Probability of emerald ash borer impact for Canadian cities and North America: a mechanistic model

Abstract

Emerald ash borer (EAB; Agrilus planipennis, Coleoptera: Buprestidae) causes large economic costs by killing ash trees (Fraxinus spp.): a process that takes several years of infestation. We suggest that the most important prediction regarding this species is not whether EAB can persist indefinitely in a new region, but whether it can persist long enough to kill trees. We use a mechanistic model of overwintering mortality of EAB prepupae to identify Canadian cities and more generally, those areas of North America at risk of impact. Although we have previously used a Newtonian cooling model to predict underbark temperatures of ash from meteorological data, we show that a linear regression model has smaller errors for the low winter temperatures relevant to EAB mortality. Using this regression model we generate distributions of predicted underbark temperatures which we then use to predict the return time of weather events cold enough to cause either complete (99%) or partial (75%) mortality of overwintering EAB prepupae. We find that most of North America does not experience extreme cold events frequently enough to prevent ash mortality from EAB (i.e., more frequently than every 6 years), and therefore conclude that large economic impacts are likely throughout the continent. However, if relatively frequent partial mortality events are sufficient to reduce ash mortality, there is a possibility of northern refugia for ash species, and some northern Canadian sites may escape the costs of this non-native pest.

Appendix

We used gridded daily extreme winter minimum data for Canada as the input to the best fit Newtonian cooling model (daily data from this source was not available for all of North America) to predict the probability of EAB mortality events. The Newtonian cooling model based on daily minimum temperatures predicted more frequent complete mortality events (99% mortality) in Canada than the linear regression model, and consequently a smaller area of economic risk. The Newtonian model predicted some small areas inside the native range of ash in Canada that would experience complete mortality every 1–3 years (Fig. 8). However, the area predicted for frequent complete mortality events excluded most Canadian urban centers. Therefore we interpret this model as predicting the potential for a small northern refuge for ash, but large economic impact for many Canadian cities. Of course, for 75% mortality events, a Newtonian cooling model predicts a much larger potential refuge for ash. This prediction suggests that that northern cities such as Edmonton and Saskatoon might escape significant economic impact of this invasive pest.

Fig. 8

Return time for a 99% EAB prepupae mortality event as predicted by the probability of − 35.4 °C underbark temperatures given by a Newtonian cooling model relating daily air temperature minimum to underbark temperature for annual winter minimum temperatures from 1993–2013. In addition to return times, areas where air temperatures cold enough to cause underbark temperatures colder than − 35.4 °C have been observed at least once between 1993 and 2013 (light blue). Circles indicate major urban centers in Canada considered (see Table 1). Purple line gives range of four major Fraxinus species in North America as indicated by Little (1971)