FIFRA Scientific Advisory Panel (SAP) Questions for the SAP: March 13 - 16, 2001 Terrestrial PRA Review

Focal Species Selection Goals: The goals for selection of bird species serving as the focus of the risk assessment were to
1. advance the assessment beyond consideration of "generic" bird types so as to consider appropriate biological conditions associated with the treated environments, and
2. identify the types of species potentially at greatest risk from ChemX exposure at the corn and alfalfa use sites identified above.
Under the current EFED status of probabilistic risk assessment development, the use of focal species in an assessment is limited. The likely lack of species-specific toxicity data engenders considerable uncertainty regarding the prediction of the magnitude of mortality in any single bird species. Rather, the use of focal species is targeted to represent a myriad of potential species of similar biological/behavioral characteristics, yet retain some specificity as to the type of organisms using a treated area.
- What is the Panel's opinion regarding this approach?
- What are the Panel's recommendations regarding alternative approaches that should be investigated?
- Drawing upon your experiences and knowledge of avian foraging strategies, habitat use, and other interactions with the agroenvironments incorporated into the assessment, what recommendations can the Panel make for alternative or additional focal species for the crop/region combinations investigated in the assessment?
Frequency of Birds in Treated Fields use in the Model: EFED recognizes that additional research on quantifying exposure of bird species in agroenvironments will be critical to the advancement of the probabilistic risk assessment approach. At this juncture, EFED is mindful of the severe limits of existing avian census data for establishing such exposure estimates. The present method for considering the avian census data has been designed not to over represent the census data to the point that sightings of birds on or off a treated field is considered commensurate with proportional feeding on and off the field in a given day. Instead, the risk assessment uses the data to determine the likelihood that a bird will be on a treated field in a given time step, based on past field study history of sightings for that species.
- What are the Panel's thoughts regarding the use of avian census data in the model and was it used appropriately?
- What are the Panel's suggestions regarding alternative approaches to using the data? Please discuss any advantages and disadvantages to these alternative approaches.
Frequency of Birds in Treated Fields, Setting Parameters: EFED elected to establish minimally biased truncated exponential distributions for this parameter for each focal species.
- Upon looking at the available field study data (see spreadsheets included in SAP package), what are the Panel's thoughts on these selections?
- What alternative approaches for these distributions would be appropriate for the data sets available?
Consideration of Drinking Water Source Selection: Because of the paucity of data for drinking water source selection in birds of agroenvironments, EFED has identified investigation of this behavior as an area meriting further research. However, in the interim, EFED has made an assumption that drinking water selection is opportunistic and that use of on-field water sources is linked with bird presence on the field during any particular time step (see questions on frequency of birds using treated fields).
- What are the Panel's views on this interim procedure?
- What recommendations can the Panel make for future alternative approaches, considering the data currently available to EFED?
Puddle Persistence: EFED recognizes that on-field puddles may be more persistent than the half-day assumption incorporated in the model. However, EFED does not currently have a way of modeling puddle duration in a field.
- What are the Panel's recommendations regarding data sources and/or modeling approaches to establish the frequency of occurrence, dimensions, and duration of puddles in agroenvironments? (Rhetorical: When does a puddle become a pond? )
- Concentrations in Drinking Water:
- In the absence of a more rigorous model for pesticide residue in dew, EFED has used a simple two-compartment partitioning model. In the Case Study document, EFED has discussed some critical limitations to this model.
  - What model modifications can the Panel suggest for improving the estimation of pesticide residues in dew, keeping in mind the limitations of the current registration data set requirements?
  - What any data sources regarding dew measurement or modeling should be considered to strengthen EFED's modeling needs?
- EFED used an instantaneous maximum approach from direct application for modeling pesticides in puddles present in a treated field on the day of application. This approach, based on experiments with buried buckets of water, does not consider partitioning kinetics with field soils.
  - What does the Panel suggest for improving this approach?
- Pesticide residues in puddles appearing on a treated field on the day after pesticide application were estimated using output from PRZM runs. The operative assumption in this approach is that water puddling on the surface of the field would be equivalent in pesticide concentration to the water potentially running off the field from a given precipitation event. EFED recognizes a number of limitations to this approach, not the least of which is the sensitivity of the PRZM model output to application date in terms of number and magnitude of run-off event pesticide loadings following the application date.
  - What are the Panel's suggestions for adapting existing tools to this task and improving the modeling of day-after treatment puddle residues?
Residues in Vegetation Food Items: In the Case Study, EFED chose to base estimates of pesticide residues on the data provided in Fletcher et al. (1994). EFED also had data on actual ChemX residues after treatment, but these data were severely limited in sample number and in a very limited number of field study areas. EFED based the decision to use Fletcher et al. on a desire to utilize a more robust, albeit non-chemical specific, data set to establish initial pesticide residues.
- What is the Panel's opinion regarding EFED's approach, which relies on larger more generalized data sets versus focusing on limited single-chemical data sets for estimating initial field concentrations of pesticide residues?
Residue Clearance in Focal Species: EFED has included a residue retention factor in the exposure model. This factor was intended to account for carryover of a proportion of a time step's body burden to the next exposure time step, allowing for some consideration of cumulative exposure.
- What are the Panel's thoughts on this approach?
- Would the Panel recommend approaches to account for uncertainty associated with extrapolating poultry metabolism data (the origin of the residue retention factor used in the Case Study) to other bird species?
- Please comment on any need for pharmacokinetic studies to improve this assessment approach. What are the Panel's recommendations on the scope of such studies (i.e., appropriate species, number of species, study design, and endpoints)?
Selection of Exposure Time Steps: The Case Study divides each exposure day into two time steps. There have been questions regarding the effect of exposure time step and feeding duration upon the outcome of the model. To investigate this, EFED used a mass balance equation with two compartments, to determine the effect of rate of food consumption combined with clearance rate, on total body burden (ingested + residual) of pesticide. The results of this exercise led EFED to conclude that duration of the exposure window within a time step had minimal effect on maximum instantaneous body burden of ChemX.
Given the lack of detailed pharmacokinetic data for ChemX in birds, what other analyses of this issue could be made?
- Selection of Acute Toxicity Standard: In the Case Study, EFED used the acute single oral dose studies (LD₅₀) as the basis for characterizing effects in preference to the acute dietary exposure studies. Both of these study designs have limitations for estimating the risk to wild avian species exposed to pesticides in the environment. Both studies have a fixed exposure period, not allowing for the differences in response of individuals to different duration of exposure. For the acute oral study, the dose administered in a single dose all at one time does not mimic wild birds' exposure. Also, for exposure through different environmental matrices, it does not account for the effect of the matrices on the absorption rate of the chemical into the animal.
  This latter criticism also applies to the dietary test for other food matrices consumed in the wild. For the dietary test, the endpoint is reported as the concentration mixed with food that produces a response rather than as the dose ingested. There are a number of study limitations that render conversion of dietary concentration to dose problematic (e.g., food spillage, and quantifying food intake for individuals). The interpretation of this test is also confounded because the response of birds is not only a function of the intrinsic toxicity of the pesticide, but also the willingness of the birds to consume treated food.
  
