Abstract:
|
Nonlinear regression is a powerful tool that can be used in many different ecological applications to make predictions and provide inferences on relevant population parameters. In this study, a nonlinear logistic growth function was utilized to model the cumulative aphid count data over time. The model was parameterized to provide inference on three relevant characteristics, onset of aphid accumulation, the rate of increase in aphid accumulation, and the maximum cumulative aphid count. A clustering algorithm was employed to group the data into similar environments. Incorporation of a dummy variable for environments allowed for the comparison of environments via likelihood ratio tests. To account for potential autocorrelation, an autoregressive structure was incorporated into the maximum cumulative count parameter. Predicted models were validated both externally, using independent data, and internally, using bootstrap simulation of the residuals. Applications are demonstrated using 17 years of wheat aphid suction trap data for four species across 12 locations in Idaho.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.