You are here:
USING POPULATION MODELS TO EVALUATE RISK IN POPULATION OF BIRDS
Citation:
NAGY, L. R. AND N. H. SCHUMAKER. USING POPULATION MODELS TO EVALUATE RISK IN POPULATION OF BIRDS. Presented at EPA Science Forum, Washington, DC, May 16 - 18, 2006.
Description:
Wildlife populations are exposed to varying habitat structure and quality, as well as an array of human-induced environmental stressors. Predicting the consequences to a real population of one perturbation (e.g. a pesticide application) without considering other human activities and naturally changing environmental conditions is unrealistic and frequently results in inaccurate predictions. The U.S. EPA has been challenged to develop risk assessment tools that predict with reasonable accuracy long-term effects of pesticides on songbird populations. Current methods address only the fate of individual animals exposed to a single stress, viz. pesticides. Here, we parameterized the PATCH wildlife simulation model (a spatially-explicit, individual based life history simulator) with data from a 3-year study of the western bluebird (Sialia mexicana) to evaluate 1) the effects of a hypothetical pesticide alone and in combination with environmental variability, 2) the population-level differences between a widespread pesticide and a pesticide that is applied locally, and 3) the time until recovery when a pesticide is applied and then discontinued. Our modeling results elucidated multiple important points with regard to pesticide applications. First, our modeled population had a greater, and unpredictable response (e.g., extinction) when the pesticide was combined with high environmental variability. Second, changes in population size were highly dependent on the habitats associated with the pesticide application; therefore, assuming a widespread application of a pesticide when only local application are more likely will overestimate the impacts of the pesticide. Finally, even when pesticides were applied for a narrow time window (e.g. 10 yrs), populations required 10 to 30 years to recover to their original size, even when no new stressors occurred. We conclude that without addressing natural stressors, the geographic extent of the pesticide application, and the timeframe of the pesticide applications, models addressing population-level impacts of pesticide applications may over or under-estimate the impacts of pesticides on the wildlife populations of interest.