, Ph.D.
James Lin, Ph.D.
|Terrestrial Team||Ed Fite, M.S. (Lead)|
Ed Odenkirchen, Ph.D.
|Other Team Members||Tim Barry, Sc.D. (Office of the Administrator)|
Douglas Urban, M.S.
Stephanie Irene, Ph.D.
David Farrar, M.S.
|EFED Technology Teams|
|Former Team Members||Kathryn Gallagher, Ph.D.|
Les Touart, Ph.D.