Submitted to: Pest Management Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 10/25/2024
Publication Date: 11/14/2024
Citation: Broadhead, G.T., Higbee, B.S., Beck, J.J. 2024. Evaluating the use of in-season measures of pest abundance to predict end-of-season damage: a study in commercial almond (Prunus dulcis). Pest Management Science. http://doi.org/10.1002/ps.8532.
DOI: https://doi.org/10.1002/ps.8532

Interpretive Summary: Effective insect pest management in agricultural commodities is critical for minimizing crop loss, and early warning signs of insect pest abundance or imminent crop loss can guide management decisions, thus enabling targeted interventions to mitigate loss. For many insect pest/crop plant systems, land-use, climate, and insect trapping data have been used in an ecological informatics approach to produce predictive models of insect pest abundance and population phenology. Accordingly, given the acquisition of appropriate data, in-season measurements of pest abundance may be used to predict damage prior to harvest and enable early management interventions. In commercial almonds, the abundance of Amyelois transitella (navel orangeworm, NOW), a primary pest of almond, is correlated with almond damage; however, accurate predictions can be hindered by the attractiveness of various trap types and the use of mating disruption. Insect lures based upon odors from the insect pest’s host plant (kairomones) have been suggested as an alternative trapping approach which may improve the accuracy of measures of NOW abundance in predicting end of season damage. USDA-ARS scientists from the Center of Medical, Agricultural, and Veterinary Entomology in Gainesville, FL, in collaboration with an almond industry expert, utilized an existing data set collected in a commercial almond orchard and evaluated the performance of multiple trapping techniques (including kairomone lures) under different management methods. Their results showed a significant correlation between pest abundance and damage across multiple trap types. However, further exploration revealed that the estimates of NOW density do not significantly enhance predictive models, yet an accuracy of > 70% can be achieved in the absence of these data. Their results provide further evidence that prediction of end of season damage can be achieved from early season trapping of the insect pest.

Technical Abstract: Pre-harvest pest management is essential to minimizing crop loss, and early warning signs of pest abundance or imminent crop loss can guide management decisions and enable targeted interventions to mitigate loss. In multiple insect pest-host plant systems, land-use, climate, and insect trapping data have been used in an ecological informatics approach to produce predictive models of insect pest abundance and population phenology. Similarly, with sufficient data, in-season measurements of pest abundance may be used to predict damage before harvest and enable early management interventions. However, sampling plans tracking insect phenology and sampling insect abundance to inform targeted, informed pest-management may not be equivalent. In commercial almonds, the abundance of Amyelois transitella (navel orangeworm, NOW; a primary pest of almond), is correlated with almond damage; however, accurate predictions can be hindered by the attractiveness of various trap types and the use of mating disruption. Kairmone lures have been suggested as an alternative trapping approach which may improve the accuracy of measures of NOW abundance in predicting end of season damage. We explore an existing data set collected in a commercial almond orchard and evaluate the performance of multiple trapping techniques (including kairomone lures) under different management methods. Our results show a significant correlation between pest abundance and damage across multiple trap types, regardless of management method. However, further exploration revealed that these estimates of NOW density do not significantly enhance predictive models and accuracy of > 70% can be achieved in the absence of these data. We discuss the implications of this outcome and avenues for improvement of NOW sampling predictive models using estimates of NOW population density.