  More importantly, there is evidence for some compounds that the laboratory derived LC50 values are poor predictors of effects in the field.
- What are the Panel's thoughts and recommendations regarding alternative approaches using LC50 data that allow estimation of dose-response relationships (critical to prediction of magnitude of effects)?
- Defining the Distribution of Species Sensitivity: EFED has selected a combination of methods to establish the distribution of species sensitivity to ChemX, from which representative points, low, medium, and high sensitivity are used to characterize the uncertainty regarding response of any particular species to ChemX exposure. The first step in the method involves normalizing all available toxicity data to a constant body weight using weighting factors established for specific and generic chemicals. The mean and standard deviation of these normalized values is taken. Then the method of Aldenberg and Slob (1993) is used to estimate the 5^th and 95^th percentiles of a log-logistic distribution with mean and standard deviation as defined above. Finally these point estimates of sensitivity are readjusted to a focal species body weight.
  - What are the Panel's views on this method to estimate the distribution of sensitivity of focal species when species-specific toxicity data are not available?
  - How would the Panel estimate the confidence interval surrounding the above estimates of the 5th and 95th percentiles?
- Other Factors Affecting Sensitivity: By relying on laboratory LD50 data, we have not accounted for a variety of physiological and environmental factors that may modify sensitivity (e.g., age, nutritional status, temperature, etc.) . EFED has concerns that not accounting for a number of these variables may lead to an underestimation of risk.
  - What guidance can the Panel offer that would allow, in the absence of chemical specific data, for a consideration of these physiological and environmental factors?

FIFRA Scientific Advisory Panel (SAP) Questions for the SAP: March 13 - 16, 2001 Terrestrial PRA Review

Local Navigation