Modeling within-tree distributions of Agrilus planipennis

Emerald ash borer (Agrilus planipennis Fairmaire), a beetle native to Asia, was first detected in southeast Michigan in 2002 and has since killed millions of ash trees (Fraxinus spp.) throughout eastern North America. Determining the extent and density of outlier infestations of A. planipennis is a challenging aspect of effective management and slowing the rate of spread. Here, we model the probability of detection of A. planipennis larvae from measurements of tree characteristics on 71 girdled trees at a low-density outlier infestion in southwestern New York State. We conducted intensive tree sampling by one meter increments, recording the presence of larvae, height above ground, diameter, and bark roughness for each individual bolt during spring of 2011 and 2012. We modeled presence or absence of A. planipennis using logistic regression with a binomial response distribution. Probability of detection was modeled with a quadratic equation to detect the maximization, or vertex, of the function. Analyzing height and diameter separately, the probability of A. planipennis detection was maximized at 18.8 cm for diameter and 13.8 m for height. There was no conclusive relationship between bark roughness and larval presence. Our results can be used by resource managers to improve efficiency of detection efforts when evaluating new A. planipennis infestations.