Modeling the dispersal of wind-borne pests: Sensitivity of infestation forecasts to uncertainty in parameterization of long-distance airborne dispersal

Modeling dispersal of wind-borne pests can be a valuable tool in a broader context of areawide integrated pest management. In such models both biotic and abiotic factors play important roles. Many aspects of the life cycle and dispersal of the focal species are typically well studied and represented with robust measures. However, some important assessments may be either lacking or based on difficult to obtain measurements. For example, investigations of population processes occurring at broad spatial scales pose logistical difficulties, and consequently the corresponding estimates may be biased. This is true in the case of long-distance airborne dispersal, in which aloft characteristics of dispersal are particularly difficult to estimate. In this study we assessed the sensitivity of regional infestation forecasts of sorghum by sugarcane aphids [Melanaphis sacchari (Zehntner)] to uncertainty in parameterization of measures meant to represent important aspects of long-distance airborne dispersal in the southern-to-central Great Plains of North America. We focused on dispersal duration, take-off height, and maximum dispersal height as three critical parameters that control wind-borne long-distance dispersal of aphids. The 63 investigated combinations of values of these parameters had largely equivalent region-wide impact on timing of first forecasted infestation (equivalence margin of 7 days) and the impact on the probability of forecasted infestation was equivalent (equivalence margin of 0.1). The spatial variability in estimates depended on local specificity of the modeled landscape, and it was larger in areas where infestations were less frequent, primarily in eastern Kansas, Oklahoma, and northern Texas.

Automated summary

Modeling dispersal of wind-borne pests is a valuable tool in integrated pest management, with both biotic and abiotic factors playing important roles. Uncertainty in key parameters can impact regional infestation forecasts, with spatial variability depending on local landscape specificity, particularly in areas with less frequent infestations.