COS 109-4 - The i-Tree Pest Predictor provides quantitative risk probabilities to identify high-impact tree-feeding insect invaders and the trees vulnerable to their attack

Abstract: Non-native insects may devastate forests over broad regions, yet not all introduced insects cause high impact. Our working group of federal, academic, and industry scientists examined the impact of introduced tree-feeding insects established in North American forest ecosystems with the goal of understanding why some cause widespread tree mortality. We hypothesized that four drivers may influence impact: 1) insect traits, 2) host tree traits, 3) whether a close relative of the invading insect already feeds on the North American (novel) host, and 4) the evolutionary relatedness between the insects’ native and novel tree hosts. We compiled data for each driver for tree-damaging conifer specialists, hardwood specialists, and generalists, resulting in a total of 248 insects established in North America. Across all insect groups, the evolutionary history between the insect’s native tree host with its novel host was found to be highly predictive of impact. We also found that phylogenetic diversity, a measure of native host breath, was an informative descriptor of an insect’s specialist or generalist feeding. Five predictive models were developed with each applicable to a specific insect group: 1) sapfeeders on conifers, 2) non-sapfeeders on conifers, 3) Scolytine specialists feeding on hardwoods, 4) non-Scolytine hardwood specialists, and 5) hardwood generalists. Predictive models also incorporated traits of the insect’s novel host (e.g., shade tolerance, drought intolerance) or insect traits (e.g., feeding guild) depending on the model. The models were incorporated into the i-Tree Pest Predictor (i-TPP) tool (https://pestpredictor.itreetools.org/). Given a suite of information for a potential new forest insect pest, the i-TPP tool provides a risk prediction of the probability the insect will cause high impact on over 400 North American tree species and the likelihood the insect will use each tree. The tool does not predict if an insect will establish should it get introduced, but it does provide quantitative guidance for prioritizing prevention of potentially high-impact insect introductions at ports-of-entry, for example, as well as eradication of newly introduced, potentially high-impact insects. We have compiled the i-TPP required input data for over 300 tree-feeding European insects not known to occur in North American forests and derived the impact and host likelihood probabilities for each. We are currently examining how these predictions can be used to understand which tree-feeding insects, should they arrive and establish, pose the greatest threats to North American forest ecosystems and which ecosystems are most vulnerable to these threats.