covered in the site selection process.
2.) EFED is recommending that pesticide use area be the primary domain. In order to better define exposure levels to specific human populations and to better understand the processes affecting pesticide concentrations in surface source drinking water, it may be useful to use smaller domains. Such domains might be spatial or based on vulnerability criteria.
Should domains other than a pesticide's use area be used (regional domains, for
What are the Panel's views on minimally acceptable data quality standards if the
study had a larger number of domains.
Are there any suggestions for optimizing survey design to maximize the number of
target pesticides represented while minimizing the total number of sampled CWSs without sacrificing data quality?
Could preferential selection of CWSs with source watersheds containing uses for
multiple pesticides be incorporated into the design?
Many of these questions could be better answered after the objectives and some design ideas have been better defined. Even though the Panel agreed that additional analysis is needed, the SAP generally concurred with the Agency that the primary domain should be the pesticide use area. With the limited sampling that is being proposed, this seems the best way that reliable estimates of exposure for the selected pesticides can be made. Strictly geographic or spatial domains seem difficult to justify given the objectives. A set of regional domains can be applied, but only if the sampling rate were much higher. It could be advantageous to create subdomains of urban and rural environments, since the exposure for these two groups could be very different. Creation of subdomains, however, requires more sampling and must be carefully evaluated and justified.
In producing the pesticide use domains, however, better pesticide use data than that proposed in the study should be used if possible. The Agency has access to confidential use statistics that might produce estimates of use for recent years and sometimes for sub-state areas. While these data cannot be made public, it might be used to generate more precise and up-to-date pesticide use areas without endangering its confidentiality. Also, recent data collected by USDA should be used in conjunction with the Agency database. A further recommendation was averaging use over the most recent three years for which data are available to capture year-to-year variation in pesticide use (which in turn reflects year-to-year variation in crops planted and pest problems that are often linked to weather).
The spatial link between pesticide use domain (on the spatial level) and the actual sampling location (the public water system) is the watershed or source water area. This link should be used to optimize sample selection and design. Rather than using simple pesticide use (mass) as a criterion, mass normalized to area (mass/area) might be better to use as the basis for identifying