|
(106) Simple Probabilistic Techniques to Determine Risk to Birds Exposed to Chlorpyrifos in Apple Orchards. Hart, Andy1, Moore, Dwayne2, Hicks, William3, Pawlisz, Andrew2, Teed, Scott2, 1 2 3 ABSTRACT- Probabilistic risk analyses are being conducted on a routine basis in higher-tier ecological risk assessments. The technique of choice has been first-order Monte Carlo analysis. However, Monte Carlo analysis is often not the best technique for propagating uncertainty, particularly in data-poor situations. Two other simple techniques, interval analysis and moment propagation, have been proposed as alternate techniques when data are scarce. In this presentation, we compare and contrast three uncertainty propagation techniques using a case study involving blue tits exposed to chlorpyrifos in apple orchards. For all three uncertainty propagation techniques, the results indicated a significant probability of mortality for only a small proportion of individuals in the most sensitive species. This result is consistent with the results of field studies, which found only limited and transient depression of plasma cholinestase activity in birds captured in orchards after spraying. The case study illustrated several advantages and disadvantages for each of the uncertainty propagation techniques. For example, interval analysis is the easiest technique to use and is useful for quickly screening out negligible risk scenarios. It cannot, however, be used to assign probabilities of effects when risk is not negligible. Moment propagation frees the analyst from having to specify distributions, is relatively easy to use, and produces probabilistic estimates of effects. The moment propagation formulae, however, become unwieldy when dependencies are introduced. Monte Carlo analysis is easy to perform, but requires more empirical data to properly specify and parameterize distributions. Key words: First Order Monte Carlo, Interval Analysis, Moment Propagation, Pesticides |
