A forecasting model for providing an early warning of corn earworm presence

The generalist lepidopteran pest, corn earworm (Helicoverpa zea), migrates northward annually from southern United States (US) inflicting damage on a diverse set of crops throughout the US and southern Canada. Substantial effort has been placed on understanding the drivers of migration and basic migration patterns, but despite this sound ecological knowledge, forecast-based decision support tools are limited. Routine corn earworm monitoring in agricultural areas has generated decades of near-weekly occupancy data dotted across much of the eastern US. Leveraging this dataset, we apply a novel machine learning forecasting model to predict weekly occupancy at the county level up to three weeks in advance. Our forecast serves as an early warning system for pest management decision support and may facilitate earlier detection, optimize the timing of early season pesticide applications, and ultimately reduce crop damage.